---
title: "Outbound Is Dying: How Spara's Multimodal AI Turns Inbound Into Pipeline (Live Demo)"
episode: 42
podcast: "The LeanScale Podcast"
publisher: "LeanScale"
guest: "David Walker"
guest_title: "Founder & CEO, Spara"
date_published: 2025-10-29
date_modified: 2026-07-22
duration: 00:29:44
word_count: 5462
topics: ["ai-in-gtm", "outbound-sales", "demand-generation", "gtm-strategy"]
canonical_url: https://leanscale-knowledge-hub.netlify.app/podcast/david-walker-spara-multimodal-inbound/
source: "LeanScale Podcast Knowledge Hub — https://leanscale-knowledge-hub.netlify.app"
license: "Free to quote and cite with attribution to The LeanScale Podcast."
---

# Outbound Is Dying: How Spara's Multimodal AI Turns Inbound Into Pipeline (Live Demo)

_David Walker on why outbound is losing signal, and how one multimodal AI layer — chat, email, and voice — converts the inbound traffic every buyer already generates_

**Episode 42 · The LeanScale Podcast**  
David Walker, Founder & CEO, Spara · Hosted by Anthony Enrico  
Published October 29, 2025 · Updated July 22, 2026 · 00:29:44  
Canonical: https://leanscale-knowledge-hub.netlify.app/podcast/david-walker-spara-multimodal-inbound/

**Topics:** AI in GTM · Outbound & Sales Development · Demand Generation · GTM Strategy


## Executive summary

The conventional AI-for-GTM story is about outbound: more sequences, more personalization, more volume. David Walker — founder and CEO of Spara — thinks that story is already over. In this live-demo episode with LeanScale co-founder Anthony Enrico, he argues that AI-powered outbound is subject to a reverse network effect: the more people who use it, the less any individual email works, so the signal-to-noise problem only gets worse. The real, durable leverage is on the other side of the funnel — converting the inbound interest that every buyer generates before they ever buy.

Walker's credibility is founder-operator, not vendor. He previously co-founded and ran a real-estate SaaS business he grew to 350 people and sold to a company with a couple thousand employees, spending most of his time on go-to-market and selling. Across both orgs he kept hitting the same wall: people were interested, they were coming to the website, and there was no fast, human-quality way to talk to them and learn the questions they actually had. He started Spara to solve that as his 'life's work' — a single multimodal model trained on a company's own data that can meet buyers in chat, email, or voice, in real time, and move them from lead to pipeline to revenue.

The heart of the episode is a working demo of the platform. Walker shows two front-end products — Navigator (the bottom-right chat everyone recognizes, but free-form with no gating or if-then logic) and Smart Bar ('ChatGPT for your website'). The through-line is a sales philosophy: don't gate the conversation behind a form. Let buyers ask questions, provide value first, and weave pre-qualification in — because the question a buyer asks is the roadmap to close, and that proprietary intent data can't be bought anywhere. From there he demos instant scheduling with no form, real-time routing (AE vs. SDR vs. self-serve), fully autonomous email that replies within seconds and continues the exact chat conversation, an AI voice pre-call that phones the buyer after they book to tailor the human demo, and lead enrichment that assembles a ready-to-close playbook and pushes it to the CRM and to the rep via Slack.

Two numbers frame the stakes. Anthony notes LeanScale's Series A–C customers manage 30+ GTM tools on average, and Series D and above run up to 70 — the sprawl Walker expects AI to consolidate the same way pre-LLM SaaS did. And Walker cites 2–4x greater engagement and MQL capture from ungated, value-first conversation versus traditional form-gated chat. Throughout, he is emphatic that this is not about replacing sellers: 'magic happens when a motivated buyer and a really good seller get together,' and Spara's whole goal is to get more qualified people onto the phone with humans, faster than humans alone ever could.

Who should listen: founders and CROs rethinking their buying experience, marketing and demand-gen leaders trying to convert website traffic into pipeline, RevOps leaders drowning in tool sprawl and dirty CRM data, and sales leaders who want speed-to-lead, real-time routing, and better-prepared reps. The biggest takeaway is a reframe of where AI belongs in GTM — not manufacturing more outbound noise, but being the always-on, multimodal layer that makes the inbound experience fast, helpful, and human enough that motivated buyers convert.


## Key takeaways

1. **AI outbound has a reverse network effect — the more it's used, the less it works** — Walker argues the signal-to-noise problem is getting structurally worse: as more teams use AI to send 'fake personalized' outbound at volume, each individual email becomes less effective. It's a reverse network effect, and it's why he built an inbound platform rather than another outbound engine.
   _Why it matters:_ Don't measure your AI investment by how much more outbound you can send. The marginal email is depreciating; the marginal inbound conversion is not. Shift leverage to where buyers are already raising their hands.
   _For:_ Founders, Sales Leaders, Marketing Leaders

2. **100% of buyers visit your website — it's the most-neglected conversion surface** — Whether a buyer came from inbound, outbound, or a referral, essentially everyone who buys a B2B SaaS (or high-consideration) product visits the website before purchase. Yet humans do a poor job there because they're too slow to talk to people in the moment they're actually engaged.
   _Why it matters:_ Treat the website — and the seconds after a buyer engages or leaves — as a first-class conversion motion, not a brochure. It's the one channel with 100% buyer coverage and almost no real-time human presence.
   _For:_ Marketing Leaders, Founders, RevOps Leaders

3. **One multimodal model beats a stack of point solutions** — Spara trains a single model on a company's data and deploys it across chat, email, and voice, so context carries seamlessly between modalities. Training a separate model per channel 'misses data in translation,' is less efficient, and gives a worse buyer experience.
   _Why it matters:_ As teams cobble together separate AI-chat, AI-email, and AI-voice tools, expect the same consolidation that hit pre-LLM SaaS. Buy for a consolidated AI layer, not a pile of single-channel bots you'll have to manage and reconcile.
   _For:_ RevOps Leaders, Founders

4. **Don't gate the conversation — value-first 'give to get' beats forms, 2–4x** — Most website chat is just a form in disguise: it demands an email before it says anything useful. Spara's unlock is to remove the gate — answer questions, show product value, and weave pre-qualification in. Walker reports 2–4x greater engagement and MQL capture than traditional gated chat.
   _Why it matters:_ Kill the email-first gate on your highest-intent buyers. Provide value before you ask for anything; buyers who get a genuinely helpful, free-form conversation are far more likely to hand over information than buyers who hit a form.
   _For:_ Marketing Leaders, Sales Leaders, RevOps Leaders

5. **The buyer's question is the roadmap to close — proprietary intent data you have to earn** — If you know the question a buyer actually wants to ask, you have the roadmap to close. Letting them ask reveals what they care about and what they're trying to solve — proprietary intent data you can't buy from any data vendor and only get by opening the conversation.
   _Why it matters:_ Instrument and capture the questions buyers ask on your site. That first-party intent is a compounding asset for messaging, product, and every downstream rep — but only if you give buyers the room to reveal it.
   _For:_ Marketing Leaders, Founders, RevOps Leaders

6. **Speed-to-lead is the whole game — humans are the latency, not the medium** — Walker's view: email is only a 'high-latency' channel because humans are slow. Remove that constraint and AI can reply within ~10 seconds, continue the buyer's exact conversation, and keep every modality real-time. Buying stalls when a buyer books, gets distracted, and never returns — the exact gap real-time follow-up closes.
   _Why it matters:_ Design for real-time across chat, email, and voice. The most common lost deal isn't a rejection; it's a distracted buyer who never came back — recoverable with instant, context-rich follow-up.
   _For:_ Sales Leaders, RevOps Leaders, Marketing Leaders

7. **Route in real time: AE, SDR, or self-serve — the way a good human would** — Spara runs qualification logic live: a highly qualified buyer gets an AE calendar, a mid-tier buyer gets an SDR calendar, and a low-tier buyer is pushed into self-service — decided in the moment from the conversation, not from a lagging lead score.
   _Why it matters:_ Real-time routing lets your most qualified, most eager buyers accelerate instead of being held back by a motion 'built for the average.' Stop slowing down the people most ready to give you money.
   _For:_ Sales Leaders, RevOps Leaders

8. **AI voice belongs in pre-call prep — not on the human demo** — Walker is blunt that if a buyer books a call expecting a human and gets an AI, that's a bad use of AI. The right place for AI voice is around the human moment: a call seconds after booking that asks a few tailoring questions, so the rep walks in knowing exactly what the buyer cares about.
   _Why it matters:_ Use AI voice to make human calls better, not to avoid them. Pre-call prep adds value to both sides — the buyer gets a tailored demo, the rep is set up to close — without pretending a bot is a person.
   _For:_ Sales Leaders, Founders

9. **Training is easier than teams fear — human-in-the-loop on assets you already have** — Spara trains on existing public and sales-enablement material — website, docs, academy content, good call transcripts — then uses a testing interface where you role-play your own customer and thumbs-up/down responses to refine tone. Walker says they're 'batting 1,000': no customer has had to create a net-new asset.
   _Why it matters:_ The blocker to deploying a GTM AI usually isn't creating content — it's structuring the data you already own and keeping a human in the loop to refine feel. Budget an hour or two of testing, not a content project.
   _For:_ RevOps Leaders, Marketing Leaders, Founders

10. **Enrichment should hand the rep a ready-to-close playbook** — On every lead, Spara runs background research and enrichment waterfalls — role, company, competitors, recent news, plus a full cross-channel conversation summary — structures it into the CRM, and proactively alerts the assigned rep via Slack and email so they arrive fully prepared.
   _Why it matters:_ The payoff of an inbound AI layer isn't just capture; it's a briefed sales team. Pair conversation data with enrichment and push it to the right rep at the right time so reps spend time closing, not researching.
   _For:_ RevOps Leaders, Sales Leaders

11. **This augments sellers — the goal is more humans on the phone, not fewer** — Walker repeatedly frames Spara as 'super pro salespeople': the AI layer exists to deliver the real-time value humans can't (instant answers across every interface, at the moment of intent) and to get more qualified buyers connected to real humans faster.
   _Why it matters:_ The right question isn't 'can AI replace the seller?' but 'where are humans too slow to add value?' Automate the latency and the routing; protect the human conversation where a motivated buyer and a good seller create the magic.
   _For:_ Founders, Sales Leaders, RevOps Leaders

12. **Would you buy through your own buying motion? Most leaders wouldn't** — Walker asks CMOs, CROs, and VPs of sales whether they'd go through their own buying motion as a buyer. Roughly 90% give a wry smile and admit no — it takes too long ('nine days to show someone our product at best') for an attention span that operates in minutes.
   _Why it matters:_ Audit your buying experience from the buyer's seat. If your own leadership wouldn't tolerate the delay, your best-fit buyers won't either — and the fix is speed and value, not more follow-up cadence.
   _For:_ Founders, Sales Leaders, Marketing Leaders


## Frameworks

### One Model, Every Channel (Multimodal AI Layer) (02:01)

**Definition:** Train a single AI model on a company's own data, then deploy it across every channel a buyer wants — chat, email, and voice — so context and quality carry seamlessly between modalities.

Training a separate model per channel loses data in translation and creates a fragmented, less efficient experience. A consolidated multimodal layer meets buyers wherever they engage, in real time, and moves them from lead to pipeline without the seams of a cobbled-together stack.

### The Reverse Network Effect of AI Outbound (05:51)

**Definition:** As more teams use AI to send high-volume 'fake personalized' outbound, each individual email becomes less effective — a network effect that runs in reverse, degrading the whole channel as adoption grows.

The signal-to-noise problem in the inbox is getting worse, not better, because AI lowers the cost of sending. Walker uses this to justify building an inbound conversion platform rather than another outbound engine.

### Don't Gate It (Give to Get) (14:31)

**Definition:** Remove the email-first form gate from website conversation. Provide genuine value and answer questions first, then weave pre-qualification into the conversation — 'give to get,' classic sales.

Most website chat is a form in disguise that gates the thing the buyer wants most. Ungated, value-first conversation makes buyers far more willing to share information and, per Walker, drives 2–4x greater engagement and MQL capture than traditional gated chat.

### Proprietary Intent Data (The Question Is the Roadmap to Close) (09:37)

**Definition:** The specific questions a buyer asks are first-party intent data that reveals what they care about and are trying to solve — a 'roadmap to close' you can't buy from any data vendor and can only earn by opening the conversation.

Websites are built to answer every possible question, but the seller's real prize is knowing which question the buyer actually has. Capturing that gives the roadmap to close and compounds as an asset for messaging, product, and every downstream rep.

### Speed-to-Lead Across Every Modality (06:31)

**Definition:** Respond to buyers in real time in whatever channel they're using — chat, email, or voice — because the latency in most GTM motions is the human, not the medium. AI can reply within seconds and continue the buyer's exact conversation.

Email is only 'high-latency' because humans are slow; remove that limit and every channel becomes real-time. Buyers stall when they book, get distracted, and never return — instant, context-rich follow-up recovers exactly that lost population.

### Real-Time Smart Routing (AE / SDR / Self-Serve) (12:40)

**Definition:** Run qualification logic live in the conversation and route each buyer to the right next step: a highly qualified buyer to an AE calendar, a mid-tier buyer to an SDR, a low-tier buyer to self-service.

Most sales motions are 'built for the average' and hold back the most qualified, most eager buyers. Routing in the moment — the way a good human would feel out a customer — lets high-intent buyers accelerate instead of being slowed down.

### AI Voice for Pre-Call Prep (Not the Human Call) (19:20)

**Definition:** Use AI voice around the human moment, not in place of it: after a buyer books a demo, an AI call gathers a few tailoring questions so the human demo is more valuable to the buyer and the rep is better prepared.

Putting AI on a call the buyer expected to have with a human is a bad use of AI. The right application is enhancing the human interaction — 'magic happens when a motivated buyer and a really good seller get together' — by making the call more targeted before it starts.

### Human-in-the-Loop Training on Existing Assets (25:07)

**Definition:** Train the model on data a company already has — website, docs, academy, good call transcripts, sales-training material — then refine it in a testing interface where you role-play your own customer and thumbs-up/down responses to tune tone and accuracy.

The fear is that training is a huge content project; the reality is that no customer has had to create a net-new asset. The work is structuring existing data and keeping a human in the loop, and teams see visible improvement in an hour or two of testing.


## Quotes

_Speakers inferred from an undiarized transcript — verify before attributing._

> "Outbound is completely dying, especially if you're in B2B mid-market to enterprise SaaS or selling AI."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 42 (00:00)

> "We are multimodal, so we allow companies to train one model on their data and then we can talk to their customers in any form factor that their customers want to use — chat, email, voice, whatever it is."
>
> — David Walker, The LeanScale Podcast Ep. 42 (02:01)

> "The signal-noise problem is getting worse. AI makes it harder for everyone. It's kind of a reverse network effect. The more people using AI to send outbound emails, the less effective each individual email will be."
>
> — David Walker, The LeanScale Podcast Ep. 42 (05:51)

> "100% of people who buy are going to go to your website — whether they're inbound, outbound, wherever they came from. And that is a key area where humans have done a really poor job, because we're just not fast enough."
>
> — David Walker, The LeanScale Podcast Ep. 42 (06:31)

> "Would you go through that motion as a buyer? And 90% of the time there's a wry smile, and the answer is always the same. Honestly, no."
>
> — David Walker, The LeanScale Podcast Ep. 42 (07:41)

> "If we know the question that our customer is looking to ask us, we have the roadmap to close."
>
> — David Walker, The LeanScale Podcast Ep. 42 (09:37)

> "That's proprietary intent data. You have to earn it. You have to give them the opportunity to have that conversation. You can't buy that data anywhere else."
>
> — David Walker, The LeanScale Podcast Ep. 42 (10:10)

> "It's give to get. It's classic sales. We've been helpful, and so Spara has earned the right to go and ask pre-qualification questions."
>
> — David Walker, The LeanScale Podcast Ep. 42 (10:54)

> "So many sales motions today hold back your most qualified and interested buyers. The people who really want to give you money are actually held back and they feel like they're slowed down."
>
> — David Walker, The LeanScale Podcast Ep. 42 (13:19)

> "Aesthetically, seeing the bar there feels so much different than a chatbot. It feels more like I'm interacting with something that's going to give me actual helpful information, not just a sales tool trying to lure me into a demo."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 42 (13:50)

> "The big unlock here was just don't gate it. Let's provide value. Let's allow the person to ask questions. Let's be helpful."
>
> — David Walker, The LeanScale Podcast Ep. 42 (14:31)

> "We're seeing literally 2, 3, 4x greater engagement and MQL capture than pre-existing chat."
>
> — David Walker, The LeanScale Podcast Ep. 42 (15:41)

> "The only reason email is a high-latency form of communication is because humans are really slow. But if we remove that limiting factor, then it can be really, really fast."
>
> — David Walker, The LeanScale Podcast Ep. 42 (16:58)

> "If I show up to that call as the buyer and it's your AI on the call, I'm going to be pissed. That's a bad use of AI. I'm expecting that human touch."
>
> — David Walker, The LeanScale Podcast Ep. 42 (19:20)

> "Magic happens when a motivated buyer and a really good seller get together to talk about the same problem at the same time."
>
> — David Walker, The LeanScale Podcast Ep. 42 (19:58)

> "We're not trying to get rid of salespeople. We're actually super pro salespeople. Our whole goal is to get more customers on the phone with salespeople."
>
> — David Walker, The LeanScale Podcast Ep. 42 (27:34)

> "We're batting 1,000, where no company has had to create a net-new asset or piece of content for us. It's about making the data they already have super useful to train a model."
>
> — David Walker, The LeanScale Podcast Ep. 42 (26:23)

> "The average go-to-market tech stack we manage is over 30 different tools. If you go Series D and above, that gets up to 70 tools."
>
> — Anthony Enrico, The LeanScale Podcast Ep. 42 (04:29)


## Practical advice by role

### Founders

- Audit your buying experience from the buyer's seat — if you (or your CRO) wouldn't tolerate your own motion, your best-fit buyers won't either. Compress the time-to-value from 'nine days to see the product' toward minutes.
- Treat inbound conversion, not more outbound, as the durable AI leverage: outbound depreciates under a reverse network effect, while the inbound traffic every buyer generates is under-served.
- Buy for a consolidated, multimodal AI layer rather than stitching together single-channel AI chat, email, and voice tools you'll have to manage and reconcile.

### Marketing Leaders

- Kill the email-first gate on website chat. Provide value and answer questions first, weave pre-qualification in, and expect materially higher engagement and MQL capture than a form-in-disguise chatbot.
- Instrument and capture the actual questions buyers ask — that first-party intent data is a compounding asset for messaging and product that you can't buy anywhere.
- Own the website as a real conversion surface: 100% of buyers pass through it, and the seconds after they engage or leave are where pipeline is won or lost.

### Sales Leaders

- Engineer speed-to-lead across every channel: real-time chat, sub-minute email replies that continue the buyer's exact conversation, and instant scheduling with no form.
- Route in the moment — AE for highly qualified, SDR for mid-tier, self-serve for low-tier — so your most eager buyers accelerate instead of being held back by a motion built for the average.
- Use AI voice for pre-call prep, never to replace the human demo. Let it tailor the call and brief the rep so the human conversation is where the magic happens.

### RevOps Leaders

- Attack tool sprawl: 30+ tools at Series A–C and up to 70 at Series D+ is unmanageable overhead — favor consolidation as AI point solutions inevitably merge, the way pre-LLM SaaS did.
- Make the AI layer earn its keep on the back end: push structured conversation data and enrichment (role, company, competitors, news, summary) into the CRM and alert the assigned rep via Slack and email.
- Deploying a GTM AI is a data-structuring and human-in-the-loop tuning job, not a content project — plan for an hour or two of testing on assets you already own, not net-new material.


## AI takeaways

**Thesis:** The highest-leverage use of AI in go-to-market isn't sending more outbound — it's being the always-on, multimodal layer that converts the inbound traffic every buyer already generates, and augmenting sellers rather than replacing them. Outbound depreciates under a reverse network effect; real-time, value-first inbound conversion compounds.

- **Outbound is a depreciating asset** — AI lowers the cost of sending, so more AI outbound means each email works less — a reverse network effect that keeps degrading the channel. The durable leverage is on the inbound side.
- **One multimodal model, not a bot per channel** — Train a single model on your data and run it across chat, email, and voice so context carries between modalities. Separate per-channel models 'miss data in translation' and are worse and less efficient.
- **Real-time is the unlock** — The latency in GTM is human, not technological. AI can answer in chat instantly, reply to email in ~10 seconds, and place a voice pre-call — recovering buyers who booked, got distracted, and would never have returned.
- **Value first, gate never** — Removing the email-first form and leading with a helpful, free-form conversation captures proprietary intent data and lifts engagement and MQL capture 2–4x over gated chat.
- **Augment, don't replace** — AI belongs where humans are too slow — instant answers, routing, enrichment, pre-call prep. It should get more qualified buyers onto calls with real sellers, not put a bot on a call the buyer expected a human to be on.
- **Training is data structuring, not content creation** — Models train on assets companies already own (site, docs, transcripts) plus human-in-the-loop tuning; no customer has needed a net-new asset. Garbage in, garbage out still applies — the work is making existing data useful.

**Agent & automation ideas**

- A no-gate website conversation agent that answers freely, weaves in pre-qualification, captures the buyer's questions as first-party intent data, and routes AE vs. SDR vs. self-serve in real time.
- An autonomous email agent that continues the exact chat conversation context and replies within seconds of any buyer reply, keeping every modality real-time.
- An AI voice pre-call agent that phones a buyer moments after they book, gathers tailoring questions, and briefs the rep so the human demo is targeted from the first minute.
- A lead-enrichment and rep-briefing agent that runs enrichment waterfalls (role, company, competitors, recent news), summarizes the cross-channel conversation, writes it to the CRM, and pushes a ready-to-close playbook to the assigned rep via Slack and email.


## Operations takeaways

### Revenue operations

- **Consolidate the stack.** 30+ tools at Series A–C and up to 70 at Series D+ is unsustainable overhead; expect AI point solutions to merge into consolidated layers.
- **One model, every channel.** A single multimodal model avoids the data loss and management burden of separate per-channel AI tools.
- **Structure and sync the data.** Push conversation summaries plus enrichment into the CRM in a usable structure, then proactively alert the assigned rep — capture is only half the value.
- **Deployment is a data job.** Standing up a GTM AI is about structuring existing assets and human-in-the-loop tuning, not producing net-new content.

### Pipeline & marketing ops

- **Website is the 100% channel.** Every buyer visits the site before purchasing; converting that traffic in real time is higher-leverage than adding outbound volume.
- **Ungated beats gated 2–4x.** Removing the email-first form and leading with value lifts engagement and MQL capture 2, 3, or 4x over traditional chatbots.
- **Capture the question.** The buyer's question is proprietary intent data and the roadmap to close — instrument and store it, don't just answer it.
- **Recover the distracted buyer.** The most common lost deal is a buyer who booked, got pulled away, and never returned — real-time, context-rich follow-up brings them back.
- **Route by intent, live.** Qualify in the conversation and send high-intent buyers to an AE, mid-tier to an SDR, and low-tier to self-serve rather than slowing everyone to the average.


## Metrics mentioned

| Value | Metric | Context |
| --- | --- | --- |
| 30+ | GTM tools managed (Series A–C) | The average number of tools in the go-to-market tech stack LeanScale manages for its Series A–C customers — a sprawl AI is expected to consolidate. |
| Up to 70 | GTM tools managed (Series D+) | Series D and above run up to 70 go-to-market tools, making stack overhead and management a growing problem. |
| 2–4x | Engagement / MQL capture lift | Ungated, value-first conversation drives 2, 3, or 4x greater engagement and MQL capture than traditional form-gated chat, per Walker. |
| ~10 seconds | Autonomous email response time | Because humans are the only reason email is high-latency, Spara can reply to a buyer's email reply within about ten seconds. |
| 350 people | Prior company scale | Walker's earlier real-estate SaaS, which he grew to 350 people before selling it to a company with a couple thousand employees. |
| ~9 days | Typical time-to-demo | Walker's shorthand for how slow legacy buying motions are — 'nine days to show someone our product at best.' |
| ~10% | Leaders who'd buy via their own motion | Roughly 90% of CMOs, CROs, and VPs admit they would not go through their own buying motion as a buyer. |
| 0 | Net-new assets required to train Spara | Walker says they're 'batting 1,000' — no customer has had to create a net-new asset or piece of content to train the model. |


## Entities mentioned

- **Spara** (company) — David Walker's company; a multimodal AI platform (chat, email, voice) that trains one model on a customer's data to convert inbound website traffic into pipeline. Walker dogfoods Spara on Spara's own site. · https://leanscale-knowledge-hub.netlify.app/company/spara/
- **LeanScale** (company) — Anthony's firm; he cites LeanScale managing 30+ GTM tools on average for Series A–C customers and up to 70 for Series D and above, framing the tool-sprawl problem AI is set to consolidate. · https://leanscale-knowledge-hub.netlify.app/company/leanscale/
- **David Walker** (person, guest) — Founder & CEO of Spara, the multimodal AI platform (chat, email, voice) that converts inbound website traffic into pipeline; previously co-founded and sold a real-estate SaaS he grew to 350 people. · https://leanscale-knowledge-hub.netlify.app/guest/david-walker/
- **Anthony Enrico** (person, host) — Co-founder of LeanScale and host of The LeanScale Podcast. · https://leanscale-knowledge-hub.netlify.app/guest/anthony-enrico/
- **ChatGPT** (tool, AI Assistant) — Reference point for Spara's 'Smart Bar' product, described as 'ChatGPT for your website' — a familiar UI buyers trust to return valuable information.
- **Slack** (tool, Team Messaging) — Channel Spara uses to proactively alert the assigned sales rep with conversation notes and enrichment once a lead is routed in the CRM; also the distraction ('then I get a Slack, then I'm out of there') that causes buyers to abandon a demo booking.


## FAQ

**Q: What is Spara?**

A: Spara is a multimodal AI platform that trains a single model on a company's own data and then talks to that company's buyers in real time across chat, email, and voice. Its goal is to convert inbound website traffic into pipeline — answering questions, capturing intent, qualifying, routing, and scheduling — and to hand sales reps a briefed, ready-to-close handoff. It was founded by David Walker, previously co-founder and CEO of a real-estate SaaS he grew to 350 people and sold.

**Q: Why does David Walker say outbound is dying?**

A: Because AI-powered outbound suffers a reverse network effect: as more teams use AI to send high-volume, fake-personalized emails, each individual email becomes less effective and the signal-to-noise problem in the inbox gets worse. Walker argues the durable leverage has shifted to converting the inbound interest every buyer generates — since roughly 100% of buyers visit a company's website before purchasing — rather than manufacturing more outbound volume.

**Q: What does 'multimodal AI' mean for inbound conversion?**

A: It means one AI model, trained on a company's data, can meet buyers in whichever channel they prefer — chat, email, or voice — with context carrying seamlessly between them. Training a separate model per channel loses data in translation and creates a fragmented experience. A multimodal layer lets a buyer start in chat, continue by email, and take a voice call, all consistent and in real time.

**Q: Why shouldn't you gate website chat behind a form?**

A: Because a form gates the very thing the buyer wants — answers — and suppresses the conversation that reveals what they care about. Walker's approach is 'don't gate it': provide value and answer questions first, then weave pre-qualification in. Buyers who get a genuinely helpful, free-form conversation are far more likely to share information, and ungated chat drives 2 to 4x greater engagement and MQL capture than a traditional gated chatbot.

**Q: What is proprietary intent data and why does it matter?**

A: Proprietary intent data is the set of specific questions a buyer actually asks on your site — a first-party signal of what they care about and are trying to solve. Walker calls the buyer's question 'the roadmap to close.' You can't buy this data from any vendor; you earn it by giving buyers the room to ask. Captured over time, it becomes a compounding asset for messaging, product, and every downstream rep.

**Q: How should AI voice be used in the sales process?**

A: For pre-call prep, not to replace human calls. Walker is explicit that putting AI on a call a buyer expected to have with a human is a bad use of AI. The right application is a voice call moments after a buyer books, gathering a few tailoring questions so the human demo is more valuable to the buyer and the rep walks in prepared — enhancing the human moment rather than avoiding it.

**Q: How much work is it to train an AI model like Spara?**

A: Less than teams fear. Spara trains on data a company already has — website, docs, academy content, and good call transcripts — then refines the model in a testing interface where you role-play your own customer and thumbs-up/down responses to tune tone. Walker says they're 'batting 1,000': no customer has had to create a net-new asset. The work is structuring existing data and keeping a human in the loop, typically an hour or two of testing.

**Q: Does Spara replace salespeople?**

A: No — Walker frames it as 'super pro salespeople.' The AI layer delivers the real-time value humans can't (instant answers across every interface at the moment of intent, sub-minute follow-up, live routing, and enrichment) and its goal is to get more qualified buyers onto the phone with real humans faster. The human conversation, where a motivated buyer and a good seller meet, is protected, not automated away.


## Timeline

- **00:00** — Cold open: outbound is dying, inbound AI wins
- **00:45** — Spara's origin: the questions buyers never get to ask
- **02:01** — What makes Spara different: one multimodal model
- **03:18** — AI adoption and the coming tool-stack consolidation
- **04:29** — 30–70 GTM tools and the problem with AI outbound
- **06:31** — 100% of buyers visit your website
- **08:19** — Live demo: Navigator vs. Smart Bar
- **09:37** — Proprietary intent data: the question is the roadmap to close
- **11:30** — No forms, instant scheduling, real-time routing
- **13:50** — Why it feels different from a chatbot
- **15:41** — 2–4x MQL capture and autonomous email follow-up
- **19:20** — AI voice for pre-call prep (not replacing the human)
- **23:17** — Lead enrichment, CRM sync, and rep alerts
- **25:07** — Training the model: human-in-the-loop, no net-new content
- **27:34** — The AI layer of your GTM engine — pro-salespeople
- **28:15** — Wrap-up and how to reach Spara


## Related episodes

- **Ep. 92: Agents That Run Outbound While You Sleep** (Mica (Amplemarket)) — The outbound counterpart to this episode's inbound thesis — agentic outbound in production, set against Walker's 'outbound is dying' argument. · https://leanscale-knowledge-hub.netlify.app/podcast/mica-ample-market-outbound-agents/
- **Ep. 29: GTM Product Demos: Exploring Ocean.io** (Michael Heiberg) — Same live-demo format and the shared through-line of micro-targeting and real-time GTM AI over mass outreach. · https://leanscale-knowledge-hub.netlify.app/podcast/michael-heiberg-ocean-io-gtm-demo/
- **Ep. 24: AI Is Breaking Sales — Here's How to Fix It** (Mustafa Saeed) — Directly pairs with the reverse-network-effect argument — AI guardrails, deliverability, and keeping humans in the loop as AI floods outbound. · https://leanscale-knowledge-hub.netlify.app/podcast/mustafa-saeed-ai-breaking-sales/
- **Ep. 88: Why AI Won't Close Your Biggest Deals** (Michael Kiernan) — A CRO's take on where AI belongs and doesn't — echoes Walker's 'magic happens when a buyer and a good seller get together.' · https://leanscale-knowledge-hub.netlify.app/podcast/michael-kiernan-nextdoor-ai-wont-close-deals/
- **Ep. 85: Why AI + GTM Engineers Can't Replace RevOps** (Tessa Whittaker) — The 'AI augments, doesn't replace' thesis applied to the operating layer, mirroring Spara's pro-salespeople framing. · https://leanscale-knowledge-hub.netlify.app/podcast/tessa-whittaker-ai-gtm-engineers-revops/
- **Ep. 95: Why AI Means More RevOps Hires, Not Fewer** (Jimmy O'Halloran) — Extends the augment-not-replace argument — productivity gains should grow revenue teams, not gut them. · https://leanscale-knowledge-hub.netlify.app/podcast/jimmy-ohalloran-new-relic-revops-consumption-revenue/


## Full transcript

_Machine-transcribed and not diarized; speaker attribution is inferred._  
_Transcript only, as a separate file: https://leanscale-knowledge-hub.netlify.app/podcast/david-walker-spara-multimodal-inbound/transcript.md_

### 00:00 — Cold open: outbound is dying, inbound AI wins

**[0:00]** Today, we have David Walker, founder, CEO of Aspara here today. So excited to hop into your platform. I think you've taken a very unique and holistic approach to making the most of anything inbound coming to your website, coming to your brand, interacting with your company. We both talked offline before this that Outbound is completely dying, especially if you're in B2B mid-market to enterprise Saas or selling AI. And your platform is making the most of implementing AI in all of the channels that you communicate with your customers to help convert along the customer journey. David, I'm so excited you're here. Thank you

### 00:45 — Spara's origin: the questions buyers never get to ask

**[0:45]** for doing this. I think a good place to start, a lot of people think of communicating with their customers through AI via chat bots or some other, you know, typical support-like channels. How is Aspara different and what gave you the inspiration to get this started? Thanks Anthony and awesome to be here. Huge fan of the pod and honored to be a guest. So I'll give you the quick origin story of Aspara why I'm doing this in the first place. So I was the co-founder and CEO of a Saas business in the real estate space and I grew it to 350 people, sold it to a really big company with a couple thousand people, and

**[1:29]** in both orgs I spent the majority of my time and focus on go-to-market and on selling. And I always felt this pain point of, you know, we have people who are interested in us, they're coming to the website, we're doing Outbound, that's driving them back towards us to come Google us and check us out. And I always just wanted to talk to them, right? Why are you here? What are you interested in? Can we answer your questions? And I really wanted to know what questions they had for us because that would give us so much understanding of how to close them. And it's just not feasible to do with humans and AI is super well positioned

### 02:01 — What makes Spara different: one multimodal model

**[2:01]** to solve that problem that I was facing myself. So I went and started Aspara to go solve that as my life's work and I'm really, really excited about it. You asked what makes us different. I think the biggest unique positioning that we have in the market is we are multimodal, so we allow companies to train one model on their data and then we can talk to their customers in any form factors that their customers want to use. So chat, email, voice, whatever it is, Aspara can be there to talk to your customers in real time and help understand what they're looking for and then convert them to get from lead into pipeline and revenue.

**[2:40]** I think that's a huge difference that a lot of people may not appreciate. So thinking about the need of training a model for chat, training a model for voice, training a model for email, training a model for every communication mode that you have with your customer, it misses data in translation and it becomes way less efficient and doesn't give the customer the best experience to help give them the information they need. And for the most part, I think people want to interact with whatever is going to give them their answers the quickest and whatever can give them information in the most thorough way possible. So being able

### 03:18 — AI adoption and the coming tool-stack consolidation

**[3:18]** to give them the information any way they want and make sure it's accurate and tailored to them is pretty incredible. Yeah. We're in this super interesting phase of AI adoption, right? Where at big companies it's board meetings where it's talking about we have to go use AI. We have to get more efficient. It's so clear to everyone that it's a powerful tool and so we're seeing this mass proliferation of adoption of application layer AI. And what we think is going to happen is the inevitable that it's happening with more pre-LLM SaaS is there's going to be consolidation.

**[3:52]** So it's like easy to go buy an AI chat, AI email, AI voice, try to get all these things and then you realize, oh my God, the overhead and management of this stack that I've just sort of cobbled together because I wanted AI as quickly as possible doesn't make a ton of sense. And so we're building for that sort of future state that we're already seeing happen where companies are starting to think, okay, there's actually work involved here and this is a really powerful part of my org. It's almost like we become an arm of the sales team and marketing team and we want to have one comprehensive platform that can do everything that we need it to do.

### 04:29 — 30–70 GTM tools and the problem with AI outbound

**[4:29]** Something you might find interesting on that point of just like cognitive load of how many tools you have at lean scale, the average number of tools and the go to market tech stack that we manage is over 30 different tools. And we primarily serve the series A to series C market. If you go series D and above, that gets up to 70 tools on the go to market space. So it's definitely getting more complex, people putting more and more technology in there. It's getting more confusing. And then the worst part, I think a lot of people are using AI in a way that's probably diminishing their brand and not helping because they're doing

**[5:10]** spray and pray outbound, high volume, high personalization that anybody can read through. I see it all the time, my inbox. So I think the oh my gosh, don't get me started on AI outbound because we could talk about that for three hours. No, but I think the real value in AI is actually where you're starting is, okay, how do I make the most of the inbound people who want to interact with us and make it the highest quality experience possible? I actually think that's where it's going to be more powerful. And the thing is, I always say to our customers, we're an inbound platform in the sense that

**[5:51]** we're not going to send out 30,000 fake personalized outbound sales emails. That's its own world. That world's really hard. The signal noise problem is getting worse. AI makes it harder for everyone. It's kind of a reverse network effect. The more people using AI to send outbound emails, the less effective each individual email will be. So there's a whole bunch of problems there. But what we do help outbound in the sense that almost any customer that is outbound to that's going to convert is going to go to your website. They're not just going to reply to a sales email and all of a sudden go buy the product

### 06:31 — 100% of buyers visit your website

**[6:31]** having not gone to the site. And so what I like to think about is that anyone who's ever going to buy your product, particularly if you're B2B SaaS or your sort of high consideration consumer where your website is a key part of your conversion flow, 100% of people who buy are going to go to your website. Whether they're inbound, outbound, wherever they came from and what's the experience like once they get there and what's the experience like after they leave. And that is a key area that humans have done a really poor job because we're just not fast enough. We're not fast enough to talk to people when they're on the website.

**[7:08]** And talking to them allows us to capture data which allows us to then influence the next steps. How do we follow up on my email? Is there a voice call that's going to be more valuable to do as fast as possible? And how do we just increase that speed to lead and that buying experience such that our buyers actually want to buy from us? Absolutely. And I know you prepared a demo for us today. I think it'd be really great. I agree with everything you're saying. And I think it'd be great to see it in action and go through because I know the first time you showed it to me, I was pretty blown away.

**[7:41]** I was like, "This is how I would want to buy." And it feels really authentic, too. So, yeah, we'd love to get under the hood and take a look. I love that you brought up. It feels like the way you'd want to buy because whenever I'm in sales conversation, I ask CMOs, VPs of marketing, CROs, VPs of sales, say, "Hey, tell me about your current buying motion, your current selling motion and the experience your buyers go through." And they walk me through the process and I say, "Great. Would you go through that motion as a buyer?" And 90% of the time, there's kind of like a wry

### 08:19 — Live demo: Navigator vs. Smart Bar

**[8:19]** smile, right? And the answer is always the same. Honestly, no. We take nine days to show someone our product at best. And frankly, no. My attention span is too low. I operate like this and we can operate like that. And so, I love that you brought that up. It's a big focus of ours. All right. So, let me jump in and show you what it's all about. >> Perfect. >> All right. So, I'm going to demo. We have--we dog food our own product, obviously. Spara is--Spara's tech is a key driver of our own sales internally. And so, I'm going to show you Spara on the Spara website. So, first product you'll see here is called Navigator.

**[9:00]** This is bottom right chat. You're used to this UI. We've seen this before. Ours is different and it's all freeform. There's no gating. There's no if-then logic. It's all true AI-powered conversation. It's super cool. It converts at a much higher rate than your traditional static chatbot. The second product I want to show you though is what we call Smart Bar. So, this is like chat GPT for your website, right? Everyone gets this UI at this point. We know that if I go in here, I'm going to get valuable information. And so, why I love this product is because one of the pain points I always felt was we built these websites

### 09:37 — Proprietary intent data: the question is the roadmap to close

**[9:37]** in my last company, this company, to answer every question that our buyer could want to ask. But what we really care about as sellers is what is the question. If we know the question that our customer is looking to ask us, we have the roadmap to close, right? Like that gives us insight into what to do next. So, by just giving your customer the ability to ask you the question, they will show you their cards. They will show you what they care about and what they're looking to solve. And that's such valuable insights. You can't buy that data anywhere else, right? That's proprietary intent data. You have to earn it, right? You

**[10:10]** have to give them the opportunity to have that conversation. So, I can go in and ask something like, "Can SPARA help me better convert leads to pipeline?" I hope we say yes because that is definitely what we do. And so, I'm going to pop into this multimodal AI interface where I can get answers. Cool. So, like, this is showing me something I wanted to show you. So, this is great. This is showing me our journey's logic. So, okay, great. You've got pipeline. You want to turn it into a scheduled call, right? We are going to host the logic and be able to deploy AI. This is what's powering this conversation I'm having right here, to

**[10:54]** do that. So, as a buyer, I asked a question. I got a helpful answer. I got to see under the hood of the product something that's relevant to me. And now SPARA has earned the right to go and ask pre-qualification questions, right? And we've provided value. It's give to get. It's classic sales. So, we've been helpful. And so, let's say I mentioned all three. I want AI chat. I want AI email. And I want AI voice. That's going to be the best thing to help my company. I can go in and say that. And perfect. We're going to, you know, answer the question. Hey, we can definitely help you with that. We're going to ask for

### 11:30 — No forms, instant scheduling, real-time routing

**[11:30]** their email. I could ignore this. And I could keep asking product-related questions, right? I could have a long conversation to better understand SPARA. But let's say I want to just answer this question. Obviously, for the purpose of this demo, we're going to use some fake data so that we're not showing it to be real. So, Sally, you know, Big Co, VP of Marketing here gives us her email. Let's offer time. There's no reason to have delay. There's no reason to put a form up. What do I want? What's the ideal buying experience? Right? Answer my questions. Give me value immediately. Speed to lead. And then let me

**[12:09]** do the thing I care about, right? So, I can go in and schedule a call right here. Make it super easy. We integrate with whatever calendar our customers want to use so you don't have to have a bunch of change management with your team. And we can just push this conversation forward from there and kick it over to a call scheduled with one of our AEs. I can exit out of here. I can show in Navigator. We're going to have the same conversation. So whether I start a conversation in one form factor, kick it over to another, it's going to be consistent. And I'm going to be able to go in and see the media assets. They can

**[12:40]** just slide right out. It's a nice little UI trick there. And same thing with our calendar. We can just go and pop in here and schedule a call just like that. So, ultimately, we've made the experience for the buyer. We've done a couple things. One is we've allowed them to give you their proprietary intent data, right? To give you that roadmap to close of what they're interested in. Two, we've been able to have logic where we can ask different pre-qualification questions. We could route, you know, one, a highly qualified person gets an AE calendar, a mid-tier qualification person gets an SDR calendar, right? A low-tier person

**[13:19]** can be pushed into a self-service motion. We handle all that logic in real time the way a good human would, right? And feeling out a customer and being able to route them to the right place. And then most importantly, we've let the customer take the next step. So many sales motions today hold back your most qualified and interested buyers, right? They're sort of built for the average and the people who really want to give you money and turn into revenue are actually held back and they feel like they're slowed down. We allow them to accelerate and speed up and just do the right next step so that they can

### 13:50 — Why it feels different from a chatbot

**[13:50]** again turn leads into pipeline into revenue. This is done in such an elegant way. This might be a minor thing, but aesthetically seeing the bar there as far anything feels so much different than a chatbot. It feels more like I'm interacting with something that's going to give me actual helpful information, not just a sales tool or trying to lure me into a demo. Totally. I feel like just the emotion of it is completely different. Yeah, I completely agree. We wanted to have both because there's a whole group of people that are used to this bottom right UI, right? They know it, they want it, they need to see it,

**[14:31]** and it's like great. And what they're expecting is most of these chats, if you've used these on websites, you go in and you click on it and it says great. Can I have your email address? It basically is just a form and it's gating the thing that I want most and you can't go in and just sort of ask a question and start a conversation, which I always found crazy because as a seller, the number one thing I always want is to be able to understand what my customer is actually looking for, what their pain points are. And so the big unlock here was just don't gate it, right? Let's provide value. Let's allow the person to ask questions.

**[15:05]** Let's be helpful. Let's talk about your product. Let's show information about your product. We will weave in pre-qualification questions to capture that data. And as it turns out, if you're willing to have a free form, high value conversation for the buyer, they are so much more likely to give you the information you're looking for than if you just do the form, the sort of the dumb version, right? Which is, hey, here's basically our form. We need your email address for you to begin. And that's the kind of, that's the old world of buying experience. The new world is dynamic. It's fast. It focuses on speed to lead. It

### 15:41 — 2–4x MQL capture and autonomous email follow-up

**[15:41]** focuses on creating value. And guess what? The results are, I mean, we're seeing literally 2, 3, 4X greater engagement and MQL capture than pre-existing chat. And then you layer in the net new interface, like a smart bar, and you really have the best of both worlds. That's huge. That's huge. So this is interacting with the website, interacting with chat. What other modes of communication can Swara get in front of your customer with? All right. So what's really cool is, you know, since I didn't actually schedule that call, I can now go in and jump into email to email and fully autonomous. So I'm going to go jump

**[16:20]** into this conversation and we can see that, you know, Sally here was interested in better lead conversion in a pipeline, interested in AI, chat, email, and voice. Awesome. You know, offered call scheduling. She didn't schedule. Let's go compose an AI-generated email. And so we can now have a hyper-personalized one-to-one follow-up with Sally. It's going to mention, we talked about lead conversion to pipeline, right? So it's literally going to speak to the conversation that she just had. So it feels like a continuation. Ask if there's any more questions and, you know, say that, hey, we're here to support our calls.

**[16:58]** So we can go off and, you know, this is Sally at Big Co, so I'm not actually sending Sally fake email, but we can go send this email. And then anytime Sally replies back, we can reply within 10 seconds. So just, just like we're built for real time with chat and with voice, we're built for real time with email. And so I always say the only reason email is a high-latency form of communication is because humans are really slow. But if we remove that limiting factor, then it can be really, really fast. So anytime Sally happens to give us her interest and reply to one of her emails, we can reply back instantaneously.

**[17:34]** Just a better experience for the buyer, right? I don't want to send an email about your product, go off and, you know, forget about it for six hours and then get an email back in my inbox. I'm doing something else at that point. I want it to be fast. So fully autonomous email, super powerful. And having it be synced with the chat, and this is, you know, to your point that we brought up earlier, there's a lot of point solutions, but when you can consolidate into one platform, you get the benefit of that seamless transition from one modality to another. Well, and I think I've a lot of times been in the exact position

**[18:07]** Sally has been in, where I'm on a website, I'm checking out the product, I go click on the link to book a demo, I'm sitting there looking for times, then I get a slack, then I'm out of there. And then I never end up actually booking. And I think that's a pretty unique and particular nurture population that you're really well positioned to help take care of. And I think you can really increase conversions because I likely, if somebody were to reach back out to me, I was like, Oh, yeah, I actually just got busy. Yep. Thanks for reminding me and bringing all the context, the questions I was asking. Yeah, I'll book

**[18:44]** a demo now. Yep, totally. And, and, you know, what's what's nice is AI is also not self conscious that it's them. I always remember early in my sales career, right, someone wouldn't respond, be like, Oh, God, did they hate me? Did they hate our product? And then, and then you start to realize, no, they just have lives, and they're busy. And they just didn't happen to do the thing in that moment. And so giving, you know, Sally every mode of communication to have the option to convert, but all synchronized to yield one positive outcome makes a ton of sense. Amazing. So let's say I did schedule a call, right? So I've got a call three days

### 19:20 — AI voice for pre-call prep (not replacing the human)

**[19:20]** from now set up with a human. If I show up to that call as the buyer, and it's your AI on the call, I'm going to be pissed, right? That's a bad use of AI. I'm expecting that human touch. But when we think about AI voice, we think about what are the moments where we can actually improve the experience for the buyer, right? Not just try to replace a human call for the sake of automation, which is not a good goal, because I think magic happens when a motivated buyer and a really good seller get together to talk about the same problem at the same time. But what if I schedule that call? And, you know, five

**[19:58]** seconds later, my phone rings, and it says, Hey, this is Spara, right? And it's asking me some questions so that my scheduled call can be more tailored and more valuable to me as the buyer, right? That's an opportunity for AI to use voice to add value to the sales team, right? They're going to know more going into the call, be able to better suit the demo to the buyer's needs. But most importantly, it adds value to the buyer, right? They know they're going to get maximum value from that call. So, you know, I can go in here and for this, we're going to use a browser call, right? Just for the sake of this demo. So I'll start

**[20:31]** the call and remember, this is pre-call prep. I just scheduled and now Spara AI is going to call me to make sure that call is as valuable to me as possible. Hello. Hi there. This is Rob and AI assistant from Spara. Thanks for scheduling your demo with us. Is it all right if I ask you a few questions so we can prepare for your call? Sure. Great. I have a few questions for you. Firstly, how many employees does your company have? We have 100. Got it. 100 employees. Thanks for that. Next question. We need to convert more of our leads into pipeline and ultimately revenue. Understood. You're looking to convert

**[21:20]** more leads into pipeline and ultimately revenue. That definitely aligns with Spara's strengths. Next question. Are there any other companies you're evaluating alongside Spara? No, just Spara. Thanks for confirming. It's just Spara you're evaluating. Next question. What time frame are you considering for implementing the solution? Not really sure, but definitely I want it set up in the next three months. Got it. You're aiming to have something set up in the next three months. Last question. What are you most excited about learning during the demo with Spara? I want to learn all about Spara's AI voice capabilities. I heard they're

**[22:00]** really cool. Thanks. I'll make sure the demo focuses on Spara's AI voice capabilities since you're excited about that. Before we wrap up, do you have any questions or concerns you'd like me to address right now? No, this was super helpful. Thanks so much. So hopefully that gives you a sense of it feels to me. Oh, cool. You want to know what I care about. And as the seller you're showing, I'm like, oh my God, I know exactly what my customer is looking for. I'm set up to close. Well, and I think something too, maybe I'm just getting older and this is like what older people do, but I have it starts to get exhausting

**[22:38]** when I have to type out every single thing that I'm looking for. So if I got like a list of those questions and I had to type out everything for someone or I wouldn't want to burn time on a discovery call with someone too. So I think having the opportunity to just communicate it via voice, I can give a long winded answer about something and you can kind of make sense of it later. Yeah, it is a better experience than needing to fill out another form or something. Totally, totally. Last thing I'll quickly show you. And this is fake, fake info, obviously, because I don't want to put anyone's real information up here. But on every lead, we're

### 23:17 — Lead enrichment, CRM sync, and rep alerts

**[23:17]** also doing a bunch of background research. So we're going in and we're looking at what's this person's role, where do they work, doing multiple lead enrichment waterfalls. We're providing a conversation summary of the conversation spars had to date. This would be, you know, via all different channels and company info, you know, competitors, recent news for different talking points, basically the playbook for the sales rep, right, to come super prepared to be able to close every deal. And so whenever we're having an interaction with someone, we can not only will be pushing all the data into your CRM. So what was talked about, you

**[23:57]** know, where we converted them to the next steps, etc, all the lead and company level data. But we can also proactively send slack alerts to sales reps. Once they get assigned within the CRM, we see who is assigned to Sally, we can email them. So they have all the notes from the history of the conversation as well as all this super helpful research. So we look at our role is, you know, we're going to create a better buying experience by allowing that buyer to engage as fast as they want in real time in whatever modality they want with the best possible AI that knows how to sell your company, we're going to take

**[24:31]** all that data, we're going to structure it, we're going to save it all in your CRM in a really useful way. And then we're going to proactively push that data to the right sales rep at the right time so that they're set up to close. It's super comprehensive and love that you're pulling some external research and enrichment on that lead and contact as well. I'm curious, I think this is probably where some of the biggest aspect of maybe implementing Spark would come from. But how do you train the model and how do you recommend your customers train the model because any model is usually

### 25:07 — Training the model: human-in-the-loop, no net-new content

**[25:07]** only as good as the data and how structured the data is. So how do you recommend people get set up with Spark? That's a great question. Every customer asks this and there's always this fear that it's going to take so much work. The reality is it takes some work because we really believe in having a human in the loop for this training. But we start by basically saying, okay, let's look at all of your public information that you host. So your website, your docs, if you have any academy, any sort of internal knowledge base that is sales related, that is okay to be that knowledge to be public facing. And we can take that data in whatever format it's

**[25:47]** structured in. We're really good at chunking that up, making it super useful and training a model. Then you go in and we have a whole testing interface within our platform to simulate the real conversation. You get to pretend to be your own customer and you get to simulate real conversations and the real interface and UI that you would see on your website. And you can provide feedback. Thumbs up, thumbs down on every message. And through that the model evolves and learns and gets smarter. And so there's a couple other ways, right? We'll take transcripts of calls that you have that were really good, any kind of internal

**[26:23]** sales rep training data. What I will say is we're batting 1,000 where no company has had to create a net new asset piece of material or piece of content for us. It's not so much just getting access to the right data where we're able to get the data that companies have off the shelf, but making it super useful to actually train a model we got really good at. And then having that extra magic of the human in the loop to really sort of refine the tone and the feel has been awesome. And our customers are surprised by, you know, they spend an hour or two testing and giving feedback and they can literally see the improvements

**[26:57]** and the changes. And they're always surprised by how easy it was, but how well trained the model gets to their specific needs. That's incredible. And as you continue to get more and more information and it gets to know your customers better, I'm sure the experience just gets better and better. And you're really getting smarter over time. Yeah. And you're really building foundationally across all your channels. So every channel in which you communicate with your customers or prospects, get the benefit of that model being trained. Yep. Our goal is to just be your, as I kind of alluded to earlier, it's

### 27:34 — The AI layer of your GTM engine — pro-salespeople

**[27:34]** to be the AI layer of your sales and marketing go to market engine, right? Where we are not trying to get rid of salespeople. We're actually super pro salespeople. Our whole goal is to basically get more customers on the phone with salespeople. But to provide all those points of value to buyers that lead to conversion that humans can't do, right? Real time conversation across any interface in the moments that buyers want to talk to your company and humans are too slow to be able to talk to those buyers so that we can get those people moving forward in the funnel and connected to a real human. Well, David, I am so impressed by the platform,

### 28:15 — Wrap-up and how to reach Spara

**[28:15]** so impressed by the intentionality of every aspect of the workflow that you presented today. And I think this is a powerful solution for anyone of LeanScale's customers, anybody who's watching this now. I think the smart bar that you have in the front, bringing information to potential customers as early as possible, genuinely, authentically helping them solve their problems, gathering that unique and proprietary intent data and then nudging them and nurturing them along the way. You do it in a way that I think is better than any other way I've seen and I'm just really impressed. And I appreciate you sharing everything today.

**[28:57]** Thank you so much, Anthony. Austin to be on. I love this podcast and it's been really, really fun to sit down and chat with you and show you Spara and a lot more fun stuff to come. Well, last thing before we wrap up, best way to get in touch with Spara, I'm guessing by interacting with Spara on your website. And any best way to get in touch with you as well. Yeah, pop on our website, Spara.co. Spara is right there, ready to talk to you. You can do all the same things that I just did here. And my personal email, David@Spara.co. Feel free to shoot me an app. Awesome. Well, thanks so much for being here, David, and

**[29:36]** I can't wait to see what you guys build next. Thanks, Anthony. Good to chat with you. Take care.


---

_LeanScale Podcast Knowledge Hub. Free to quote and cite with attribution to The LeanScale Podcast (https://www.leanscale.team)._