Outbound & Sales Development

Agentic GTM is composed from a few interchangeable building blocks: a very smart intern (the LLM), web search, a data platform, and a workflows agent that can take or suggest actions on your behalf.

A maturity ladder for AI in GTM: using skills (precise, installable instruction sets) puts you in the top 1% of users; using scheduled tasks (routines that run agents on a cadence) puts you in the top 0.1%.

A research-and-enrichment agent (running in the LLM, plugged into the web) hands its output to an engagement agent living inside the GTM platform (harnessed to owned channels and CRM), which has its own instructions for how to act.

Personalization that generates a bespoke sequence per person: the agent knows whether you're connected, whether you emailed or cold-called before, and what the CRM says, and decides both the copy of each step and the channel each step starts on.

To design agentic flows, ask what recurring work you would hand a very smart intern you just hired — then build agents around that answer, letting the LLM interview you and write the instructions.

Across Amplemarket customers since January 2026, opportunities generated by channel split roughly 40% email, 40% calling, and 20% social — evidence that multi-channel, contextual outbound still works.

AI creates a widening productivity gap between adopters and non-adopters, but market pressure keeps teams employed: if you can go faster, competitors force you to reinvest that speed rather than fire people to move at the old pace.

A framework for evaluating a 'backwards' move from CEO to CRO: weigh product-market fit, founding-team fit, and investment thesis against the compensation, equity, and personal-fulfillment math of joining a high-growth build with strong culture.

The idea that the go-to-market motion — talent profile, process weight, and metrics — must be rebuilt at each ARR band rather than scaled linearly. Different stages leverage different areas of the process.

Gate sales hiring on two leading indicators: how well reps are ramping against a defined ramp curve, and what percent of quota (and ramped-quota capacity) they are attaining. Below threshold, pull the plan back.

The second-order damage of over-hiring sales: diluting territories and top-of-funnel demand across reps who won't stick, which starves top performers and eventually drives your A-players out.

A hiring heuristic for specialized verticals: emulate your customer base and screen for the common denominator of work ethic and mission alignment rather than a specific sales or industry pedigree.

An AI-adoption operating model: allow broad, decentralized experimentation to reduce fear and prove ease of use, then centralize the valuable skills, agents, and data pipelines — governed by RevOps — for anything mission-critical.

A pipeline philosophy that favors thoughtful, researched, use-case-specific outreach to a narrow buyer over high-volume, low-hit-rate blasting.

Treating in-person events as an operational play with three phases — pre-plan (targeting, pre-set meetings), execute (on-site, ideally with stage presence), and post-plan (structured follow-up tracked through CRM) — not as a booth you show up to.

The primary barrier to AI adoption is a mental model, not a skill gap: people onboarded in a pre-AI world treat AI as the next, harder evolution of technology and cling to point-and-click UI notions.

A model of selling as two blended disciplines: the art (psychology and influence — moving many stakeholders in the same direction) and the science (methodically progressing a deal through a rigorous process to signature).

The reframe that managing a sale is like managing a project — bringing operational and project-management rigor (sequencing, stakeholders, milestones) to progressing a deal to close.

Sort work by urgency and importance: do the highly-important-and-urgent first, delegate the urgent-but-low-importance, and protect time for the highly-important-but-not-urgent — always weighing level of effort per initiative.

In an AI world, the cost of being wrong has collapsed, so iteration velocity — not first-time correctness — is the dominant competitive advantage in go-to-market.

Do the reps yourself with an AI-built tool and an expert on call, so you compress the failure that an outside agency would spread over months into a couple of fast, cheap weeks.

Any task that looks like one step ('start Google Ads') is really dozens of steps, and not knowing where to start — or which step is next — is what actually stops people, not the tool.

Reverse-engineer every revenue goal from the sales cycle and pipeline coverage, then act with the urgency that math demands.

For AI-native companies at scale, the customization math tips from buy to build once one engineer can build the tool in a month and any coding agent can maintain it.

The durable value in a GTM tool isn't the code — it's the product manager's hard-won knowledge of what people actually do with it: the workflows, sequence, guardrails, and integrations.

Get a paying customer on an MVP or front-end-only demo before building the real product, so you validate that someone will actually pull out a credit card.

AI gets you ~99% of the way, but the final mile still takes human taste, focus, and follow-through — a six-month project becomes four days, not four minutes.

You can only automate a motion well if you understand it cold — which is why AI SDRs built by people who never ran a sales process miss the mark.

The quality of any tool reflects how core its job is to the business that built it — so for jobs that matter, choose the obsessed point solution over the all-in-one chasing more NRR.

Match every hire, tool, and system to your actual stage instead of copying what much larger companies do.

The best GTM advice is focus — narrow the ICP and the product, and get comfortable saying no early.

Since go-to-market process breaks whenever it depends on human compliance, the fix isn't more enforcement (mandatory fields, stage gates) but removing people from the data-capture loop entirely — letting AI listen and populate the system automatically.

Once AI captures CRM data automatically, the salesperson stops being a producer of data (data entry) and becomes a consumer of it — served a prioritized view of what's healthy, what's slipping, and what to work on next.

Adopt in stages: crawl (RevOps connects CRM, call recorder, Slack/Teams, and email, sets team structure and field mappings for a two-way sync), walk (auto-create and enrich records, remove humans from data entry), run (deal-health analysis, agents, and cross-functional data products).

Plot every opportunity on two axes: horizontal = how healthy the deal is (likelihood to win), vertical = how likely it is to close when the rep expects. The quadrants surface safe bets, acceleration opportunities (will close but not this quarter), and firm-decision-date deals where you may not be selected.

A full-reasoning agent wired to every captured touchpoint plus external research tools that executes deal tasks — building a custom ROI calculator to the prospect's own metrics, pulling industry benchmarks, and drafting the decision-maker email.

A per-contact score from -10 to +10 that identifies champions and blockers at a glance, with the reasons and specific quotes behind each. Filtering contacts by ICP persona and a high promoter score produces a live, shareable list of advocates.

Product roadmap signal should come from live sales conversations with the market — use cases, friction, competitors mentioned — not primarily from customer success and existing customers, who are biased because their problem already feels solved.

RevOps problems are hard to solve but remarkably consistent across similar-stage companies — a Series B sales-led company has the same problems and the same fixes as its peers — which is why the function outsources well while sales and product must stay in-house.

As AI and no-code make building products easy, the hard problem shifts from creation to distribution — getting a great product in front of its rightful customers in an attention (eyeball) economy.

An outbound platform architected for AI from the ground up as a multi-agent system — each agent using the model it's best at — rather than a pre-AI product with AI 'slapped on top' via chatbots or plugins.

A workflow where you paste a domain, the system scrapes it, infers your ICP and buyer personas, writes them out as reusable context files, and converts them into a targeted lead list that also powers copywriting and qualification.

Use intent signals (job changes, hiring, department growth, 10-K priorities, life events) to decide who to contact and when — but keep them out of the message. Mentioning the signal wastes scarce email real estate and doesn't impress the buyer.

Replace the standard chain — Sales Navigator for lists, Apollo and other enrichment tools, a verifier, and ChatGPT deep research — with a single AI-native system on a fair, usage-scaled credit model.

Bridge the gap between sellers who understand angles but not tooling and 'GTMEs' who understand tooling but not selling by giving one strategy-fluent operator an easy-but-sophisticated execution system.

Model your own mind as a large language model: a database with a learning layer whose training data is the people you interact with. The more people you meet, the richer the dataset and the better your predictions, analysis, and instincts.

Because your brain-as-LLM has no safety team to filter bad inputs, deliberately curate what you consume: stay away from the noises of no's and stay close to the noises of yes, filling your dataset with people who believe things are possible.

Any capacity you stop using atrophies — a finger immobilized for months, eyes covered for a year, and the brain the same way. Offloading creative and critical thinking to AI quietly weakens the very muscles you depend on.

The outreach-spam era is a supply glut: pre-2019, ~50 companies chased ~100 US/Europe customers; the tech boom pulled ~500 sellers from across the globe toward a shrinking buyer base, so desperate automated outreach was inevitable.

There are only two ways to succeed: consistent, disciplined long-term work, or compromising ethically to appear successful overnight. Growth hacks are nonsense; you have to move through the natural process with patience.

Modern life moves people between three boxes — office, home, and club — which caps thinking and energy. Humans were designed for the outdoors, so escaping into nature once or twice a month restores perspective, energy, and ideas.

A VC framework Anthony relays: evaluate a startup on the wave (market), the surfboard (product), and the surfer (founder) — and the surfer matters most, because a great surfer finds the right wave even on mediocre equipment.

Treat your network as a decade-long asset instead of a job-hopping byproduct. Build history with people and do it with integrity, and those relationships become your most durable, lowest-cost pipeline.

A founder raises capital only three to five times in a lifetime while an investor does it every single day — so the founder is structurally the amateur. Closing that gap with structure, data, and network intelligence is the mission.

Flowlie's positioning: a behind-the-scenes operating system for a raise — not a marketplace, broker, or middleman — that helps founders uncover the right investors and the right people in their own network to reach them.

The two pillars of Flowlie: a predictive fit-scoring model (version five) that ranks how likely a firm or partner is to be interested, and a network-analysis engine that maps warm-intro paths and ranks each with a 'path impact score.'

The core fundraising philosophy: the outcome is decided mostly by the preparation — target lists, investor updates, relationship-building, and warm-path lining-up — that happens before you ever say you're raising.

Deliberately forward-loading warm-intro requests — scheduling connectors to introduce you weeks out — so investor meetings cluster into a single window instead of trickling in one at a time.

Valley's model for a full outbound loop: (1) find high-intent leads, (2) research them across 60–70 data points, (3) score them for ICP fit, (4) craft messaging in your voice, and (5) send it on LinkedIn autonomously at scale.

V1 outbound: take a filter-based list, enrich the contacts, drop them into an email or LinkedIn sequencer, and wait for meetings at the bottom of the funnel. V2 outbound: intent signals, warm outbound, and enrichment/research/qualification — send 30 hyper-relevant messages and book 10 instead of blasting 30,000.

A converting outbound message must make the recipient feel two things: relevance (it's clear why you reached out to me specifically, not the person next door) and investment (that real time was spent on me, not a one-to-many blast).

Person-level website-visitor identification is only valuable when paired with a layer that researches, personalizes, and sends. Intent should be weighted by visit frequency (repeat visitors over single visitors) and by high-intent pages like pricing, customer, and case-study pages.

Run Valley on autopilot (fully autonomous research, drafting, and sending) for broad, high-volume ICPs and transactional sales; run it as a copilot / human-in-the-loop 'AI SDR intern' (research and drafting, human approves send) for narrow, high-value, account-based motions.

Because LinkedIn caps monthly message volume, the only levers to improve outbound results are acceptance rate and reply rate — which forces teams to improve the relevance of their prospects and their messaging rather than sending more.

A GTM platform should be designed from the ground up as one system spanning data, engagement, and machine learning — not assembled by bolting point solutions together — the way a self-driving car is engineered whole rather than by strapping cameras and radar onto an ordinary vehicle.

AI augments the seller rather than replacing them. Duo, launched September 2024, is a human-in-the-loop companion — the rep's Pokemon or Iron Man suit — that learns each individual through reinforcement learning and grows with them.

Amplemarket is 'in the business of matchmaking' — connecting buyers who have problems with sellers who have solutions, so that every time a problem exists the buyer is made aware of the best possible solution.

In non-transactional, high-consideration buying, the purchasing experience — the craft, the care, the reverence for the product — is part of the value itself, and that care transfers to the buyer.

AI lets far more people build, so there will be more companies ('planets') to connect, each with smaller sales teams, and the space between them grows more opaque as creating information drops to near $0.

Every 24 hours the rep lands on a fresh feed of the most relevant accounts and buying signals in their book of business (the Spotify Daylist), paired with a recommended action for each — swipe the lead in or out (the Tinder system of action).

Low-quality, high-volume outbound is not a small positive but an active negative — it burns your domain, your leads, and your single chance at a first impression, signaling that your company doesn't care.

The highest-value trigger is timing — reaching a buyer when the problem you solve is already the last thing on their mind before sleep. You find that moment by composing signals (e.g., 100%+ team growth plus ten open AE roles) rather than relying on any single one.

Attio's product philosophy: the CRM should adapt to how your organization already does business, not force you to change your process to fit the tool.

By syncing your inbox and calendar on signup, Attio auto-builds your network of companies and people, enriches it, and layers on last-touch, contact ownership, and relationship strength — with no separate tool.

A custom record attribute powered by an AI prompt: you write your ICP in plain language and Attio evaluates every inbound lead against it, flagging fit for the rep.

Five standard objects plus unlimited custom objects and attributes, including Workspaces and Users objects that pull product data in, so the schema mirrors your actual business.

An automated workflow that, on every new signup, uses a research agent to summarize and ICP-tag the company, then routes: enterprise to round-robin, non-ICP to self-serve, and ambiguous mid-market/startup leads to a Slack channel for a human to route via buttons.

Attio is both where customer data lands and where you take action on it — you can report on live data, drill into the underlying records, and immediately sequence, task, list, or route them without leaving the tool.

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.

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.

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.

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.

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.

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.

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.

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.

Measure the business by revenue per full-time employee rather than by headcount hired or money raised. Top performers run $500K+ per head (versus an old $150–200K benchmark), driven by AI-leveraged operators.

Build marketing, brand, a reliable pipeline channel, and your own sales process before hiring a salesperson. Reps are harvesters of pipeline and closers — not creators of demand.

Land your first sales inside your existing network, then narrow to a hyper-specific micro-niche for whom the product is an absolute no-brainer, and make the economics the best deal of their lives early on.

In an AI-driven sea of sameness, brand generates demand. Aesthetics signal seriousness and a content strategy (written, tutorials, or podcasts) is the modern equivalent of commercials and billboards.

Start on HubSpot as an affordable, pre-built, scalable CRM; stand up Snowflake as the data warehouse for sales, product, and financial data; and report from there (e.g., Looker) rather than overloading the CRM.

Before buying any onboarding or CSP tooling, define exactly what first-time-to-value is for your product and sprint to reach it as fast as possible.

Let AI take work to roughly 90% and reserve the last mile for a human, so output sounds authentic and nothing goes out that doesn't resonate. The goal is producing better, not just producing more.

Treat go-to-market like health and fitness: track leading-indicator 'biomarkers' (onboarding speed, churn by segment, new-rep ramp, pipeline created, conversion) instead of reacting to lagging results after they break.

Each stage of growth — validation, product-market fit, product-channel fit, scale — is exponentially harder than the last, and you can lose product-market fit at every technology wave (on-prem to cloud, cloud to SaaS, SaaS to AI-native).

Design your offering as the thing you personally wished existed in your prior role, then scale the 'love' by hiring people better than yourself, guarding culture and integrity, and getting process and finances tight early.

RevOps is the business's family-clinic generalist — no single specialty, but a stream of problems from every function daily. Its job is to diagnose root causes by stepping into each function's shoes, not to treat the presenting symptom.

Take a reported symptom and break it into workflows and steps from first principles — for a conversion drop: lead source, count, region/quality, marketing activity, routing, scoring, and product pitch — then benchmark whether it's isolated (~20% of reps) or across the board.

Step 1: give the person comfort and let them talk (avoid seeding your bias). Step 2: validate the hypothesis quietly against the data in the background. Step 3: talk to other stakeholders of the platform, process, and functions to triangulate where the problem truly lies.

RevOps solutioning is a blend of people, process, and platform — never numbers alone. The revenue outcome can come through personal relationships, process, or systems, and usually a combination.

Four defenses that stop problems before they surface: (1) automation and AI to keep leaders out of low-value work, (2) learning and development so the team understands how the GTM machine fits together, (3) data hygiene with restrictive write-access to core systems, and (4) weekly/biweekly checks with real-time reports and fix-on-the-spot remediation.

A deliberate, agenda-less block of time spent exploring the data — the opportunity module, lead behavior, Slack signal — just to sense how the business is behaving, without a specific question to answer.

The two skills that carry a RevOps career: being a genuine people person who can build relationships with extroverted sellers and senior cross-functional leaders, and curiosity paired with a doer attitude — because the problems are new every day.

Prioritize RevOps work by identifying the few major 'pillars' or 'boulders' that create the biggest business impact for a given week, month, and quarter, and aligning them to the company roadmap and OKRs.

Most refusals aren't a hard no but a 'not yet or not now' — the request is acknowledged, logged into OKRs and weekly planning, and sequenced behind what the revenue-generating teams need right now.

An informal personal scale that rates each task by how easy or hard it is for you specifically to address, used alongside deadlines to decide what to work on and when.

Do the hardest, biggest task ('the frog') earliest in the day, so the rest of the day is easier to navigate.

The scoreboard isn't tasks completed but tasks completed that have purpose — work tied to a real company or RevOps priority.

Design departments and systems that run self-sufficiently without you — reducing your role to maintenance — so the business survives your absence.

Stand up a good-or-great process quickly and iterate on it, rather than trying to architect a perfect one up front.

Effective RevOps leadership requires all three at once: having strong opinions, voicing them, and having good opinions backed by data and field experience.

Direct disagreement at the plan and the best direction for the organization, never at the individual — because you're all on the same team.

Treat initiatives as calculated risks with an explicit hypothesis, a plan B/C, and a shared understanding of the odds — so a failed experiment that proves something still counts as a win.

A research-derived methodology that distinguishes two selling environments — low-complexity/transactional and high-complexity/enterprise — and, in the complex environment, wins by Teaching the buyer something new, Tailoring it to their situation, and Taking Control from an advisory, expert frame.

The Challenger research profiled sales-rep archetypes and measured which won. Weiss names four: the Hard Worker (outworks everyone), the Challenger (teaches and pushes change), the Relationship Builder, and the Lone Wolf.

Neil Rackham's 1970s discovery framework: understand the buyer's Situation, the Problems inside it, the Implications of not solving them (cost of inaction), and the Need-payoff of solving them, to drive urgency around why change now.

A riff on SPIN designed to combat product-led, feature-and-benefit selling by first understanding the buyer's problems before offering a solution — best used to help a buyer see a problem they didn't know they had.

The discipline of reading where a buyer is in their own process and matching your motion to it: a mature buyer knows the problem, the solution criteria, and the competitors; an immature buyer knows little and needs co-creation.

A holistic system organized around engaging the buyer, qualifying, and closing — including 'upfront contracts' (pre-agreeing to a next step) and 'pain funnels' that probe first-, second-, and third-level pain.

The view that sales mastery is the deep execution of a full set of fundamentals — hunting, reaching decision-makers, discovery, creating needs, business-case building, presenting, multi-threading, negotiating, handling rejection — assembled in your own way, rather than any single methodology.

Gartner's framework, rooted in research on buyer decision fatigue and decision confidence, in which the seller's job is to help an overwhelmed buyer make sense of an overload of information and competing options so they can decide with confidence.

The time it takes a new user to 'get it' after logging in — a core PLG success metric Ocean actively drives down by putting the product's aha moment directly on the landing page.

Vectorize both companies (65M) and LinkedIn profiles (230M), then combine them in one search: input an example person's LinkedIn handle and find lookalike people, by role and context, inside the lookalike companies of a target account.

Target by a contextual understanding of what an individual actually does for a company, not by their title — because titles vary with company size (CMO vs. head of growth vs. VP marketing) for the same real role.

Build, filter, and preview the target list in Ocean without spending a single credit; export to Clay for enrichment only once the list is validated.

Gen 1 is an LLM wrapper around an analog/non-normalized database — a pretty face on messy data with broad, imperfect targeting. Gen 2 models the actual GTM process and automates the flow end-to-end, with human validation between steps.

Automation's real strength is micro-targeting: overlay intent and third-party data on a tightly defined audience so every message is highly relevant, producing 5–10% conversion instead of 0.1%.

Instead of sending from a massive shared server (a shared IP pool) full of thousands of unvetted senders, give each user an isolated mini-server ('cluster') with its own IP address so one sender's behavior can't affect the others.

Regularly send emails from a customer's mailboxes to known reference mailboxes to observe where they actually land — inbox, spam, promotions, or undelivered — as the true indicator of email infrastructure health.

Simulate natural, two-way activity across real corporate mailboxes to balance the unnaturally low response rates of cold outreach, so email service providers don't flag the account.

Cap daily send volume per mailbox to what platforms now tolerate (15–25/day, down from hundreds), have the platform control the cap rather than the rep, and scale volume only after messaging is validated on a small sample.

Treat cold outreach like a paid-ad platform: give the system many message variations, test each against a small subset of the audience, and scale only the versions that generate positive engagement.

Use reinforcement learning — a distinct branch of AI from LLMs — as the optimization layer that looks at what has and hasn't performed to predict which hooks, lead magnets, and offers will resonate, and recommends new variations over time.

Blend autonomous AI (which ingests large data sources) with human review checkpoints — a sales rep reviews certain AI-generated copy before it reaches a prospect, and an admin reviews certain content before it reaches the rep.

The main lenses for cutting a sales team's territories: (1) geographic — international/domestic regions, time zones, states; (2) product or service specialization; (3) industry/vertical; (4) firmographic tier — enterprise / mid-market / SMB; and (5) a fair round-robin or named-accounts approach when the others don't apply.

Balancing territories is not only a morale-and-attrition safeguard; it's a pure-business lever. Equalize a strong territory and a weak one and, in aggregate, the same sales headcount produces more revenue.

Design territories from evidence: mine historical SQLs and closed-won deals sliced by each segmentation bucket, backfill any data you failed to capture, and gather feedback from reps, product, and marketing before drawing the lines.

Roll new territories out at natural calendar breaks (a new quarter or month) and define explicit holdover criteria governing which prospects a rep can keep working after the reshuffle.

Anthony's go-to sequence for a typical B2B SaaS company: start firmographic (enterprise vs. SMB motions need different sellers), then geographic by time zone (buyer availability), then product or industry only if they genuinely differ, and fill the rest with round-robin or named accounts inside bigger buckets.