#The challenge
Marketing had run on the same automation platform for years. Everything worked, and nothing could change: the stack blocked the personalization, predictive scoring and AI-assisted workflows the company wanted next, and years of accumulated programs, custom objects, webhooks and integrations meant nobody could say what would break if it moved. The migration also could not stall campaigns while it happened — demand generation had to keep running through the whole thing.
#The approach
Audit everything, with an explicit keep-or-kill call
Inventoried activities, objects, assets, integrations and webhooks with a keep-or-kill recommendation on each, so the migration carried forward what was actually used and retired the rest — rather than faithfully replicating years of accumulated debt in a new platform.
Sandbox before production
Built a sandbox testing plan so the migration was validated somewhere safe before anything touched live sending.
Protect deliverability across the move
Planned DKIM setup and IP warming ahead of cutover, so sending reputation carried across the platform change instead of resetting at go-live — the failure mode that quietly costs a migration its first quarter.
Preserve behavioural data
Set up custom events so behavioural data survived the move and scoring did not have to start from zero on the other side.
Load brand assets so AI output is on-brand from day one
Loaded brand assets — logos, colours and fonts — into the destination platform so its AI generation features produced on-brand output immediately rather than after a cleanup pass.
Run marketing operations in parallel
Ran embedded day-to-day marketing ops throughout the migration: weekly pipeline and MQL reporting, an automated daily pipeline report, lead-hygiene and data-quality dashboards, event lead upload and attribution, and campaign and newsletter email builds.
Strategic sessions on the AI operating model
Monthly working sessions with the head of revenue operations on the AI operating architecture — governance, operationalizing the ICP, and an evaluation of AI agent and customer-data products from platform vendors already in the stack.
#Outcomes
Migration scoped with audits delivered
The full platform audit was delivered with keep-or-kill recommendations, and the migration was underway toward a target cutover roughly a quarter out. The cutover is a future milestone, not an executed result.
Marketing kept running through the migration
Weekly pipeline and MQL reporting plus an automated daily pipeline report were stood up and run in parallel with the migration work.
Lead-hygiene and data-quality dashboards live
Hygiene and data-quality dashboards were built so database problems surfaced during the migration rather than after it.
Deliverability and behavioural data planned for, not discovered
DKIM setup, IP warming and custom-event configuration were planned ahead of cutover rather than handled as post-launch surprises.
Brand assets loaded for AI generation
Logos, colours and fonts were loaded into the destination platform so its AI features generate on-brand output from day one.