Most teams trying to move from rules-based to agentic CRM start in the wrong place. They pick the journey that’s easiest to build, or the one a platform demo made look impressive, rather than the one actually worth the investment. Before you touch a single line of automation, run this audit. Five steps, an afternoon’s work, and you’ll walk away with a shortlist worth building.
Step 1: List every journey currently running on rules
Every welcome flow, win-back trigger, birthday message, tier promotion and re-engagement send you’ve got, however manual or automated it already is. Include the ones nobody’s touched since the platform went live. You can’t pick a good candidate from a list you haven’t actually written down.
Step 2: Score each by commercial impact
For every journey on that list, ask the same question: if this moved from rules to agentic, would it actually move a revenue or retention number, or would it just run the same campaign slightly faster? If you can’t answer that, it doesn’t make the shortlist. Seasonal sends and one-off promotions rarely do. Lifecycle journeys that touch a customer repeatedly over time usually do.
Step 3: Score each by data readiness
Check the signals behind each journey. Are the identifiers clean across POS, app, and loyalty, or is the customer scattered across three partial records? Is the event data flowing in real time, or does it land a day late? A high-impact journey running on stale, fragmented data isn’t ready. It’ll just make confident, wrong decisions faster than the rules-based version did.
Step 4: Check how safely you can test it
Can you run the agentic version alongside the existing rules-based journey as a control, and compare the two properly, or does this require betting the whole programme on day one? The journeys you can de-risk incrementally are the ones worth starting with. The ones that only work as an all-or-nothing switch go on the list for later.
Step 5: Pick the overlap, not the shiny one
Your best first candidate isn’t the journey with the highest score on any single dimension. It’s the one that scores reasonably well on commercial impact, data readiness, and testability all at once. That overlap is small on purpose, and it’s why most teams skip it, revert to magpie mode and pick the shiny one instead.
Run this audit properly and the answer is rarely the journey you expected at the start. We’ve built the full method behind this, the five steps and how they compound over time, into the Agentic CRM Playbook.
