The Agentic Renewals Maturity Model
Five stages from reactive chaos to a governed, agent-assisted book. Find where your org actually sits, why most teams stall in the middle, and what to build before you automate.
Most renewal orgs want to "add AI" to their motion. Almost none of them are ready for the version they imagine.
The pitch is seductive: an agent that watches your book, catches every risk, drafts every save. But drop that agent onto a book where renewal dates live in a stale spreadsheet and health is a gut feel, and it does not save you. It automates the guesswork. The gap between a messy book and an agent-run one is not a single leap. It is five stages, and each one is built on the foundation of the last.
This is a maturity model for exactly that climb. Read it, find where your org honestly sits, and (most importantly) resist the urge to skip rungs.
The one rule: you cannot skip a stage
Before the stages, the rule that governs all of them: each stage requires the previous one's foundation. You cannot jump from a spreadsheet to an agent layer, because automation does not fix a broken process. It runs it faster.
Bill Gates put it best, in a line widely quoted from Business @ the Speed of Thought: "automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency." That is the trap this model exists to prevent. It is not just a renewals problem, either: in McKinsey's 2024 survey, even the AI high performers cited data as their biggest obstacle, with 70% reporting difficulties with data governance, integration, and quality. If the companies best at AI are held back by ungoverned data, a renewal team cannot skip that work either.
Every stage below is a prerequisite for the next, which is why "where are we, really?" is the only question that matters before you buy anything.
The five stages
Click a stage to see the tell-tale signs
Data is clean enough to trust. The book is segmented. The team works a consistent process.
Health scoring is defined, documented, and agreed on
The book is segmented by risk, size, and timing, and those segments drive different plays
A rep can pull their at-risk accounts in under two minutes
Forecast is built from data, not rep confidence
The threshold. Reach Stage 3 and you have earned the right to ask "what should we automate." Before it, automation just makes the mess faster.
Walking the climb
The interactive ladder above has the tell-tale signs for each stage. Here is the narrative: what each stage feels like, and what it actually takes to move up a rung.
Stage 1 → 2 is about capture. You get renewal data out of inboxes and spreadsheets and into a system with a single owner. It is unglamorous and it is foundational: you cannot govern what you have not captured. Most teams clear this rung the moment they buy a CRM and put someone in charge of renewals.
Stage 2 → 3 is the hard one, and it is where most teams stall. Having a CRM is not the same as trusting it. The middle stage (tooled but ungoverned) is a graveyard of half-filled fields, health scores no one acts on, and playbooks that live as documents instead of enforced workflows. Moving to Stage 3 is not a purchase; it is the discipline of making the data clean enough to trust, defining what health actually means, and segmenting the book so different accounts get different plays. This is the real work, and there is no tool that does it for you.
Stage 3 is the threshold. Data is clean, the book is segmented, the team works a consistent process. A rep can pull their at-risk accounts in under two minutes, and the forecast is built from data rather than confidence. Only here have you earned the right to ask "what should we automate?" because now automation has something trustworthy to act on.
Stage 3 → 4 is deterministic automation. Score changes trigger workflow steps. Tasks generate from data instead of a manager assigning them. Segmentation re-runs on a schedule and the book re-sorts itself. The team's capacity shifts from generating the list of accounts to look at to actually working them.
Stage 4 → 5 adds the agent layer, with a human in the loop. An agent monitors the book continuously, flags shifts in risk before a rep would have noticed, and drafts the work (outreach, briefs, escalations) for a human to approve, edit, or override. The "human-approved" constraint is not a limitation; it is the governance that makes an agent trustworthy enough to deploy on real revenue.
Where do you actually sit?
Be honest about the rung. The most common mistake is a team that has bought Stage 4 tools while its data is stuck at Stage 2, paying for automation on a foundation that cannot support it.
Diagnose your org
Where do you actually sit?
Check the statements that are true today. Be honest: the first one that is not true is your ceiling.
Why the middle is so crowded
If you landed in Stage 2, you are in good company, and it is worth understanding why the middle is where teams get stuck, and why it is worth the climb out.
The reason it is crowded is that Stage 2 is comfortable. You have tools. You have dashboards. It looks like maturity. But the data underneath is not trusted, so none of it drives action, and the team stays busy without the book actually being worked.
The reason it is worth escaping is that renewals are where the economics of a SaaS business live. The classic Bain research, published in HBR, found that cutting customer defections by just 5% could lift profits anywhere from 25% to 85%. And net revenue retention has become the number that most separates the best cloud companies from the rest: Bessemer's benchmarks show top performers sustaining well above 100% NRR while the bottom quartile falls below it. Every rung you climb turns more of your book from "hope it renews" into a governed, visible, worked motion, and the compounding value of that shows up directly in net revenue retention.
How to move up: build in order
The model is a diagnostic, but it is also a sequence. Three rules for climbing it:
- Diagnose honestly. Use the tell-tale signs, not your aspirations. Where you are determines what to build next; where you want to be does not.
- Do not skip. If your data is ungoverned, do not buy automation. Fix the governance first, or you will automate the mess. The rung below is always the prerequisite.
- Build the foundation before the ceiling. Stages 3 and 4 (clean data, health scoring, dashboards, reports, governed workflows) are the load-bearing work. The agent layer is the payoff, not the starting point.
This is exactly why BaseCommand is built the way it is: the foundation first, and free to start. Health scoring, dashboards, and renewal reports are computed from the CRM you already have (the Stage 3-and-4 work), with no data project required. (The Assembly Tax is the tax you pay for staying below Stage 3; here is why it compounds.) The always-on agent layer, Stage 5, comes later, deliberately, because it is only trustworthy on a foundation you have actually built.
You cannot buy your way to Stage 5. You climb to it. The good news is that the first rungs, the ones that matter most, are the ones you can start on today.
Want to see where your book sits? Start free on your CRM, or run the Renewal Reality Check on a CSV export first.