Is your sales team AI-ready? Start with the CRM.

The answer, firstOnly if the CRM is. AI tools reason over the data you already have. If half your deals have no next step, no owner and no stage-entry date, an AI will summarise the mess faithfully and confidently, and you’ll trust it because it’s well written. AI-readiness for a sales team is a data-quality question before it’s a tooling question: consistent stages, required fields that get filled, and a process the team actually follows. Build that and the tools work. Skip it and they hallucinate your pipeline.

Chanel Greeff · 23 August 2026 · 3 min read

What the tools actually need

Every AI feature being bolted onto CRMs right now (deal summaries, next-step suggestions, forecast commentary, call notes pushed into records) does the same thing underneath: reads the fields, reads the notes, writes a sentence. The quality of the sentence is capped by the quality of the fields. A deal with a stage, an owner, a dated next step, a decision date and three logged touches produces a useful summary. A deal with a name and an amount produces a plausible one, which is worse.

The five readiness checks

Stage definitions are written down and agreed. If “qualified” means three things to three reps, no model can infer which one applies.
Next step and owner are required, and filled. Check the blank rate. Over 20% and the AI is guessing on a fifth of your pipeline.
Stage-entry dates exist. Without them, no days-in-stage, no stall detection, no honest forecast commentary.
Notes go into the CRM, not email. The model can only read what’s in the record. A team that lives in WhatsApp has an empty CRM, however clever the tool on top.
Closed-lost has a reason. Every pattern-finding feature depends on it. “Other” as the top reason means there is no data.

In what order

Process first: the stages and their entry evidence, agreed by the team. Then the CRM configured to enforce it: required fields, stage gates, deals that can move back. Then a month of the team running it with somebody owning hygiene. Then, and only then, switch on the AI features, because they’ll have something true to read. It’s the order the fixes run in, Reset my stages before Clean up my CRM, which is why the data is on my menu and the AI features aren’t.

What I’m not saying

That you need a particular tool, or any tool yet. The businesses I work with are mostly under thirty people, and for most of them the honest AI strategy for the next year is a clean CRM and the features already inside monday.com or HubSpot. The data discipline is the investment; the tooling is a toggle. If you’re not sure which side of that line you’re on, the three-salespeople test will tell you in five minutes.