"AI" in sales tools covers three different things. Knowing which one a vendor is selling is the first skill a revenue leader needs.
1. Predictive AI: scores and forecasts
Models trained on your history estimate what will probably happen: which lead is likely to convert, which deal is at risk, which customer may churn, what the quarter will close at. This has existed for years (lead scoring, forecasting tools). It is only as good as the CRM data it learns from.
2. Generative AI: drafts and summaries
Large language models (LLMs) such as those behind ChatGPT, Claude, Gemini or Copilot predict the next words of a text. That makes them good at drafting (emails, proposals, account plans), summarizing (calls, meetings, long email threads) and answering questions over documents. They can also be confidently wrong. They do not "know" your customer unless you give them the data.
3. Agentic AI: taking actions
An agent is an LLM connected to tools: it can look up a contact, write an email, send it, read the reply and book a meeting, with little or no human step in between. AI SDR agents and customer-service agents are the best-known examples. Agents multiply both value and risk, because mistakes go out at machine speed.
How common is it?
Salesforce's State of Sales 2026 found that 87% of sales organizations use some form of AI, for tasks like prospecting, forecasting, lead scoring or drafting emails. So the question for a CRO is no longer whether the team uses AI, but where it moves a revenue number and where it only adds activity.
A simple test for any AI pitch
Ask three questions:
- Which kind is it — predict, draft, or act?
- Which funnel metric should move — and by how much, against what baseline?
- What does a human check before it reaches a customer?
If the vendor can't answer the second question with your metrics, you are buying a demo, not revenue.