How do you use AI agents in the sales process?
An AI agent in the sales process is AI support that carries out several steps on its own: watching signals, researching, preparing meetings, drafting and updating the CRM. It helps most before and after the customer meeting, not in it. The salesperson keeps the relationship, approves what goes out and makes the decisions.
How is an AI agent different from a chat assistant?
A chat assistant answers when you ask. An agent gets a task and carries out several steps to solve it: fetching data, comparing, drafting and putting the result in the right system. It works towards a goal, not just a question. That is why it needs clear limits on what it may do on its own and what a person must approve.
Where in the sales process do AI agents help most?
| Step | What the agent does | What the salesperson does |
|---|---|---|
| Targeting and buying signals | Reads up on companies, looks for signs that the need is current, prioritises | Chooses who to contact |
| Research and meeting prep | Compiles the company's situation, the people and previous contacts | Decides the angle and goal of the meeting |
| Follow-up | Structures notes, suggests next steps, drafts emails | Reads, adjusts, sends |
| Quotes | Interprets the request, finds similar deals, drafts a proposal | Sets price and terms, approves |
| CRM and reports | Keeps records up to date, compiles weekly reports | Uses the input to prioritise |
What does it look like in practice?
43 emails. 9 meetings. 3 deals. That is the best cold campaign I have been near. And it was not about what we sold for a second.
We chose a single problem, one the recipient recognised in a second. The email was three sentences: the problem, a number, a question. What made the difference was that we had decided on the problem before we started writing.
Today I let the AI support do the targeting first. It reads up on the companies and looks for signs that exactly that problem is live for them now: headcount growth, new offices, a change in how they work. A list of 500 becomes 40 where the question is probably on someone's desk this month. Then I write the email myself.
The effect is not that we reach more people. The effect is that we send fewer emails to more people who actually reply.
What should the salesperson always keep?
The relationship, the judgement and the decisions. It is the salesperson who hears that “we'll get back to you after the summer” means no. The AI agent should give the salesperson more time for the conversation, not take it over.
Which rules apply?
- If a customer meets the agent directly, for example in a chat, it must be clear that it is AI. The requirement is in Article 50 of the EU AI Act and has applied since 2 August 2026. Internal agents where the salesperson reviews are normally not covered. Read more in the guide to the AI Act.
- Details about contact persons are personal data. You need a legal basis, and the recipient must be able to opt out.
- Anything that goes to a customer is approved by a person.
Where do you start?
With one step, not all of them. Follow-up after meetings asks the least change. Targeting and buying signals often give the fastest effect in the pipeline. See Where should a B2B company start with AI?, Atlas, DealRadar and Proposo.
FAQ
Common questions about AI agents in sales
Can an AI agent handle cold prospecting on its own?
It can do the targeting and draft the messages. The salesperson should own the sending and the conversation, both for accuracy and for regulatory reasons.
Do we need to change CRM?
The agent works with the CRM you have.
How do you measure the impact?
Response rate, meetings per email sent and time per week on administration, before and after.
Where in the sales process do you lose the most time?
Bring a normal sales week to a conversation. Together we will see where an AI agent helps most.
Book 30 minutes