AI in a sales team

Why AI in sales starts with enquiries, not leads

Sales teams are sold AI from the acquisition side: automated "SDRs", cold-email sequences, lead scoring. Those promises have two problems the offers stay silent about. First, mass cold outreach is a legal and reputational risk — consent rules, electronic-communication regulations, and a sending domain nobody answers after a quarter. Second, a model that scores leads has to learn from something, and most companies do not have their sales history in a form a model can learn from.

Meanwhile, in the same team, revenue leaks daily in far less spectacular places: a customer enquiry waits in a mailbox for two days because the salesperson is on the road; an offer takes an hour because prices have to be gathered from the price list, emails and previous quotes; the CRM does not reflect reality because updating it loses to making calls; the follow-up never goes out at all. These are document- and text-based processes — the data exists, the rules can be written down, and a salesperson approves every result anyway.

That is why a good first AI project in a sales team looks unspectacular: it does not promise new customers, it shortens the path from enquiry to offer for the ones you already have. In return it comes together in weeks and pays for the next steps.

What we actually automate in sales teams

• Enquiry qualification and registration. Reading enquiries from email, forms and portals, creating the opportunity in the CRM with complete data and a proposed priority — for the salesperson to approve. The enquiry stops waiting for someone to come back from the road and retype it.
• Quote and offer preparation. Assembling a quote from the price list, the customer’s terms and the history of similar orders — as an offer draft for approval, not as an email sent behind the salesperson’s back. It shortens the stretch where deals are lost today to a faster competitor.
• RFP and tender-question responses. Preparing answers to the recurring questions in requests for proposals from earlier responses and documentation — with a list of the questions that need a human decision, instead of a hundred pages to read from scratch.
• CRM upkeep. Preparing the record after a call or an email thread — contact, agreements, next step — approved with one click. The CRM starts reflecting reality without the evening note-typing session.
• Follow-ups and opportunity tracking. Prepared reminders and follow-up drafts for opportunities that have stalled — based on what the CRM says was agreed, approved before sending. Nothing goes out by itself; what was supposed to go out stops getting lost.

The common denominator: a person approves the result, the system prepares it. Which processes qualify for this kind of automation at all — the five conditions a good candidate must meet — is described on the AI automation page; in sales those conditions are checked exactly the same way.

What we do not promise

• A cold-email machine. AI-generated mass outreach is a legal risk — consent and electronic-communication rules — and a reputational one that no dashboard shows: a burned domain and a spammer’s reputation outlast the campaign. We do not build this, including on explicit request.
• "AI will close the deal for you." The sales conversation, negotiation and the discount decision stay with people. AI takes the repetitive work around that conversation off the salespeople — retyping, assembling, formatting — not the conversation itself.
• Lead scoring without a data history. A model predicting which opportunities will close fastest has to learn from past opportunities — consistently described, with outcomes. If the CRM does not have that, an honest offer starts by putting the data in order and says plainly that scoring comes later.

How we work

The same way as in every implementation: one process, a number measured before the start — for example hours from enquiry arrival to the offer going out — the narrowest possible pilot, and parallel running before the switch. At first the system prepares and the salesperson approves every result. The full course — four phases and what you get at the end of each — is described by the delivery methodology, and the full scope of the service, from the process audit to maintenance, by the AI implementation page.

A separate topic that comes back in every sales conversation is data: customer contacts and order history, price lists, commercial terms. Where data goes in a model integration, and what to ask any provider, is set out on the security and GDPR page.

What it costs

Three things drive the cost: the number of systems to connect, the state of the data, and the level of certainty required. Team size matters less — what counts is whether the price lists and discount rules are written down or live in salespeople’s heads. All the components, maintenance included, are broken down in our article on what an AI implementation costs.

Frequently asked questions

Where should a sales team start with AI?

With the enquiries you already receive, not with generating new ones. Registering enquiries from email and forms in the CRM, assembling quotes from price lists and the history of similar orders, answering recurring questions from requests for proposals — these are document- and text-based processes where the data already exists and a salesperson approves every result anyway. The effect shows in weeks: a shorter path from enquiry to offer. Cold-outreach automation is a different category of risk — legal and reputational — and the wrong first step.

Will AI write to our customers on its own?

Not from day one, and not without a human decision. Everything that goes out to a customer — offers, replies to enquiries, follow-ups — is produced as a draft for the salesperson to approve. Only pilot data shows which categories of reply are repetitive and safe enough to loosen the oversight; that decision is made deliberately, on numbers, not on assumption. The salesperson stops retyping and formatting, but still decides what goes out and to whom.

Will this work with our CRM?

That is a question the audit at the start of a project settles, but the principle is simple: if the system has an API or file import, integration is a matter of scope, not possibility. The more common problem is different: part of the customer knowledge never reaches the CRM at all, because it lives in salespeople’s inboxes and notes. Automatically preparing the record after a call or an email thread — approved with one click — is often the first project precisely because the CRM then starts reflecting reality.

How much does an AI implementation cost in a sales team?

The cost is driven by the number of systems to connect — CRM, email, price lists, quoting tools — the state of the data, and the level of certainty required. Team size matters less than whether the price lists and discount rules are written down or live in salespeople’s heads: the first case is a matter of weeks, in the second the project starts by writing them down. Maintenance is a separate line: usage-dependent fees and periodic tests after model version changes. We broke the cost components down in our article on what an AI implementation costs.

What about customer data and price lists?

Enquiries and order history are personal data of the contacts on the customer side, and price lists and commercial terms are the information your margin depends on. Before the integration, not after it, you need to settle which data leaves the company at all, where it goes and on what basis — from the choice of provider and processing region to limiting the data to the minimum a given step needs. The rules, and the questions to put to any provider, are set out on our security and GDPR page.