My objection wasn't that AI was useless. It was that sales is fundamentally a relationship business. The parts of the job that matter — building trust, reading a room, navigating complex buying dynamics — aren't things any model can replicate. What I missed was that those weren't the parts of the job I was spending most of my time on.
The first experiment
I was preparing a proposal for a €2.5M deal. Five stakeholders, a technical procurement process, a competitor with a stronger price position. I usually spent a full day on a proposal like this. I tried drafting the executive summary, risk section, and pricing justification with AI assistance. It took three hours. The quality was better than my usual first draft.
What surprised me
The proposal wasn't better because the AI was smarter than me. It was better because the prompting process forced me to be more precise about the client's actual problem, our specific differentiation, and the objections I expected. That discipline produced sharper thinking — which produced sharper writing. The AI accelerated the process and improved the output simultaneously.
The current workflow
I now use AI in three parts of every sales cycle. In research: building account intelligence before the first call. In communication: drafting outreach, follow-up emails, and proposals. In debrief: analysing lost deals for patterns I might have missed. None of this replaces judgment. All of it gives me more time and better information to apply it.
Advice for other sceptics
Don't try to AI-ify your highest-leverage activities first. Start with the work that feels like overhead: research, email drafts, CRM updates, proposal templates. When you recover three hours a week from that overhead, you can decide how to reinvest it. For me, that reinvestment went back into relationship work — the part of the job I'm actually best at.
The question isn't whether AI belongs in sales. It's whether you'd rather spend your time on the 20% that builds the relationship or the 80% that supports it.