AI belongs in account research, not just email

Most AI-for-sales projects start with outbound email. That is understandable: email is easy to measure and easy to automate. But in enterprise selling, the bottleneck is rarely writing the email. It is knowing who to write to, what they care about, and why now.

In 2023 at AWS, I designed a go-to-market motion that was adopted in North America and Japan. It used ML-based account research, stakeholder mapping and prioritization. The result was a 13x increase in C-level engagements and a 3x increase in average deal size.

Where the leverage actually is

Enterprise sellers spend a large share of their time on research: reading filings, mapping org charts, figuring out which initiatives are funded, and deciding which accounts deserve attention this quarter. That work is essential, and it is exactly the kind of work machine learning can accelerate.

When you use AI to do the research and prioritization, sellers get time back for the work only humans can do: building trust with executives, shaping the problem, and orchestrating the internal team.

Three uses that matter

Account prioritization. Use signals to rank accounts by likely need and timing, so sellers spend their best hours on the best accounts.

Stakeholder mapping. Build a picture of the buying group, including executives the team has never met, and identify the right point of entry.

Executive-ready points of view. Give sellers a well-researched starting hypothesis for each executive conversation, so the first meeting is about the customer's business, not your product.

Why deal size goes up

When you engage higher in the organization with a sharper point of view, the conversation moves from features to outcomes. Outcomes are bigger than features. That is why the same motion that raised C-level engagement also raised average deal size.

What this means for mid-market companies

Mid-market enterprises often compete against larger rivals with far bigger sales teams. AI-driven account research is one of the few levers that lets a smaller team punch above its weight. It does not replace sellers. It makes each seller more effective at the highest-value part of the job.

How to start without a big platform project

You don't need a large data science team to begin. I usually start with a pilot on a single segment. First, define the signals that matter for your business, such as funding events, leadership changes, new initiatives or technology shifts. Second, use AI to monitor and summarize those signals for the pilot accounts each week. Third, have sellers turn the best signals into executive outreach with a clear point of view. Fourth, measure what changes: executive meetings, stakeholder coverage and the size of the opportunities created.

If the pilot works, expand it. If it doesn't, you have learned cheaply which signals don't predict anything for your market.

Keep the seller in the loop

AI research is a starting hypothesis, not a script. The best sellers use it to prepare and then listen carefully in the meeting. The research earns the conversation; the seller earns the trust. I'm wary of any AI program that tries to remove the seller from the moments where judgment and relationships matter most.

What to measure

I track a small set of metrics: the share of target accounts with an executive relationship, the number of stakeholders engaged per opportunity, average deal size, and the time sellers spend on research versus customers. Those numbers tell you whether AI is changing how the team sells, or just adding another tool.

If those numbers move together, the program is working. If only activity goes up, it isn't, and it's time to revisit the signals.

When I work with a mid-market company on AI in GTM, I start here, not with email automation.