How to Sell AI Products: Position Against Labor, Not the Tool

We make about a thousand cold calls a week for early-stage AI companies. Different products, different verticals, different buyers. And over the past several months, we noticed something we weren’t expecting.

The AI products that land in cold conversations aren’t the ones described as better tools. They’re the ones described in terms of the specific work that a prospect’s team is already doing — and positioned as taking that work on.

That’s the first part. The second part matters just as much: when a prospect still defaults to “we already have a tool for that,” the conversations that convert are the ones where the SDR gets the prospect to articulate what their people are actually doing day to day — and in doing so, acknowledge that no, their tool isn’t doing it.

The Default: Prospects Hear “Tool” and Compare

When an SDR describes a product by its category — platform, solution, software — the prospect’s brain does something efficient and destructive: it reaches for the nearest comparison. Not the nearest competitor. The nearest thing they already own that sounds like it’s in the same area.

None of these prospects actually had what was being sold to them. But the category label gave them permission to stop evaluating. Tool language triggers tool comparisons, and tool comparisons end conversations before the product’s actual value ever comes up.

The Insight: Describe the Work Their People Are Doing

The shift that worked wasn’t simply calling the product a “worker” instead of a “tool.” It was more specific than that. The conversations that converted were the ones where the SDR described the exact work that someone on the prospect’s team is currently doing — and positioned the AI as doing that same work.

“AI policy analyst” is better than “legislative tracking platform,” but only because it points to a real job with real deliverables. What actually lands is describing what that job looks like: reading every bill introduced, writing plain-language summaries, flagging amendments the day they happen, producing the stakeholder briefing. That’s not a product description. That’s a description of what someone on the prospect’s team spent Tuesday doing.

The Second Move: When They Still Say “We Have a Tool”

Even with labor-first positioning, some prospects still default to the tool comparison. “We already have something for that.” “We’re covered.” “We just bought a platform that handles this.”

This is where the second move matters — and it’s the one that turned the most conversations. Instead of arguing the comparison or stacking features, the SDR asks a question that forces the prospect to separate what their tool does from what their people still do.

Questions that don’t work

“Do you have an AI tool for this?” — invites a simple yes. “Have you looked at solutions in this space?” — too abstract. “What tools are you using?” — lets them hide behind a product name. Asserting “your tool doesn’t do this” — puts the SDR on offense and the prospect on defense.

Questions that work

Data Eng: “Are your engineers writing the code, or is something generating it for you?” Gov Affairs: “Who reads the bills, writes the analysis, and produces the briefing?” E-Commerce: “Is your chat handling pre-purchase selling, or handling support tickets?” Contact Ctr: “What % of calls does your system handle end-to-end without a human agent?”

Prospects who had just said “we already have that” would pause and say: “A little bit of both,” or “some manual, some automated.” That concession is the moment the tool comparison falls apart — because the prospect acknowledged it themselves. The SDR didn’t have to argue it.

Where It Breaks Down

Three Things to Build Into Any AI Outbound Motion

We figured this out the way we figure everything out — by making thousands of calls, reviewing every conversation, and paying attention to what actually moves prospects from “we’re covered” to “tell me more.” That’s the advantage of phone-first outbound for AI companies: every conversation produces signal.

The AI products we work with are genuinely new. They do things that didn’t exist two years ago. The challenge is that prospects evaluate them with categories that are ten years old. The best thing we’ve found to break through isn’t a better pitch — it’s getting the prospect to describe their own reality, and then showing them exactly which part of that reality the product was built to handle.