Most of us are using AI wrong in prospecting. Not because we don't know the tools — but because of what we're asking them to do.

We form a hypothesis. We paste it into AI. We get back a polished version of what we already believed. We call it research.

That's not intelligence. That's a mirror with better grammar.

Stop using AI to validate. Start using it to challenge.

‍

The Default Mode (and Why It Fails)

Here's what I see most reps do. They look at a target account, form an opinion about what's going on, then ask AI something like:

"This company is expanding fast — what are some good ways to position ours solution?"

AI obliges. It produces confident, well-structured reasons why the rep is right. The rep feels ready. They send the outreach.

And buyers don't respond — because the message was built on an assumption that was never tested.

The problem isn't the tool. It's the direction of the question. We asked AI to agree with us. So it did.

‍

Flip the Frame: AI as Challenger

What if instead, we asked AI to poke holes?

Not to validate the angle — but to stress-test it. To tell us what we're missing, where the logic breaks down, what the buyer might push back on.

Think of it like a sparring partner before a big pitch. You don't want someone who nods along. You want someone who finds the weak spots before the buyer does.

This shift changes how we prepare, how we prospect, and how credible we sound when we finally reach out.

‍

The 3-Part Account Intel Routine (Built to Challenge, Not Confirm)

Here's a practical workflow to build into your day. Every step is designed to surface what you might be wrong about.

Step 1: Build a sharp account snapshot

Give AI the raw material: the company website, LinkedIn page, and any recent news. Then ask it to summarize — but also to surface constraints.

Prompt:

Using the material above, summarize this company in 6 bullets: who they sell to, how they make money,their key offers, target market, competitive positioning, and likely KPIs —only include KPIs if supported by the material above. Then list 4 plausible growth constraints specific to this company’s size, stage, and sector. If the source material doesn’t support a specific answer, say so rather than generalizing.

What you're looking for: a summary sharp enough to trust, and constraints that sound commercial — not a generic list of challenges any company could have.

Note: If the constraints feel broad or obvious, push back. Ask AI to get more specific to the company's size, stage, or sector. Vague constraints produce vague outreach.

Step 2: Identify the trigger — then challenge it

This is where I see most reps stop. They spot a trigger and run with it. We want to go one step further: form a hypothesis, then ask AI to break it.

Prompt:

Based on the account summary above, identify 3 triggers that could create urgency for this company right now. Prioritize triggers supported by recent signals — hiring activity, leadership changes, funding, new products, or market shifts. Rank them by likelihood. For each: (a) the ‘why now’ logic, (b) what evidence I should look for to validate it, and (c) one reason this trigger might not be a priority for this buyer right now.

That last part— (c) — is the one most reps never ask for. It's also the most valuable. If AI can't find a flaw in our thinking, we're probably asking the wrong question.Push it harder.

This isn't about doubt. It's about credibility. When we walk in having already considered the counter-argument, our outreach lands differently.

Step 3: Build outreach angles — and make AI argue against them

Different personas respond to different pressures. A Head of Sales cares about growth and conversion. Rev Ops wants efficiency and process. Finance and Leadership are watching risk and predictability. Before we draft anything, we need to know which angle fits — and whether it actually holds up. This prompt does both.

Prompt:

Using the account summary above, identify which of the following three angles is most likely to resonate and why: growth/revenue, efficiency/process, or risk/quality. Then build out the top one or two only — with a one-sentence hypothesis, two supporting points, a question that tests it, and one reason a buyer here might push back.

We know which angle fits the account. And we know where it’s vulnerable before it goes out.

‍

Drafting the First Message: Short, Specific, Human

Once we've done the thinking, the email itself should be easy. But here's something worth pausing on: this prompt only works because of the two steps that came before it.

Without the account summary and the tested hypothesis, AI has nothing real to work with — and it will fill the gap with generic observations that buyers see through immediately. Prompt quality depends on what came before it, not just the words in the prompt itself.

Prompt:

Using the account summary and trigger hypothesis above, write a first-touch email under 100 words. Confident, direct, human — no buzzwords. Include: (1) a specific observation drawn from the account context, (2) the hypothesis we identified, (3) one thought-provoking question that tests it, (4) a soft CTA — yes or no only.

Write two versions: one direct, one warmer in tone. Do not invent details that weren’t in the account summary.

Then — and this is the part AI cannot do — run the human lens:

•      Add one real detail you noticed that AI wouldn't have

•      Remove anything that sounds polished but hollow

•      Ask yourself: does this sound like someone who gets their world, or someone running a sequence?

As Brene Brown says, clear is kind. The goal is not to impress. It's to surface something worth responding to.

One Goal: Get a Response

All of that preparation comes down to: the first message. It isn’t trying to close anything. It’s trying to find out: do they see what we see?

The ask is small on purpose. A binary question works because it lowers the barrier to reply:

•      "Is this something you're experiencing?"

•      "Is this worth exploring?"

•      "Is this on your radar right now?"

We also need to manage our expectations. Cold outreach can take up to 16 touches. One well-crafted message won't always move the needle. But one careless message — one that reads like we haven't done the thinking — will close the door faster than silence will.

The goal is to be the signal, not add to the noise.

The Mindset Shift That Makes All of This Work

AI is efficient. It moves fast, summarizes well, and generates options quickly. What it cannot do is replace judgment, curiosity, or the ability to ask a sharp question at the right moment.

So use it to get smarter faster — not to feel ready before we are.

The reps who stand out aren’t the ones using AI to produce the most outreach. They’re the ones using it to pressure-test the thinking behind a smaller number of better messages. That takes discipline — because it’s slower at first. But it compounds. The rep who sends ten well-reasoned messages will always out perform the one who sends a hundred that sound the same.

The best prospecting doesn't feel like prospecting.

It feels like: "This person gets my world — and they're asking something worth answering."

That’s the bar. AI can help us get there — but only if we’re willing to be wrong first.

‍

Want to bring this thinking to your team?

If your sales team is using AI but not seeing it translate into better conversations, it might be time to look at the prompts — and the mindset behind them. If that’s a conversation worth having, let’s have it.

‍

‍