Trends and Reports

The Next Big Shift in Sales Isn't AI Tools. It's AI Audiences.

In an industry built on humans helping humans, our client's biggest pitch of the quarter had an audience of one. It wasn't human.

The Company Advice Team

Posted on

September 9, 2026

Key Takeaways

One of our clients recently delivered a full sales pitch to an AI system standing in as the buyer. No human ever joined the call, and no human followed up afterward. In an industry built on personal missions and human trust, that's a bigger signal than any adoption stat.
There's a real difference between an AI notetaker, which just summarizes what you say, and an AI agent, which is built to score the conversation, act as a gatekeeper, and decide what happens next. Health tech and B2B sales are colliding with the second kind faster than most founders realize.
Your pitch now has two audiences: the AI that summarizes or scores it, and the human who reads that recap later. Pitches that are specific, clearly structured, and built around one repeated idea survive that trip intact. The personality that doesn't survive it has to move somewhere a human will actually see it.

When the Buyer Didn't Have a Pulse

‍

One of our clients logged into a scheduled pitch call not long ago. Right time, right link, calendar invite accepted by the other side days earlier. The call opened. A window populated. And after a minute or two of the kind of throat-clearing pleasantries you make while waiting for someone to unmute, it became clear that no one else was joining. Not the exec who'd booked the meeting. Not an analyst sitting in. Not even an assistant standing in for someone running late. Nobody.

What was on the other end was an AI agent, present enough to count as an attendee, running the meeting without a camera, a name plate, or a pulse behind it. We don't know exactly which tool it was. What we know is that our client didn't get a warning beforehand, and didn't get a callback afterward, either. No human ever circled back to explain, debrief, or so much as confirm the meeting had happened.

They were caught off guard, understandably. And they did what any of us would do in that position: they delivered the pitch anyway, as well as they possibly could, to whatever was actually listening on the other end of that call.

‍

The Business of Humans Helping Humans

‍

Here's what makes that story sting more in health tech than it would almost anywhere else. This industry runs on the idea that the sale is personal. Founders build these companies out of a diagnosis, a broken referral process, a system that failed someone they loved, and the pitch is supposed to be the moment that story lands on another human being who gets it. “Humans helping humans” isn't just a tagline in this space… it's close to the entire premise of why these companies exist and why buyers say yes.

So when the room a founder walks into for the biggest pitch of the quarter turns out to have no one in it, that's not just an awkward Zoom call. It's a preview of a much bigger shift: the moment where the emotional core of the sale has to survive a step that has no emotions at all.

‍

Recruiting Got There First (And It's Already a Punchline)

‍

AI isn't just taking notes in meetings anymore. It's sitting in on them, and in some cases, running them — and recruiting got there first, loudly. Ninety-one percent of hiring professionals now use AI somewhere in their process, and one of the biggest names in AI recruiting, Mercor, jumped from a $10 billion valuation to talks of $20 billion in under a year.

It's also become a punchline. There's a clip going around of a candidate's AI interviewer glitching mid-question, stuck looping “tell me about a time when when when…” on repeat like a scratched CD– and the candidate posting it, thrilled, because it got him out of a bad interview and into a viral moment instead. That's the genre right now: AI-interview fails as slapstick, comedy built entirely out of the gap between how seriously the system takes itself and how little it actually understands.

Watch the clip — TikTok, @leohumpsalot.

@leohumpsalot I WAS SO HAPPY YALL 😭😭😭😭 #fyp #jobs #jobsearch #ai ♬ original sound - Leo Humps

But funny doesn't mean harmless, and it doesn't mean being contained within recruiting. It's gotten common enough across customer success, healthcare, legal, education, and field sales that an AP report this July found lawyers and privacy experts actively coaching people on how to handle a bot showing up to a meeting: what to say to get it turned off, how to protect your voiceprint (the biometric signature built from your unique speech patterns — basically an audio fingerprint), and what happens to the transcript afterward. We laugh at the glitch clip and then, a beat later, remember our own client had a version of this happen with nothing funny about it at all.

‍

The AI Use Case Lineup

‍

So here's the lineup as it stands: AI recruiters are already running live interviews. Notetakers are everywhere. Chances are you've had to admit one into your last Google Meet, or had a scribe running at your last doctor's appointment. Agents that run meetings themselves, asking questions and flagging follow-ups, are starting to show up too. And now, as our client found out, agents built to stand in for the buyer and take the meeting entirely are already here as well. The open question isn't whether AI shows up to your sales calls anymore. It's which role it's playing when it does.

‍

Is the Next Step Agents Selling to Agents?

‍

What surprised us about our client's call wasn't that an AI system was in the room.. we’ve gotten used to a notetaker lurking by now. What surprised us is that it was running the show: standing in as the buyer, on the receiving end of the sale, entirely on its own. And that raises a stranger question we don't have a full answer to yet. If AI can sit in on a sales call as the buyer, how long until AI is running the seller's side too?

We haven't seen it yet — a fully agent-to-agent sales call, where one company's AI rep pitches another company's AI buyer with no humans on the line at all — but the pieces are already sitting on the table. An AI system that reads a deck, scores a conversation, and produces a recap is functioning as a procurement filter. It isn't a huge leap from there to a seller building an AI counterpart trained to hit that filter's criteria directly, optimizing not for a human's attention span but for whatever heuristics the buyer's agent is running underneath it.

That's not just a hypothetical, either. In March 2026, Stripe and the payments company Tempo launched the Machine Payments Protocol, an open standard built specifically so AI agents can pay for things without a human clicking "confirm purchase." The infrastructure for one agent transacting with another, no human in the loop, is already live.

If the agent that scores a pitch and the agent that pays for it are both already real, the recipe that wins a deal shifts again. Not “did the human like us,” not even “did the summary read well,” but “did we structure our claims, our proof points, and our objection handling so a machine scores us correctly.” That has real implications past the sales call, for marketing and brand too: positioning built for how a person remembers you starts to matter less than positioning built for how a system parses you — consistent language across every asset, claims that are specific enough to be verified or cross-referenced, less weight on tone and more on substance a machine can actually retrieve.

We don't think this replaces the human, relationship-based sale, especially in a field like health tech that's built on trust between people. But we'd be lying if we said we weren't watching for the first agent-to-agent deal to close, because from where we're sitting, that clock already started.

Now that we know what we're up against, let's talk pitch.

If your pitch only works because someone's in the room reacting to your charm, it's already outdated. Your pitch now has two audiences: the AI that summarizes or scores it, and the human who reads that summary later. Which means, whether we like it or not, we have to win over the robot first.

‍

How to Pitch to a Robot 101

‍

Lead with the point. Bury your value prop in an elaborate build-up, and it might get summarized right out of existence. Say the thing before you lead up to it.
Be specific, not vibey. “We saved a client 30% on X” survives compression. “We're passionate about delivering value” gets cut; there's nothing in it to extract.
Make the ask impossible to miss. These tools spit out action items. Bury your next step in a story and it might not make the list. Say it obviously, and say it twice.
Give it a sticky name. A pitch with a clear, memorable label (“the 30-day rollout,” “the no-code fix”) is easier for a summarizer to hang onto than a paragraph of good ideas with no headline.
Relocate the personality, don't drop it. Your delivery, your timing, your presence– none of that survives transcription. And if your business is built on the human mission behind it, that mission can't live in the part of the pitch a machine is compressing. Put it somewhere a human will actually see it: the follow-up, the recap, the proposal.

‍

We think about this stuff for a living. If you want a second opinion on whether your pitch can survive contact with a robot, let's get 20 minutes on the calendar and find out. Book a call with The Company Advice.

‍

References

Spencer, Keith. “AI and Hiring Trends 2025.” Resume Now, 20 Nov. 2025, www.resume-now.com/job-resources/careers/ai-trends-for-2026-report

“Mercor Quintuples Valuation to $10B With $350M Series C.” TechCrunch, 27 Oct. 2025, techcrunch.com/2025/10/27/mercor-quintuples-valuation-to-10b-with-350m-series-c

Nieva, Richard. “AI Data Labeler Mercor In Talks To Raise $500 Million At $20 Billion Valuation.” Forbes, 9 July 2026, www.forbes.com/sites/richardnieva/2026/07/09/mercor-fundraise

Bussewitz, Cathy. “AI Notetakers Promise Easy Meeting Recaps, but Some Professionals Question Their Use.” Associated Press, via U.S. News & World Report, 9 July 2026, www.usnews.com/news/technology/articles/2026-07-09/ai-notetakers-promise-easy-meeting-recaps-but-some-professionals-question-their-use

Weinstein, Jeff, and Steve Kaliski. "Introducing the Machine Payments Protocol." Stripe, 18 Mar. 2026, stripe.com/blog/machine-payments-protocol.

Frequently asked questions

What's the difference between an AI notetaker and an AI agent?

A notetaker listens and summarizes. It's not judging your pitch, it's compressing it for a human to read later. An agent goes further: it's built to score the conversation, flag objections, and decide what happens next — including, as we learned firsthand, standing in for the buyer entirely. Recruiting already has both. Sales is catching up fast.

How many people actually use AI in meetings?

Roughly 75% of professionals now use AI in their meetings, according to a 2025 Fellow.ai survey cited in Laxis's 2026 industry report — about double the adoption rate from 2023. That figure covers AI in meetings broadly (notetaking, summaries, scheduling); a separate, narrower figure — 91% of hiring professionals using AI somewhere in recruiting specifically — shows just how much further along that one function is.

Can an AI agent attend a meeting without a human present?

Yes — we've seen it directly. AI tools and agents have become common enough in meetings generally that, as of July 2026, lawyers and privacy experts are actively coaching professionals on how to handle one being present, and in our client's case, present was the only thing it was.

How do you pitch effectively to AI?

Lead with your core point instead of building up to it, use specific numbers and outcomes instead of vague claims, state your ask clearly more than once, and repeat one sharp idea rather than spreading several across the call — summarization tools tend to extract what's clearest and most repeated. Learn more strategies in the blog post above.

Will AI eventually replace sales meetings and pitches entirely?

Probably not entirely, based on current research. Industry analyses from firms like BCG and McKinsey describe AI as more of a co-pilot in B2B sales — handling scheduling, prep, and follow-up — rather than a full replacement for human sellers, especially on complex or high-stakes deals. Sales leaders report that fully automated, AI-led outbound tends to underperform on show rates, qualification accuracy, and brand impact, which is why a hybrid human-AI approach is currently considered the more effective model. There's a further wrinkle worth watching, though: if AI is already sitting in as the buyer, it's plausible sellers respond in kind by building AI counterparts of their own — agent-to-agent sales — though that hasn't emerged as a real pattern yet, as of the time of writing.

Your Cart