Straight answer: AI notetakers like Otter, Fireflies and Fathom do save you time, but they change the way you listen on a call, and most people using them have not noticed the drop in how present they are with a client. The fix is not ditching the tool, it is using it on purpose instead of on autopilot.
The call where I realised something had gone wrong
I was on a discovery call with a prospect last year, a small ecommerce brand owner looking at AI for her customer service. Forty minutes in, she stopped mid-sentence and said, “Sorry, is that bot listening to everything I say?” I’d forgotten to mention it. Fireflies had joined the Zoom automatically, sitting there in the participant list with its little icon, and she’d only just noticed it.
She wasn’t angry, she was just unsettled. And the call never quite recovered its rhythm after that. She got shorter with her answers. She stopped telling me the messy, off-brief stuff, the bit where people usually reveal what they really need. We didn’t close that deal.
That call made me go back through my own notes from the previous six months. I’d been running an AI notetaker on almost every discovery call, telling myself it freed me up to be more present because I wasn’t scribbling. My close rate on those calls had dropped from roughly 40 percent to 28 percent over that stretch. Not a scientific study, just my own pipeline, but it was a big enough drop that I stopped pretending it was a coincidence.
What these tools change about a call
The pitch for Otter, Fireflies, Fathom and Gong is always the same: never miss a detail, get perfect notes, follow up faster. All true. What nobody selling these tools tells you is what happens to your brain once you know the recording exists.
You relax. Not in a good way. You stop taking the small mental notes you used to take because you know the transcript will catch it. That sounds efficient, but listening isn’t just information capture, it’s the little nods, the pause before you answer, the follow-up question you only ask because you were really tracking what someone just said rather than half-tracking it while thinking about your next talking point. Once the AI is doing the remembering, a lot of people quietly stop doing the noticing.
Clients feel this even when they can’t name it. It shows up as you asking a question they already answered five minutes ago, or missing the throwaway line that was the real objection. Small business owners tell me all the time, “the salesperson seemed nice but I didn’t feel heard.” Nine times out of ten that salesperson had a perfect transcript and a mediocre memory of the actual human in front of them.
The trust problem nobody puts in the case study
There’s a legal layer to this too, and it’s not the interesting bit, it’s the annoying bit that people skip because it slows down the demo. In the UK you need consent to record a call under data protection law, and in plenty of US states you need it too. Most people running an AI notetaker click “join” and move on, and the AI bot announces itself in the participant list, but announcing isn’t the same as asking.
I now say a version of this at the start of every call: “I’ve got an AI notetaker joining so I don’t miss anything, that alright with you?” Takes four seconds. Nobody has ever said no. But the four seconds does something else too, it resets the tone of the call to one where I’m clearly still the one paying attention, the AI is just backup.
That distinction matters more than most AI marketing content admits, because the entire pitch of these tools is “let the AI handle it so you don’t have to.” For admin, brilliant. For the moment where a prospect is deciding whether they trust you with their business, letting the AI handle it is the wrong instruction to give yourself.
Where AI notetakers earn their place
None of this means throw the tool out. I still use one on:
- Internal team calls, where the point is action items, not connection
- Long client onboarding sessions with lots of technical detail I can’t hold in my head
- Calls where multiple people need the notes afterwards and typing them up myself would eat an hour I don’t have
- Any call I need a verbatim record of for compliance or a dispute later
Where it earns its keep is on volume and detail. Where it costs you is on the calls that decide whether someone hires you at all. First discovery calls, pricing conversations, anything where the client is still working out if they like you as much as they like your service.
Five rules that fixed it for me
- Tell the client it’s there, out loud, every single time, no exceptions
- Turn it off for the first ten minutes of a first call. That’s the bit where rapport gets built, and you don’t need a transcript of small talk
- Take three of your own notes by hand even with the AI running. It forces you to decide what matters instead of outsourcing that judgement
- Read the AI summary within an hour, not three days later, so you can spot what it missed while the conversation is still fresh in your own head
- Never let the AI write your follow-up email word for word. Read the summary, then write the email yourself in your own voice
The bit AI marketing content keeps getting backwards
A huge amount of AI for small business advice treats “save time” as the whole story. Save time on notetaking, save time on emails, save time on proposals. Nobody asks what you’re doing with the time you saved, and in my experience most people aren’t reinvesting it into being more present with the next client. They’re just cramming in one more call.
That’s the uncomfortable bit. The tool doesn’t make you better at your job, it makes you faster at the parts of your job that were never the hard part anyway. The hard part was always the listening, the reading the room, the noticing that someone’s tone changed when you mentioned price. AI can’t do that for you, and treating it like it can is how you end up with a beautifully organised transcript of a deal you lost.
Brands that build trust at scale understand this instinct even when they’re not selling anything one to one. Look at how GoPro built its brand on capturing real, unscripted moments rather than polished summaries of them, or how Figma’s whole growth story was built on people feeling seen inside a shared workspace, which I broke down in how Figma built a brand people trust. The pattern in both is the same one I learned the hard way on that Zoom call: people can tell the difference between attention and the appearance of attention, and they reward the real thing.
What I’d tell a small business owner starting out with this
If you’re running a service business from home, taking client calls between school pickups and everything else, the temptation to let AI carry more of the call is completely understandable. You’re busy. But I’d rather you use the AI notetaker on your second and third calls with someone, once trust already exists, and go in unassisted on the first one. That first call is where you’re being auditioned as a human being, not just as a service provider.
If you want a second opinion on which parts of your client process are safe to hand to AI and which ones will quietly cost you deals if you do, that’s exactly the kind of audit an AI consultant for small business should be doing with you before you roll a tool out across your whole sales process, not after you’ve noticed your close rate has slipped and you’re not sure why.
It’s the same discipline that shows up in good content strategy, which is a subject I go back to often because it never stops being true. Joe Pulizzi’s whole career has been built on the idea that relationships beat tactics, and a call where the client feels like a transcript rather than a person is a tactic winning over a relationship. Same with email. Mailchimp built an entire business on the idea that automation should feel more human, not less, which is worth remembering next time you’re tempted to let an AI summary do your follow-up thinking for you, an approach I unpacked in how Mailchimp built a brand people trust with automated email.
The honest test to run on yourself
Here’s the check I now run once a month. I pull up three recent call transcripts and ask myself, without looking at my notes, what the client’s actual concern was underneath what they said. If I can answer that instantly, I was listening. If I have to read the transcript to remember, the AI was doing my job for me and I was just nodding along on camera.
That test costs nothing and takes ten minutes, and it’s told me more about where my business is leaking deals than any dashboard has. Marketing tools like the ones I’ve written about, from Instagram’s approach to building an engaged audience to interactive calculators that pull visitors into a real conversation with your brand, all share one thing in common: they work because they make someone feel like the exchange is about them specifically. An AI notetaker running silently in the background does the opposite unless you’re deliberate about it. It makes the exchange about efficiency. Efficiency doesn’t close deals. Feeling understood does.
Frequently asked questions
Should small businesses stop using AI notetakers on client calls?
No, but stop using them by default on every call without thinking about it. Save them for internal meetings, onboarding, and calls after trust is already built, and go without one on first discovery calls where the client is still deciding whether they like you.
Is it legal to record a client call with an AI notetaker without asking?
In the UK you need consent under data protection rules, and many US states require it too. The bot appearing in the participant list is not the same as consent, so say it out loud at the start of the call rather than assuming the icon covers you.
Do AI notetakers hurt sales calls?
They can, mainly because knowing the AI is capturing everything makes people listen less carefully in the moment, not more. Over six months of running one on nearly every discovery call, my own close rate dropped from around 40 percent to 28 percent, which was enough to make me change how I use them.
What’s a better way to use AI on client calls?
Tell the client it’s there, skip it for the first ten minutes of rapport building, take a few notes by hand anyway so you’re still actively deciding what matters, and write your own follow-up email in your own words rather than pasting in the AI summary.
If you would rather not build this yourself, here is AI automation for ecommerce stores.
Related reading: What Your AI Meeting Notetaker Is Doing With Your Client Calls and The AI Notetaker Sitting In On Your Client Calls Might Be Breaking Your NDA.
