AI Product Management Lesson #2
Humanizing AI by Reducing Cognitive Load
We all hate taking meeting notes. Or we don’t take them … and forget everything. In most meetings, you’re either (not counting distractions / multitasking):
- paying attention or
- writing things down
Anyone claiming to do both well is lying (or not listening).
This isn’t a productivity issue. It’s biology.
The original sin: asking humans to multitask
Meetings ask us to:
- listen
- think
- react
- decide
And then we add:
“Also, please document this perfectly.”
For years, we treated this as a discipline problem.
It wasn’t.
It was cognitive overload.
A brief, slightly embarrassing history of trying to fix this
Phase 1: The Moral Phase (pre-2010)
“Take better notes.”
Translation: try harder to be less human.
This gave us notebooks, templates, and the dreaded “note-taker” role.
Phase 2: The Tool Phase (2010–2016)
Evernote. OneNote. Shared Google Docs.
Founders believed better tools would solve the problem.
They didn’t.
They just made multitasking prettier.
Phase 3: The Recording Phase (2016–2019)
“Just record the meeting.”
We did.
And then never watched it again.
We didn’t fix memory.
We created archives of regret.
Phase 4: When AI finally learned to be polite
The breakthrough came when a few founders asked a dangerous question:
“What if the user shouldn’t do this at all?”
That’s when products like Otter and Fireflies clicked.
They didn’t say:
“Here’s a smarter way to take notes.”
They said:
“Don’t take notes.”
The AI listens.
Quietly.
Patiently.
Like a background character doing the hardest job in the room.
That’s humanizing AI.
Now here’s where Tell Mel quietly teaches the same lesson
At first glance, Tell Mel and meeting-note tools seem unrelated.
They’re not.
Tell Mel starts with a similar truth:
People want to preserve stories … but the moment you ask them to:
- install an app
- learn an interface
- type or navigate
You lose them.
Especially if they’re elderly.
The founders didn’t ask:
“How do we make AI smarter?”
They asked:
“What’s the least demanding way a human already knows how to share stories?”
The answer was obvious … and almost no one takes it seriously anymore:
A phone call.
Just talk.
No learning.
No remembering.
No cognitive strain.
Same philosophy, different domain
Otter. Fireflies. Tell Mel.
Different products.
Same product wisdom:
Humanizing AI means removing effort, not adding intelligence.
- Meetings → stop writing
- Memories → stop typing
- Stories → stop learning software
The AI does the work so the human doesn’t have to.
The AI Product Management lesson hiding in plain sight
As an AI PM, this is the uncomfortable question you must ask:
What task am I asking the user to keep doing, simply because “that’s how it’s always been done”?
If your product still requires:
- attention
- training
- constant interaction
You may have built a capable system … but not a humane one.
A quiet acknowledgment to the founders who got this right
It’s worth acknowledging the founders who saw this early:
- Sam Liang (with Yun Fu and Eli Calderon Morin), founder of Otter AI, for realizing that presence matters more than perfect notes.
- Krish Ramineni (with Sam Udotong), founder of Fireflies, for letting people participate instead of transcribe
And of course,
- Charlie Graham and Dr. Justin V. Graham, for trusting something as unfashionable … and as human … as a phone call
Different journeys.
Same instinct.
The simple rule that keeps showing up
If your AI product demands attention, it’s not humanized yet.
The best ones feel boring.
Invisible.
Almost obvious in hindsight.
That’s not a lack of ambition.
That’s product maturity.
Originally published at
https://www.linkedin.com/pulse/ai-pm-masterclass-2-humanizing-reducing-cognitive-load-zztec
