A founder said yes to a discovery call last Tuesday because the email was nice and the ask sounded reasonable. By Sunday evening she had lost three hours over six days to a conversation that was never going anywhere. The meeting itself was forty minutes. The other two hours and twenty minutes were the back-and-forth before it, the calendar reshuffle when his assistant changed the time twice, the mental warm-up the morning of, and the quiet ten minutes afterwards trying to remember why she had agreed in the first place.
That is the shape of the problem this post is about. The cost of a wrong yes is rarely the meeting on the calendar. It is everything that happens around it. AI as a triage layer between you and the inbound asks closes the gap by handling the front door before you open it.
What is AI as a triage layer for inbound asks?
A triage layer is a thin model running between your inbox, your DMs, and your eyes, reading every inbound ask against your declared yes/no criteria for the quarter and drafting either a polite no, a clarifying ask, or a green light to a real reply. You are still in the loop, you read the draft and send it, but the categorisation work is done. The pattern is closer to an editor than a gatekeeper.
The criteria are the load-bearing piece. Greg McKeown’s Essentialism and Derek Sivers’s “Hell Yeah or No” both rest on the same idea, that a clear binary filter beats a vague one every time. Warren Buffett’s 25/5 method goes a step further, name your top five and treat the other twenty as actively avoid. Without something concrete written down, the model has nothing to triage against and it decays into vibes.
Why does it matter for your business?
It matters because the inbound surface for an owner-operator has no natural ceiling, and the cost of a wrong yes compounds harder than the headline numbers suggest. Bain has knowledge workers at roughly a quarter of their week on email, and Microsoft’s Work Trend Index documents the after-hours creep that pushes founders well past that. Two to four hours vanish into asks that should not have got through.
The hidden tax is not the meeting time, it is what Cal Newport calls attention residue. Each yes leaves a cognitive trail. You think about the meeting before it, you carry it for a stretch afterwards, and the deeper work you meant to be doing runs at half speed in between. Paul Graham’s “Maker’s Schedule, Manager’s Schedule” is the same point from a different angle. A single interruption inside a maker block does not cost you an hour, it costs the whole afternoon, because the depth that was building takes time to rebuild and the rebuild rarely happens in the same day. The triage layer protects the block before the block is broken, by deciding at the door whether the ask is worth the spillover or not, and writing back accordingly. Sam Altman’s framing is useful as a check, you can have anything if you are willing to pay the price for it, and saying no is the price of focus. The triage layer is what makes the price affordable to pay every day rather than once a quarter when the burnout shows up.
Where will you actually meet it in your week?
You will meet it first in the inbox, the most common entry point and the place the tooling has matured fastest. Wade Foster at Zapier walked through his workflow on Lenny’s Newsletter, the model reads inbound, classifies against his criteria, and drafts replies he then approves. Tools like SaneBox and Superhuman package the same idea, and the DIY pattern with ChatGPT or Claude is well-documented for founders happy to wire it up themselves.
The second place you will meet it is in DMs and warm intros, where the asks arrive with a relationship attached and the cost of a clumsy no is higher. The triage layer here is softer, it tags rather than rejects, and it surfaces context, who introduced this person, what we have agreed about cold pitches this quarter, what the standard polite-no template sounds like. The third place is calendar requests routed through a scheduling link, where the triage layer sits behind the link rather than in front of the inbox. None of these is exotic. The whole stack is two years past early adopter and the operator class has normalised it.
When to ask the filter and when to ignore it
You ask the filter on every speculative pitch, every casual coffee with no stated thesis, every “quick chat” that turns out to be longer, and every ask from a stranger that opens with three paragraphs about themselves. The default is to let the layer do its work, and the criteria do the deciding. Naval Ravikant’s heuristic is the right one for ambiguous cases, if you cannot decide, the answer is no.
You ignore the filter for the exception class, and the exception class is the bit that earns the discipline its trust. Warm intros from people whose judgement you have already proved bypass the filter on purpose. The “interesting human, no thesis” coffee, occasionally, deserves twenty minutes anyway, because optionality and serendipity are real and a founder who filters them out entirely ends up isolated. The filter is a forcing function for clarity, it is not a substitute for human warmth. The discipline is calm, not curt. If a no goes out and the relationship matters, you write a longer reply by hand and the layer steps aside.
Related concepts and where to go next
The triage layer sits inside a wider Eliminate-quadrant practice, alongside auditing your week for work that earns nothing and killing meetings that never should have been booked. The cluster opens at AI for your own work, which frames the four quadrants Eliminate, Automate, Delegate, and Do. The framework explainer at the EAD-Do framework recast for AI walks the rule deeper.
The criteria document the triage layer reads is the same artefact you brief a delegated assistant against, and the same shape as the contractor brief that sits at the heart of the Delegate quadrant. The piece on briefing AI like a contractor goes through how to write one without it sliding into a wishlist. Treat the triage layer as the front door, the contractor brief as the workshop, and the rest of the week looks different inside a quarter.
If you want a second pair of eyes on the criteria you should be triaging against this quarter, book a conversation.
