Weekly Check-In Triage: 3 Questions That Beat Better AI

Weekly Check-In Triage: 3 Questions That Beat Better AI

Why decision quality in coaching starts with appetite, glucose, and timing—not more chat.

Justin Harris told a client trying retatrutide that 2 mg on Friday produced “no appetite whatsoever” through Saturday and Sunday, with more fatigue than normal; his response was to use it as a practical appetite lever, not a moral victory. The mechanism is simple appetite suppression, and that matters because weekly check-ins are really triage: if you cannot decide whether to push food, hold steady, or pull back, the “AI coach” is just generating words. My thesis is blunt and falsifiable: better AI coaching will not come from more automation, but from sharper weekly triage rules that turn three inputs—appetite, glucose response, and recovery status—into one decision.

The most useful thing a coach can do in a weekly check-in is not narrate the athlete’s week. It is decide. That sounds obvious until you watch how much time gets burned reviewing details that do not change the next seven days. If the check-in does not end in an action, it was probably a report, not coaching.

The KB gives us a clean example of why. In the Rory Lazowski exchange, Justin did not treat the retatrutide response as abstract data. He noticed the appetite drop, the fatigue, and the fact that more food was likely coming. His call was not “keep stacking tools” or “trust the process.” It was: if calories are going up, pause or reduce the dose; if the athlete is leaning out, lean out while it is easy. That is triage in its plainest form: identify the dominant constraint, then pick the next move that preserves future options.

That logic matters because weekly check-ins are usually dominated by noisy middle cases. Not the obvious disaster. Not the obvious win. The middle case where body weight is moving, hunger is weird, training feels acceptable, and the athlete wants reassurance. In those cases, a coach who cannot rank the problem list will overfit to the loudest complaint. A coach who can rank the problem list can make one clean change and wait for the next signal.

Justin’s view of off-season nutrition says the same thing from another angle. In the podcast excerpt, he frames the off-season as teaching the body to digest and assimilate a massive amount of clean food, then getting to a place where the athlete can basically eat whatever they want without gaining weight—moderately, not in two days. The mechanism is digestive and metabolic tolerance, and the intended outcome is not just growth now, but better contest prep later. The practical point for check-ins is that “can they handle the plan?” is not a side issue. It is the plan. If the athlete cannot assimilate the food, the next tweak is not philosophical; it is operational.

This is where AI coaching either helps or gets in the way. AI is good at pattern storage and repetition. It is weaker at judgment unless the judgment is explicitly structured. A weekly check-in triage system should force the following sequence:

1. What is the limiting factor right now?
2. What changed since last week that actually moved that factor?
3. What is the smallest adjustment that tests the next step without creating new noise?

That sequence is boring. It is also what prevents bad decisions.

Joe Webb’s check-in illustrates why. He reported that the same insulin dose on a high day was dipping blood sugar more than before, forcing him to bring meals closer together; he planned to reduce the dose further on the next high day. Justin’s reply, as shown in the exchange, is truncated in the source, but the setup itself already contains the decision-quality lesson: the meaningful observation is not “high day went fine” or “I felt okay.” It is that insulin sensitivity changed enough to alter meal timing. That is the signal. The adjustment is to respond to the signal, not preserve the old dosage out of habit.

That is exactly how weekly check-ins should work across the board. One metric changes, one plan changes. The coach should not wait for a crisis to act, and should not make three changes because two things looked off. The point is not maximal intervention. It is controlled intervention.

Now compare that to the kind of AI coaching people expect: long summaries, motivational framing, and generic next steps. Those can be pleasant, but they do not improve decision quality unless they sharpen a threshold. If a client says appetite is down, the useful follow-up is not “how do you feel about that?” The useful follow-up is: is appetite suppression helping the current phase or blocking it? If the answer is blocking it, the triage answer is to reduce or pause the appetite-lowering tool. If the answer is helping, keep the tool and use the window. Same with blood sugar. Same with recovery. Same with food tolerance.

Justin’s nutrition casework also shows that not every detail deserves equal weight. In the fruit case from the nutrition module, he says he would be surprised if fruit choice alone made a noticeable difference over a year, because most results come from nailing macros; the “last few percent” live in the details. But he still puts guardrails around where fruit belongs, and he adjusts that advice by day type. That is the right model for check-in triage: macro-level compliance first, detail-level optimization second. Do not let the tail wag the dog.

For coaches using AI, this creates a useful test. If your weekly check-in workflow cannot answer these three questions in under a minute—what is changing, what matters most, what gets adjusted now—you do not have a coaching system. You have a conversation machine. And if you cannot reliably tell the difference between a signal and a story, the AI will only help you sound more confident while making the same bad decisions faster.

The strongest coaching habit in the KB is not “be more data-driven.” It is more specific: spot the constraint, make one change, and watch the response. Retatrutide lowering appetite in a gaining phase, insulin sensitivity changing on a high day, food tolerance becoming a long-term adaptation target—these are all triage problems, not content problems. AI can summarize them. It cannot rescue weak judgment.

So the sharp take is this: weekly check-in triage is the highest-value use of AI in coaching only when the system is built to improve decision quality, not to replace it. If the check-in does not change the next action, it did not help.

Sources Used

• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.md
• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.json
• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.json
• modules/03-knowledge/kahunas-coaching-deep-nutrition.md
• raw/Justin_TT1.txt