Weekly Check-In Triage: 3 Decisions That Beat AI Guessing

Weekly Check-In Triage: 3 Decisions That Beat AI Guessing

Decision quality rises when coaches triage changes by urgency, leverage, and reversibility instead of treating every check-in like a full rebuild.

Justin Harris’s coaching pattern in the Rory Lazowski exchange is blunt: when retatrutide crushed appetite with 2 mg on Friday, he didn’t frame it as a universal win or loss; he said to keep moving leaner while it was easy, then reassess how it behaves when food rises. Mechanism in one phrase: appetite suppression changes the food-intake constraint. That is the right kind of mechanism for weekly check-in triage because the thesis is simple and falsifiable: coaches make better decisions when a check-in is used to sort problems by leverage and timing, not to react to every variable with equal urgency.

Check-ins are not status reports; they are triage

Most weekly check-ins fail for the same reason most generic AI coaching systems fail: they flatten very different situations into one response. A slight drop in appetite, a small insulin-sensitive change around a high day, and a real pattern of worsening blood glucose are not the same class of problem. They do not deserve the same intervention speed, and they definitely do not deserve the same size of change.

The strongest coaching examples in the KB all point the same direction. When Joe Webb reported that the same insulin dose on a high day was now dipping his blood sugar more than usual, Justin did not turn it into a philosophical debate. He told him to reduce the dose further on the next high day. That is triage: identify the variable that changed, judge the near-term risk, and adjust the smallest lever that solves the problem.

That sounds obvious until you watch coaches miss it. They either overreact and change food, training, and drugs at once, or they underreact and wait for a problem to become a mess before calling it “trend data.” Weekly check-ins should do neither. They should answer three questions in order: what changed, how dangerous is it, and what is the smallest reversible move that restores control.

Decision quality starts with classifying the problem

Justin’s comments in the sources separate problems into at least three categories.

First, there are leverage problems. In the nutrition case from the deep nutrition module, he says the last few percent lives in the details, but the vast majority of results come from nailing macros. That is a useful filter for weekly check-ins: if the athlete is still missing the basics, the correct response is not advanced optimization.

Second, there are constraint problems. In Rory’s retatrutide thread, appetite suppression was immediate and obvious after a 2 mg dose. Justin’s reaction was not to worship the appetite effect; it was to note that appetite forcing lower may not be a good fit if food intake needs to rise. That is exactly how a coach should think when a variable changes the athlete’s ability to execute the plan. If the constraint is food volume, appetite changes matter more than a minor food swap ever will.

Third, there are risk problems. In the Joe Webb example, the issue was not “interesting data”; it was that a fixed insulin dose no longer behaved the same on a high day. That is the kind of check-in detail that deserves immediate action because it affects whether the next meal needs to come 30 minutes sooner or not. The operational question is not whether the athlete is “doing well overall.” The operational question is whether today’s dose still matches today’s response.

A coach using AI well should preserve these distinctions. If the check-in mechanism treats appetite changes, glucose changes, and macro adherence as interchangeable notes, it will produce mush.

The weekly triage order should be: safety, execution, leverage

The KB gives a practical hierarchy even when it does not spell it out as a framework.

1. Safety or clear dysfunction comes first.
2. Execution blockers come second.
3. Fine-tuning comes last.

That order matters because coaches waste the most time on the least important tier. They obsess over small nutrition choices before they have established whether the athlete can actually get the food in, recover, and follow the plan consistently.

The nutrition module makes this point quietly but strongly. Fruit versus blueberries, pasta once a day, pre/post-workout carb sourcing: Justin’s answer is that the macro structure matters far more than the exact food choice, and that the last few percent lives in the details. In weekly check-in triage, that means food-level tweaks should only become the focus after the larger execution patterns are stable. If the athlete is still fighting appetite, GI tolerance, meal timing, or total intake, the check-in should not spend its entire budget on carb-source perfection.

That is also why the Rory exchange is useful. Retatrutide lowered appetite so much that it could become a prep tool, but Justin was cautious about assuming it would be helpful for gaining. That is a decision-quality move: do not let a useful response in one phase get mechanically copied into another phase where the constraint is different.

The best check-in question is not “what happened?”

It is “what needs to change because of what happened?”

That single shift improves AI coaching more than adding more natural language or more data fields. The athlete can report six things: hunger, sleep, scale trend, pumps, training performance, and digestion. The coach still has to decide which of those items is signal and which is background noise.

The KB examples show how Justin handles that sorting:

• When appetite is dramatically suppressed at 2 mg retatrutide, he notices the mechanism and treats it as a variable that affects future food intake.
• When insulin sensitivity appears improved on a high day, he adjusts the dose instead of talking around the issue.
• When macro execution is the real bottleneck, he focuses on the food framework rather than over-optimizing fruit choice.

That is the anatomy of good check-in triage. The coach reads the report, identifies the limiting factor, and selects the smallest change that preserves momentum.

What AI should do in weekly check-ins

AI is useful here if it does less, not more.

A good weekly check-in system should:

• separate red flags from normal adaptation,
• distinguish a constraint change from a preference change,
• recommend one adjustment at a time when possible,
• and avoid rewriting the whole plan when a single lever moved.

It should not turn a small appetite shift into a nutrition ideology. It should not turn a high-day insulin adjustment into a full metabolic theory. It should not treat a fruit question as if the client is already maxed out on execution.

That is the practical lesson from the KB: decision quality comes from ranking the issue before recommending the fix. Coaches who skip ranking end up with noisy, bloated check-ins and unstable plans. Coaches who triage well make smaller, cleaner changes and keep the athlete moving.

The falsifiable test

Here is the test this thesis makes: if a coach’s weekly check-ins consistently lead to fewer unnecessary plan overhauls, faster correction of clear execution problems, and tighter responses to changing constraints, then triage improved decision quality. If the coach is still making large changes from small signals, the system is not triage — it is reaction.

That matters for AI fitness coaching because AI is especially good at collecting information and especially bad at deciding what deserves attention. The best use of it is not more commentary. It is a triage layer that forces the coach to answer: does this change the plan now, later, or not at all?

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/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.json
• modules/03-knowledge/kahunas-coaching-deep-nutrition.md
• modules/08-voice/kahunas-coaching-deep-voice.md

Sources Used

• /Users/justinharris/TroponinIQ/kb/supertrop/modules/08-voice/kahunas-coaching-deep-voice.md
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.md
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.json
• /Users/justinharris/TroponinIQ/kb/supertrop/modules/03-knowledge/kahunas-coaching-deep-nutrition.md