# Integrated LLMs vs. Algorithmic Coaching: Utility, Guardrails, and Safety **URL:** \ **Category:** Training\ **Created:** [9 October 2026 19:00 UTC](https://forum.intervals.icu/t/integrated-llms-vs-algorithmic-coaching-utility-guardrails-and-safety/132885 "2026-10-09T19:00:29Z")\ **Posts on this page:** 1\ **Showing post:** 9
**Author:** ![Vincent\_Barbaix](https://forum.intervals.icu/user_avatar/forum.intervals.icu/vincent_barbaix/32/92934_2.png) [@Vincent\_Barbaix](https://forum.intervals.icu/u/Vincent_Barbaix)\ **Post date:** [16 October 2026 14:35 UTC](https://forum.intervals.icu/t/integrated-llms-vs-algorithmic-coaching-utility-guardrails-and-safety/132885/9 "2026-10-16T14:35:38Z")
Coach here, with a disclosure: Nick and I build a deterministic (non-LLM) planner on top of Intervals.icu, so I’m not neutral. But on @Gabriel_Vargas’s bad-data question, what we learned is that the model matters less than what happens when data is missing or wrong: - **Show where every number comes from.** A threshold measured in a lab, one calculated from a field test and one estimated from eFTP should not look the same on screen. - **Don’t fill gaps silently.** If a weight or a test is missing, say so instead of plugging in a default. - **Keep a person between the plan and the calendar.** The coach or the athlete reviews before anything is published. - **Same inputs, same output.** Then you can trace a strange week back to the input that caused it. Bad files (heart-rate spikes, power dropouts, duplicates) are still the hard part. We flag activities that arrive without detail, but bad data in still hurts. I agree with WindWarrior: the principles are simple, and tools should make them easier to follow, not replace judgement. If you want to see how we handle it: [our beta thread](https://forum.intervals.icu/t/132881). Grts, Vince & Nick --- _[View the full topic](https://forum.intervals.icu/t/integrated-llms-vs-algorithmic-coaching-utility-guardrails-and-safety/132885)._