The best nutrition coaching in the world used to require two things: a qualified sports dietitian and money. A lot of both. Elite coaches charge $200–400 per hour and are booked months out. For recreational athletes, even serious ones, personalised nutrition guidance has always been out of reach.
AI is changing that. Not by replacing good nutritionists, but by making intelligent, context-aware nutrition guidance available at the moments it's actually useful: 11pm when you're logging your third meal, Sunday morning when you're meal prepping for the week, or 30 minutes post-workout when you need a recovery meal recommendation fast.
What makes AI coaching different from a nutrition app
Traditional nutrition apps are calculators. They take your profile data, apply a formula (Harris-Benedict, Mifflin-St Jeor), and output a target. They don't adapt. They don't reason. They don't understand context.
Ask MyFitnessPal what to eat after a hard 10km run and it will show you a food search bar. Ask the Jonno Agent and it will:
- Check your Strava data and see you ran at 5:15/km average, burning 650 kcal
- Calculate how much protein and carbohydrate you have remaining for the day
- Factor in your food preferences from your history (you logged Thai food 6 times last month)
- Search nearby restaurants on Uber Eats for high-protein recovery options
- Build you a cart: 45g protein, 80g carbs, under 700 kcal
This is the difference between a tool that stores data and one that reasons about it.
The three gaps AI fills
1. Availability
Human coaches are available during business hours. Athletes train at 6am, race on weekends, and make food decisions at midnight. AI is available at the exact moment the decision happens. The value of nutrition guidance scales massively when it's delivered at decision-point, not in a scheduled weekly check-in.
2. Contextual memory
A good coach remembers your history: the foods that gave you GI issues, the protein sources you prefer, the weeks you over-reached and crashed. AI systems that integrate your training logs and food history can build the same contextual model, and recall it instantly.
Jonno's agent is given your training history, macro targets, and food preferences with every interaction. It doesn't start from scratch each conversation.
3. Cost and access
This one is obvious but important. Sports nutrition coaching for recreational athletes has historically been a luxury. AI democratises that access. Serious athletes earning median incomes can now have the same quality of nutrition guidance previously available only to professional athletes with sports science support teams.
Where AI nutrition still requires human judgement
We should be honest about what AI gets wrong.
AI nutrition tools are not medical devices and should not replace clinical nutrition assessment for people with disordered eating, specific medical conditions, or complex dietary requirements. The Jonno Agent operates within a well-defined domain: helping athletes hit macro targets to support training and recovery. It is not a dietitian, and it doesn't pretend to be.
Where precision matters most, competitive professional athletes, people with medical dietary requirements, complex eating disorder recovery, human expertise remains essential. AI accelerates good habits. It doesn't replace professional clinical care.
The model behind Jonno
The Jonno Agent is built on Anthropic's Claude. Anthropic is one of the leading AI safety companies in the world, and Claude is specifically designed to be honest, helpful, and harmless. That matters for nutrition advice, where overconfidence or incorrect recommendations can genuinely harm performance or health.
Claude is given a structured context at the start of every conversation: your profile, your day's macros, your recent training, and your preferences. It reasons within that context rather than generating generic advice.
The result is guidance that feels like it comes from a coach who knows you, because it's built on data that does.
What this looks like in practice
Athletes using AI nutrition coaching consistently report two outcomes: they hit their protein targets more consistently, and they spend less cognitive energy on food decisions. That cognitive load reduction, not having to think through "what should I eat" from scratch every meal, turns out to be a meaningful performance advantage in itself.
Decision fatigue is real. Athletes already make hundreds of micro-decisions in a training session. Having a system that eliminates the nutrition decision is, functionally, a performance enhancement.