Abstract claims about AI programming quality are easy to make and hard to evaluate. The more useful exercise is looking at a specific example: a defined trainer, a defined client, and the plan the AI generates for that combination. Here's one.
The trainer profile
The trainer in this example has twelve years of experience working primarily with general population clients, with a focus on strength development and body composition. Their programming philosophy prioritizes compound movement patterns, uses linear progression for beginners and undulating periodization for intermediates, and keeps accessory work targeted — two to three movements per session rather than long lists of supplemental exercises. They prefer trap bar deadlifts over conventional for most general population clients, use RPE-based loading from week four onwards in a beginner block, and structure sessions as warm-up, primary compound movement, secondary compound movement, and two to three accessories.
The client profile
The client is a 34-year-old woman, new to structured resistance training. She has trained inconsistently for two years — mostly cardio with occasional machine-based strength work. She has no significant injury history but has reported mild knee discomfort during high-volume squat patterns in initial assessment sessions. Her goal is to build overall strength and improve body composition. She trains three days per week with access to a full commercial gym. Sessions need to be completed in 60 minutes.
The generated plan: Week 1
Session A begins with a barbell back squat as the primary movement, programmed at three sets of five at a conservative load — roughly 60% of estimated one-rep maximum — with the rationale that technical proficiency is the primary objective in week one. The secondary movement is a dumbbell Romanian deadlift, three sets of eight, chosen over a conventional deadlift to reinforce hip hinge mechanics without the technical complexity of a barbell pull at this stage. Accessories are a seated cable row, three sets of ten, and a dumbbell incline press, three sets of ten. Total working sets: twelve.
Session B is an upper-body focus day. Primary movement is a dumbbell bench press, three sets of eight. Secondary movement is a trap bar deadlift, three sets of five — introduced here rather than session A because the trainer's profile indicates a preference for this implement with general population clients, and early exposure at lower intensity familiarizes the client with the movement before it becomes a primary loading tool. Accessories are a lat pulldown, three sets of ten, and a cable face pull, three sets of fifteen. Total working sets: twelve.
Session C returns to lower body emphasis. Primary movement is a goblet squat, three sets of ten — chosen over the barbell squat to manage cumulative knee stress in week one given the client's reported discomfort, while keeping the squat pattern as a training priority. Secondary movement is a barbell Romanian deadlift, three sets of eight, progressing from the dumbbell version in session A. Accessories are a dumbbell split squat, two sets of ten each side, and a plank variation, three sets of thirty seconds. Total working sets: eleven.
Programming rationale
The week one volume is intentionally conservative — total working sets across the week sit at the lower end of the minimum effective volume range for a beginner. This reflects two factors: the client's training age and the need to establish movement quality before increasing load. The AI generated this loading approach because the trainer's profile indicates that they front-load technical development in beginner blocks and introduce RPE-based loading from week four, not immediately.
The knee management consideration is visible in the session structure: the barbell back squat appears in session A at low volume and low intensity, and session C substitutes the goblet squat variation. This reflects the client's knee discomfort flag in her profile — the AI distributed squat volume in a way that keeps the pattern trained while managing cumulative stress in the first week.
The exercise selection across all three sessions reflects the trainer's stated preferences: trap bar over conventional for the deadlift pattern, compound-primary structure with targeted accessories, and no circuit or superset format — straight sets throughout, consistent with the trainer's documented approach for beginner clients.
What the trainer would review
A trainer reviewing this output would likely confirm the loading scheme for session A's squat, verify that the dumbbell RDL in session A is appropriately loaded for week one, and check that the trap bar deadlift in session B is introduced at a load that prioritizes mechanics. These are the normal checks a programmer would apply to any first week of a new client block — the AI's output doesn't change the review process, it just means the programmer is reviewing rather than constructing.