Training-Plan Copilot AI Agent

What this agent does

Personalised to the schedule, not to a predicted response

Ingest the athlete's schedule, equipment, training history and preferences from supported coaching or club systems.

Normalise exercise names, units and session formats, and keep each input tied to its source.

Reason

Draft the block — sessions, exercises, sets, load targets and the deload week.

Apply the progression, volume and substitution rules the club configured.

Personalise on schedule, equipment, history and preference — never on predicted response.

Decide

Flag anything in the intake that belongs to screening, not programming.

Route the drafted block to the coach of record for sign-off.

Out

Retain the inputs, the draft, the coach's edits and the sign-off against the athlete record.

Execute write actions only inside the approval boundaries agreed during implementation.

→ Product statement

The agent proposes the block; the coach of record decides what is trained. Progression is configured, bounded and reviewable — never silent.

Example workflow

One block, intake to sign-off

AgentHuman

1Athlete intake receivedScreening form, coaching notes, training history or club-system record

2Inputs assembledSchedule, equipment, history, preferences and screening flags, each with its named source

3Block draftedSessions, exercises, progression, deload and confidence

4Controls appliedScreening-flag checks, progression and volume ceilings, substitution-library rules and confidence threshold

No human action required

Stages 1 to 4 run unaided, and nothing reaches an athlete at any of them — the agent is drafting, and the coach's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold

High confidence

Goes to the coach of record to sign.

Low confidence

Adds a senior-coach read first.

Coach sign-off

The block is held with its inputs, its flagged lines and the confidence.

Approve · Edit · Send to senior review

Signed — released to the athlete▼

6Coaching systems updatedOnly where write access and approval policy allow it

7Outcome evaluatedEdit distance, flagged-intake outcomes, completion and post-session corrections

Edits

Every coach edit is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items

Deciding whether a member is fit to train.

Progressing a plan past an open screening flag.

Changing a return-from-injury or post-surgical plan.

Programming for a member held for clearance.

Automation boundaryAgent acts unaided

✓Draft the block from the schedule, equipment and history on file.

✓Substitute an exercise inside the configured library.

✓Write the session notes back to the athlete record.

✓Flag intake that needs screening, and hold the block for the coach.

Any write happens inside the boundaries agreed at implementation, never ahead of sign-off.

Return-to-play after a suspected concussion.

Reading a reported symptom as a training input.

Pregnancy, paediatric or cardiac-history programming.

Changes to progression ceilings or approval rules.

Example output

One session, annotated

Whatever the agent drafts stays attached to the inputs behind it.

Training-plan output · single sessionIllustrative example

Athlete

Drafted session

Working load

Source of record

Confidence

Progression

Returning member

Lower-body strength, three compound lifts then accessory work

Last logged set

Logged training history

93%

Within the configured ceiling

As receivedTaken from the training history and the intake form on file — the agent wrote none of this side.

Inputs used Logged sessions Coach intake form Equipment and schedule

Why this progressionIt moves inside the ceiling a coach set, not toward a predicted response.

ActionApproveEditSend to senior review

What the score decidesBelow the configured threshold the block picks up a senior-coach read before the coach signs.

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value landsValue 01 – 04

Every blockFrom the intake record

03Drafting

Draft from what is on file

Draw on the schedule, equipment and history on file, inside the ceilings a coach configured.

01Approved path

Draft the block, not the verdict

Routine blocks, progressions and deloads arrive already written.

02Human review

Send review to the risky plans

Screening flags and low-confidence blocks are marked, so the coach's read starts where the risk sits.

04Build an evidence trail

The plan, the inputs behind it and the coach who signed it stay attached to the athlete.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

Club managementMindbody · ABC Glofox
Xplor Recreation · Perfect Gym

Coaching platformsTrainHeroic · TeamBuildr
TrueCoach · Trainerize

Athlete recordsIntake · screening
Session logs · notes

Agent

Training-plan copilot

Reads the inputs
Drafts the block
Holds for sign-off

Scheduling and commsCalendars · email
Member app · portal

Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six layers between the model and the athlete

The layers nest. What each one misses is named in the map below.

L6 · Outermost — last line of defenceInward → L1 · closest to the model

L6Rollback / safe modePull the copilot back to the standing template when evaluation or production signals degrade.Roll back

L5Version monitoringTrack model, prompt, progression-rule and library-configuration changes.Track

L4TraceabilityRecord the inputs, the draft, the flags, the edits and the sign-off time.Record

L3Coach sign-offHold blocks for the coach of record; it governs release, not whether a signed block is right.Gate

L2Policy guardrailsTest blocks against the configured ceilings, volume limits and substitution library; a failure returns the block.Restrict

L1Confidence thresholdsRoute low-confidence blocks to a senior-coach read before the coach signs.Require review

Model coreBlock produced — sessions, progression, deload, flagged intake and confidence

L1 – L2Test whether a block may stand

L3Puts the release in a coach's hands

L4 – L5Show what the session was built from

L6Reverts to the standing template when signals degrade

How Nestack evaluates it

Evaluate the drafting workflow — not only the finished block.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the session the athlete trains

Depth of coverage ▼

E1Final-output evaluationDid the block hold inside the configured ceilings and equipment?

E2Step-level evaluationDid the agent use the right history, screening record and rule set?

E3Tool evaluationDid it read the right athlete record and write the correct field?

E4Confidence calibrationDo low-confidence blocks actually attract more coach edits?

E5Slice evaluationHow does performance change across specific training groups?

E6Business outcomeHow many blocks needed a coach edit or a correction after the session?

Floor — the outcome the club answers for

Failure modes

Where each failure originates in the agent

Seven failure modes, placed at the stage each one originates.

Agent lifecycleDirection of processing →

01 · Retrieval1 mode

TG-03

Stale training history

A layoff is missing, so the athlete reads as continuously trained.

Stage gathersSchedule, equipment, history and screening flags

02 · Reasoning2 modes

TG-04

Response-based personalisation

The block is shaped around a predicted response rather than the record.

TG-06

Ceiling breached on substitution

A swapped exercise carries more volume than the rule allows.

Stage proposesThe block, the progression applied and confidence

03 · Tool / write2 modes

TG-02

Flagged intake programmed anyway

A block is drafted past a screening flag no one cleared.

TG-05

Symptom read as a training input

A reported symptom is answered with a deload instead of a person.

Stage writesOnly where write access and approval policy allow it

04 · Output1 mode

TG-01

Unfamiliar volume spike

A block adds unfamiliar work faster than the ceiling intends.

Stage returnsThe block the coach signs and the athlete trains

05 · Change / Version1 mode

TG-07

Silent progression drift

A model or rule change widens what the agent will progress.

Stage tracksModel, prompt, progression rules and library config

Sev-1 · a block trains past a screening flag Sev-2 · a wrong progression reaches an athlete Sev-3 · input degrades, block routes to review

See the complete Sports & Fitness failure-mode catalogue →

Affected slices

The average holds while one group carries it

Base rates differ by training group, so one aggregate rate describes none of them well. Nestack reports the coach-edit rate by slice, and names the groups that carry it.

Slice performance — reported separately, not only in aggregateIllustrative example

Slice	Failure rate	Lift	Status
Returning after a layoff	5.3%	3.9×	Review
First-block novices	3.9%	2.9×	Review
Minimal-equipment setups	2.2%	1.6×	Watch
Members in continuous training	0.9%	0.7×	Normal

Bar: coach-edit-rate lift vs. continuous-training baseline · scale 0–4.0× · tick marks the 2.0× review threshold 2 of 4 slices over threshold

Evidence-linked improvement

The cycle closes on a test, not a note

A cycle is done when the miss has become a test the next release has to survive. That suite is what the next block written by the copilot is measured against.

Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn

01Detect

Coach-edit rate rises in one training group.

02Diagnose

Within the week it went out, a coach reads the block and its inputs until one of them accounts for it.

03Improve

Version-stamp the change and attach the plans that exposed it.

04Verify

The affected cases run again, and a fail stops the release.

05Learn

It becomes a standing test, and the progression guardrails change with it.

Learn → DetectThe return edge. The next detection runs against a suite that is one case longer.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, drafting workflow, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6

01Coaching workflow discovery and boundary definition.

02Club and coaching-platform assessment.

03Progression, volume and substitution-library mapping.

04Athlete-record ingestion and normalisation.

05Drafting logic and source-input binding.

06Confidence scoring and flagged-intake routing.

07Coach sign-off workflow.

08Coaching and club-system integration.

09Progression regression cases.

10Guardrails and progression controls.

11Plan-trail instrumentation.

12Deployment, documentation and Agent Care handover.

12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tier PilotOne club, one squad ProductionProduction coaching systems AdvancedMultiple sites / brands

Introduced at Pilot

Drafting to your rules and library✓✓✓

Coach sign-off✓✓✓

Plan-quality baseline✓✓✓

Introduced at Production

Reporting by training group—✓✓

Sign-off workflow in your systems—✓✓

Approved write-back—✓✓

Coaching-platform integration—✓✓

Introduced at Advanced

Multi-sport progression rules——✓

Multi-stage coaching approvals——✓

High athlete volume——✓

Multi-squad programme controls——✓

Build price From $5,000 From $8,000 Custom quote

Final build priceConfirmed after discovery based on integrations, workflow complexity, athlete volume, approval controls and deployment requirements.

Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

You bringWe build with it

01Your intake and screening-form structure → Athlete-record ingestion and input mappingWeek 1

02Representative programmes you already write → Drafting baseline, template extraction and input bindingWeek 2

03Your progression, volume and substitution rules → Progression, volume and substitution-library mappingWeek 1

04Access to relevant APIs, feeds or exports → Club and coaching-platform assessment, then integration setupWeek 2

05Plans that should not have gone out → Progression cases and the evaluation suiteWeek 4

06What must reach a coach before an athlete trains it → Confidence scoring, flag routing, guardrails and progression controlsWeek 3

07Named coaches of record to sign blocks → Coach sign-off workflow, then pilot and production validationWeeks 5–6

Nothing else is required Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Phases are drawn over the weeks they actually occupy, which is why week 5 doubles up rather than padding.

Phase W1W2W3W4W5W6

Discovery W1

Build W2 – W3

Evaluate W4 – W5

Pilot & Launch W5 – W6

Week focus W1Coaching workflow discovery, rule mapping and the automation boundary W2Source integration and the drafting baseline W3Drafting workflow, confidence logic and progression controls W4Evaluation suite, screening-flag checks and failure-mode testing W5Coaching-platform integration, pilot blocks and targeted corrections W6One training block written under the coaching staff, then handover

Reading the bandA bar covers only the weeks its work is named in. Both bars in week 5 are real work running together.

At the end of W6The last checks clear on real blocks and monitoring moves to Agent Care.

DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Sports & Fitness AI agent

Build a training-plan copilot around the coaches who sign the block.

Show us your intake form, your progression rules and who signs. Get it wrong and an athlete trains a block no coach cleared — so we set the boundary before we write anything.

Discuss this agent →

Nestack Agents · Training-plan copilotAGT-SF-02 · Agent Care available after launch