Fitness AI agent · Lead conversion

Lead-Conversion & Sales Follow-Up AI Agent

Answer inbound enquiries, book tours and trials, and follow up only where the consent record allows it — then hand the lead to the named sales rep, who decides what is sold.

Typical delivery
4–6 weeks
Deployment
Your stack
Rep decides
Consent-gated
After launch
Agent Care

What this agent does

Automates the reply, not the permission

In

  1. 01

    Ingesting enquiries, consent records and provenance from supported CRM, phone and messaging channels.

  2. 02

    Normalising channel identifiers and consent wording, and carrying each forward with its source.

Reason

  1. 03

    Answering inbound enquiries from the timetable, the pricing card and club information.

  2. 04

    Treating any AI voice call as an artificial-voice call that consent governs.

  3. 05

    Checking where each lead's consent came from, and suppressing outreach it cannot support.

Decide

  1. 06

    Flagging any consent record that is missing, stale or revoked in channel.

  2. 07

    Routing each enquiry to the named sales rep for handover.

Out

  1. 08

    Retaining the consent relied on, the message sent, the opt-out and the handover per lead.

  2. 09

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

Product statement

The agent answers and books; the named sales rep decides what is sold, and the operator stays responsible for the consent.

Example workflow

One enquiry, arrival to handover

  1. 1Enquiry receivedWeb form, class-booking page, phone call, text or social message
  2. 2Consent checkedSource, wording and date of the consent on record, and any revocation in any channel
  3. 3Reply preparedAnswer, tour and trial options, flagged lines and confidence
  4. 4Controls appliedConsent-record checks, DNC and suppression scrubs, local-time windows, frequency caps and confidence threshold

No human action required

Stages 1 to 4 run unaided, and no outbound contact is made at any of them — the agent is answering, and the rep's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold
High confidence

Goes to the named sales rep to take over.

Low confidence

Adds a consent review first.

Sales handover

The lead is held with its consent record, its flagged lines and the confidence.

Accept · Amend · Send to consent review

Accepted — the rep has the lead

  1. 6Sales systems updatedOnly where write access and approval policy allow it
  2. 7Outcome evaluatedReply accuracy, opt-out handling, consent exceptions and corrections made after sending

Amendments

Each rep amendment is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required

8 items

  • Outbound contact where no consent record exists.
    • Contacting a lead who has revoked, in any channel.
    • Quoting a price or term not on the published card.
    • Enrolling a lead in a membership or a recurring charge.
  • Automation boundary

    Agent acts unaided

    • ✓Reply to an inbound enquiry from the club's published information.
    • ✓Book a tour or a trial visit in the slots held open.
    • ✓Answer documented questions on hours, classes and kit.
    • ✓Check the consent record before any outbound step, and hold the rest.

    Any outbound step runs inside the boundaries agreed at implementation, behind the consent record.

    • AI-voice calls without documented written consent.
    • Treating a purchased lead's consent as verified.
    • Serving statutory contract or health-club notices.
  • Changes to consent, window or handover rules.

Example output

One enquiry, annotated

Everything the agent sends is attached to the consent it was sent under.

Lead-response output · single enquiry

Illustrative example
Lead
Timetable enquiry
Drafted reply
Answered the class-timetable question and offered two tour slots
Consent on file
Written, web form
Source of record
Enquiry form on file
Confidence
91%
Channel
Reply in the channel it arrived on

As received

Taken from the enquiry and the consent record it arrived with — nothing here is written by the agent.

Consent evidence

  • Web-form consent capture
  • Channel and timestamp
  • Opt-out history

Why this reply

It answers what was asked and books inside the consent — the line the rep weighs.

What the score decides

Below the configured threshold the reply picks up a consent review before it reaches the rep.

Action

AcceptAmendSend to consent review

Value

Where AI adds value

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

Where the value lands

Value 01 – 04

Every enquiry

From the enquiry channel

03Response

Reply from the record

Draw on the club's published information and the consent captured with the enquiry.

01Approved path

Answer fast, inside the consent

Routine questions about hours, classes and tours are answered as they arrive.

02Human review

Stop where the consent stops

Missing, stale or revoked consent suppresses the outreach and puts the lead in front of a rep instead.

04

Build an evidence trail

The enquiry, the consent record it relied on and the rep who took it over stay together.

Integrations

Typical integrations

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

  • Club management systems

    • Mindbody
    • ABC Glofox
    • Zen Planner
    • Club Automation
  • Sales CRM

    • Salesforce
    • HubSpot CRM
    • Pipedrive
    • Zoho CRM
  • Messaging and voice

    • Twilio
    • SMS gateway
    • Voice and IVR routing

Agent

Lead conversion and follow-up

  • Reads the consent
  • Answers the enquiry
  • Hands to the rep
  • Consent and suppression

    • DNC
    • opt-out log
    • Consent-record store
  • Observability & evaluation

    • OpenTelemetry
    • 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 lead

Each layer contains the next. What gets past the set is listed in the map below.

L6 · Outermost — last line of defence

Inward → L1 · closest to the model

L6Rollback / safe modePull the agent back to reply-only when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, consent-rule and calling-window configuration changes.Track
L4TraceabilityRecord the consent relied on, the message sent, the channel and the handover time.Record
L3Sales handoverHold each lead for the named rep; it governs handover, not whether the reply was right.Gate
L2Policy guardrailsTest each message against consent, suppression, window and frequency rules; a failure holds it.Restrict
L1Confidence thresholdsRoute low-confidence replies to a consent review before the rep sees them.Require review
Model coreReply produced — answer, tour and trial options, flagged lines and confidence
L1 – L2
Test whether a message may go
L3
Puts the handover in a rep's hands
L4 – L5
Keep the enquiry and the permission it rested on
L6
Reduces to reply-only when signals degrade

How Nestack evaluates it

Evaluate the follow-up workflow — not only the reply that went out.

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

Surface — the message the lead receives

  1. E1Final-output evaluationDid the reply answer what was asked, from documented information?
  2. E2Step-level evaluationDid the agent read the right lead, the right consent and the right rules?
  3. E3Tool evaluationDid an opt-out suppress the lead in the channel it arrived on?
  4. E4Confidence calibrationDo low-confidence replies actually attract more rep amendments?
  5. E5Slice evaluationHow does performance change across specific lead sources?
  6. E6Business outcomeHow many replies needed a rep amendment or a correction after sending?

Floor — the exposure the operator carries

Failure modes

Where each failure originates in the agent

Seven failure modes, each placed where it first appears.

Agent lifecycle

Direction of processing →

  1. 01 · Retrieval

    1 mode

    • LM-03

      Superseded consent read

      The consent returned is older than the lead's opt-out.

    Stage gathers

    Enquiry, consent record, opt-out list and club facts

  2. 02 · Reasoning

    2 modes

    • LM-04

      Vacatur read as consent

      A removed restriction is read as a removed consent requirement.

    • LM-06

      Opt-out treated as pending

      A revocation is queued as though the duty were not yet in force.

    Stage proposes

    Reply, tour options, flagged lines and confidence

  3. 03 · Tool / write

    2 modes

    • LM-02

      Contact outside the window

      A message goes at a local time the state does not allow.

    • LM-05

      Reply sent to the wrong lead

      An answer lands on another enquiry's thread.

    Stage writes

    Only where write access and approval policy allow it

  4. 04 · Output

    1 mode

    • LM-01

      Term quoted off the card

      A price or term is stated that the published card does not carry.

    Stage returns

    The reply the lead reads and the rep answers for

  5. 05 · Change / Version

    1 mode

    • LM-07

      Silent outreach regression

      A model or rule change widens who the agent will contact.

    Stage tracks

    Model, prompt, consent rules and window config

  • Sev-1 · contact sent outside the consent
  • Sev-2 · a wrong statement reaches the lead
  • Sev-3 · source degrades, lead routes to review
See the complete Fitness failure-mode catalogue

Affected slices

Overall response quality can hide one lead source

Break the amendment rate down by lead source before you read it in total. Over every enquiry it looks small; concentrated in the sources whose provenance is thinnest, it does not.

Slice performance — reported separately, not only in aggregate

Illustrative example
SliceFailure rateLiftLift vs. thresholdStatus
Purchased and aggregated leads5.1%3.5×Review
Shared comparison-form leads3.0%2.1×Review
Reactivated lapsed enquiries2.3%1.6×Watch
Leads from the club's own forms1.3%0.9×Normal

Bar: amendment-rate lift vs. the all-lead baseline · scale 0–4.0× · tick marks the 2.0× review threshold

2 of 4 slices over threshold

Evidence-linked improvement

Each cycle ends with one more case

Nothing closes because it was understood. It closes when the next release has to pass a case, and that suite is what the next reply out is measured against.

Improvement cycle · five stages

Switchback — the path turns at Improve and returns at Learn

  1. 01

    Detect

    Amendments and post-send corrections rise in one lead source.

  2. 02

    Diagnose

    If the consent record was wrong, the reply was wrong before it was written — so both are read.

  3. 03

    Improve

    Changes ship against a version with the conversations that caused them.

  4. 04

    Verify

    A failing case blocks release until it clears.

  5. 05

    Learn

    The case is permanent, and the outreach rules move with it.

  6. Learn → Detect

    The return edge. The next detection is measured against the longer suite.

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, response workflow, evaluation, integration, then production validation and handover.

Workstream

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

  1. 01Enquiry workflow discovery and boundary definition.Week 1
  2. 02CRM, telephony and messaging assessment.Week 1, Week 2
  3. 03Consent, calling-window and state mini-TCPA mapping.Week 1, Week 2
  4. 04Enquiry ingestion and consent normalisation.Week 2
  5. 05Response logic and consent binding.Week 2, Week 3
  6. 06Confidence scoring and handover routing.Week 3
  7. 07Sales handover workflow.Week 3, Week 4
  8. 08CRM and messaging integration.Week 5
  9. 09Consent-path regression cases.Week 4, Week 5
  10. 10Guardrails and outreach controls.Week 4
  11. 11Contact-trail instrumentation.Week 5, Week 6
  12. 12Deployment, documentation and Agent Care handover.Week 6

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 tierPilotOne site, one lead sourceProductionProduction lead volumeAdvancedMultiple sites / brands
Introduced at Pilot
Replies from your published information✓✓✓
Sales-rep handover✓✓✓
Response-quality baseline✓✓✓
Introduced at Production
Reporting by lead source—✓✓
Handover workflow in your CRM—✓✓
Approved write-back to the CRM—✓✓
Sales-CRM integration—✓✓
Introduced at Advanced
Multi-state rule sets——✓
Multi-stage sales approvals——✓
High enquiry volume——✓
Multi-site sales controls——✓
Build priceFrom $5,000From $8,000Custom quote
Final build price
Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from build
Build 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.

  1. 01Your enquiry channels and the consent captured on each

    Enquiry workflow discovery and boundary definition

    Week 1
  2. 02Representative answered enquiries

    Response baseline, source binding and disclosure wording

    Week 2
  3. 03Your published membership card and club information

    Consent, calling-window and state mini-TCPA mapping

    Week 1
  4. 04Access to relevant APIs, feeds or exports

    CRM, telephony and messaging assessment, then integration setup

    Week 2
  5. 05Follow-ups that should not have been sent

    Consent cases and failure-mode testing

    Week 4
  6. 06Where follow-up must hand over to a person

    Confidence scoring, handover routing, guardrails and outreach controls

    Week 3
  7. 07Named sales reps to take leads over

    Sales handover workflow, then pilot and production validation

    Weeks 5–6

Nothing else is required

Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Each band covers the weeks it genuinely occupies, so the fifth week holds two kinds of work.

Phase

W1

W2

W3

W4

W5

W6

  1. DiscoveryW1
  2. BuildW2 – W3
  3. EvaluateW4 – W5
  4. Pilot & LaunchW5 – W6

Week focus

W1Enquiry workflow discovery, consent mapping and the automation boundary

W2Channel integration and the response baseline

W3Response workflow, confidence logic and handover controls

W4Evaluation suite, consent checks and failure-mode testing

W5CRM integration, pilot enquiries and targeted corrections

W6A live enquiry queue worked under the sales team, then Agent Care handover

Reading the band
Each band covers the weeks its work is named in, and no others. The week 5 overlap is real, not padding.
At the end of W6
Live enquiries close the validation and monitoring moves to Agent Care.
Duration
Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Fitness AI agent

Build a lead-conversion agent around the consent you already hold.

Show us your enquiry channels and who takes a lead over. If you can show us where each lead's consent came from, we'll map the follow-up, set the automation boundary and name what stays with the rep.

Typical delivery
4–6 weeks
Deployment
Your CRM and phone system
Rep decides
Consent-gated
After launch
Agent Care available

Nestack Agents · Lead conversion and sales follow-up

AGT-SF-07 · Agent Care available after launch

Build a lead-conversion agent

Four questions so the call starts from your actual setup — then pick a time that suits you.

About you
Your leads

No obligation. We’ll map the workflow, name the automation boundary and tell you if an agent is the wrong answer.