AMP Integrated AI Roadmap — Done by Q2 2027

The Goal

$2B Platform Valuation

Company value = earnings × the multiple a buyer pays. Seven months of execution works both: higher earnings at every location from four operating engines, a proven playbook for absorbing acquisitions quickly, and the story that re-rates the multiple — an AI-operated platform running on data AMP owns.

Lever 1 · Higher Earnings Per Location

Four engines: fuller provider schedules, labor cost matched to actual demand, lower cost of goods through smarter purchasing, and more patients returning on time. Today, the share of consultations that become paid treatments ranges from 5% to 36% depending on the practice — closing that gap is the single biggest near-term prize.

Lever 2 · Faster Acquisitions

24 brands, ~100 locations (including Heyday's 11 corporate-owned + 18 franchise studios), 883 employees, 14+ states today. Two live migrations this window — Heyday onboarded in 45–60 days, Reflections complete by December on the standard playbook — prove that every future deal plugs into AMP's systems in weeks, not quarters.

Lever 3 · A Richer Multiple

Patient, staffing, and inventory data AMP owns and can prove — powering predictions, automation, recurring membership revenue, and a purpose-built patient app. By April, that story is demonstrable in diligence, not aspirational.

What We're Building — In Plain Terms

One data foundation with four layers. Today the same patient can exist as five different records across systems, and three versions of "utilization" circulate at once. This fixes both — permanently — and everything else on this page runs on top of it. The workstreams don't wait for it to finish: they launch now and get smarter as each piece lands.

1 · Getting the Data In

Automatic, reliable feeds from every systemZenoti, payroll, AesthetixCRM, ad platforms — flowing in continuously, with errors caught instead of silently dropped

One identity per patientDuplicate and mismatched records (10–12% of all patient records today) get merged into a single, verified patient — with a person reviewing the uncertain cases

Privacy built in from day oneHIPAA-grade protections; each brand sees only its own data; every access logged

2 · One Source of Truth

Complete patient recordWho they are, what they've had done, what they've spent, how to reach them

Marketing & lead funnelEvery lead, its source, and whether it became a booking

Schedules & capacityAppointments, open slots, no-shows and cancellations

Labor & payrollWhat staffing actually costs, connected to what it produces

Inventory & purchasing + revenue & paymentsWhat we buy, use, sell, and collect

3 · The Intelligence

ForecastsExpected demand by location and day; staffing needs before the schedule is built

Early warningsWhich patients are likely to lapse, which appointments are likely to no-show, where spend looks off

AI assistants that actTrigger a win-back campaign, offer a reschedule, flag a purchasing anomaly — automatically, with human sign-off

4 · Where People Use It

One number, everywhereEvery report and board pack reads from the same agreed definitions — no more three versions of the truth

Messaging that's personal and relationalAesthetixCRM automatically receives each patient's treatment dates, history, and risk scores — so every text and email reads like it came from their practice, about them, at the right moment

Provider & patient toolsVisit-note drafting, pre-visit prep, and the patient app all read the same trusted record

The Roadmap — Seven Months, Four Sprints, Run in Parallel

Each block states the work and what it unlocks for the business. No lane waits on another: while the data foundation is being built, everything else keeps moving and plugs in as pieces land.

Workstream

Sprint 1 · Mobilize

Decisions & Fast Wins

Sep – Oct 2026

Sprint 2 · Cutover Quarter

Migrations Land

Nov – Dec 2026

Sprint 3

Intelligence Online

Jan – Feb 2027

Sprint 4 · Land It

Complete & Harden

Mar 2027

01 · The Foundation

Data Foundation for AI & ML

One trusted home for all patient, scheduling, labor, and inventory data — built by parallel teams, not one long line

Measures: duplicate patient records below 2% · every report on one agreed set of numbers

Platform decision; build partner already selected Decision · Sep 21Agreement signs the week of Sep 21 — the build clock starts immediately

Connect the core systemsZenoti, payroll, and AesthetixCRM flow in automatically and reliably — replacing today's fragile, undocumented plumbing

Begin merging duplicate patient recordsUnlocks: marketing stops emailing the same person three times, and every count becomes trustworthy

Complete patient record + marketing funnel liveUnlocks: for the first time, a lead can be followed from ad click to paid treatment · Nov

Schedules connected to payrollUnlocks: the true cost and true output of every provider hour, in one place · Dec

Reporting moves onto the new foundationUnlocks: one agreed definition for every number in every board pack

Inventory, purchasing & revenue data completeUnlocks: what we buy connected to what we use and what we sell · Jan–Feb

First predictions liveDemand forecasts, lapsed-patient risk scores, no-show likelihood — feeding campaigns and schedules the moment they're ready

Heyday's history folded inUnlocks: the acquisition playbook — proven, documented, repeatable

First AI assistants pilotedCampaign triggers and purchasing alerts that act automatically, with human sign-off

Built to run without heroesDocumentation, backups, and monitoring so no single person is ever the point of failure again

02 · Critical Path · 45–60 Days

Heyday Migration

All Heyday locations move onto AMP's systems (Zenoti, Sage Intacct, Centime, WooCommerce) in one coordinated cutover — live within 45–60 days of kickoff

Measures: live on AMP systems by early Nov · old contracts terminated by Dec 31 · zero inherited balances or terms · reconciled to the penny

Week 1: request stored payment cards from the old system #1 RiskA weeks-long vendor process; without it, member billing breaks at cutover. Backup runs in parallel from day one: members update cards ahead of the move

Week 1: audit every legacy contract; termination letters outRenewal dates and notice periods lock the exit — nothing gets inherited

Configuration locked, then dress rehearsal #1 (mid-Oct) and #2 (late Oct)The entire migration practiced end to end, twice

Cutover weekend — every location at once Day 45–60Go/no-go checks first: patient counts match, gift-card balances match to the penny, memberships intact · by early Nov

Four weeks of intensive support, books run in parallelDaily check-ins across operations, service, and accounting until everything is proven clean · Nov

Old systems terminated — by Dec 31 at the latestTheir recurring costs come off the books before FY27 starts

Heyday's history joins the data foundationUnlocks: Heyday locations get the same campaigns, alerts, and reporting as everyone else · Dec

Patient Journey Automation — AesthetixCRM

Marketing and follow-up move off Zenoti + HubSpot onto AesthetixCRM: pilot proves it, waves roll it out, and every message becomes personal and relational

Measures: every lead answered in under 5 minutes · more consults becoming treatments · no-shows recovered · rebooking rate

Pilot live — Destination Aesthetics Mid-OctInstant response to every inquiry, no-show and cancellation follow-up, and win-back campaigns (90-day tox, 6-month filler) — none of which exist today

Pilot punch list closed ahead of go-liveWebsite and domain updates, texting registration per brand, call-center users loaded, location-based lead routing, team live-training — the concrete items between today and a proven pilot. Patient opt-ins/opt-outs stay governed at the Zenoti profile level and flow to AesthetixCRM through its native integration

Pilot proves out → wave rollout begins early Nov Exit check · late OctSpeed-to-lead, routing, and campaign results verified at Destination Aesthetics before the next brands onboard

Full rollout: every AMP-brand location live on AesthetixCRMWaves begin early November; Heyday's studios join after their migration lands; one phone/text system per brand replaces today's patchwork; local teams trained and running their own campaigns as they onboard · Nov–Dec

Messaging becomes personal and relationalUnlocks: "your Botox is due" arrives at the right week for that patient, from their own practice, referencing their actual visits — not a batch blast · Dec

HubSpot retired — with full records preservedComplete history archived, forms and pages redirected, then the contract ends by Dec 31

Campaigns get predictivePatients likely to lapse get outreach before they're gone; likely no-shows get confirmations and backups; timing and channel tuned per patient

Rescheduling assistant pilotedWhen a provider's day won't fill, the system offers patients a move — automatically, with staff approval

Every email, text, and call traceable to the revenue it produced

Labor Optimization

Staffing matched to real demand — fuller schedules, lower labor cost as a share of revenue

Measures: one official utilization number · labor as % of revenue · revenue per provider hour

Real-time utilization alerts live ShippedManagers see under-booked providers as it happens, not in a month-end report

Connect utilization to payroll cost The blockerUntil this lands, no claimed labor savings can be verified — it's the first fix, not the last

Pick one official definition of utilization Leadership call · Oct 15Three versions circulate today and compensation rides on one of them

Hardened-build review — NxGPToday's utilization tooling gets assessed and roadmapped into a hardened, production-grade software build

Provider scorecard completeUnlocks: seeing which providers convert, upsell, and hold price — and coaching the gap · Nov

First verified savings number publishedA real figure with the math shown — the labor business case, proven

Demand forecast by location and dayUnlocks: staffing recommendations before the schedule is built, not after shifts sit empty

Tackle partially-empty schedulesWhere most of the real money is — beyond the obvious fully-empty shifts

Schedule recommendations live in pilot regionsEvery staffing decision shows its predicted revenue impact before it's made

Procurement & Inventory — AMPly

Our in-house platform already runs network ordering; this window extends it from injectables to every spend category — and toward purchasing that thinks ahead

Measures: days of inventory on hand · cost of goods % · spend visibility across all three categories

AMPly runs ordering across the network today LiveReorder recommendations sized by booked appointments, center-to-center transfers, cycle counts, and vendor ordering — injectable days-on-hand already cut by more than half

Baseline every spend categoryInjectables are controlled; retail is consolidating; back-of-house supplies are the least visible and the fastest-creeping — discovery and a consumption baseline close this month

Invoice coding automated in SagePurchases flow to the right ledger lines without manual re-keying — believed complete; confirming this month whether the integration is full or partial

Hardened-build review — NxGPAMPly and the procurement stack get assessed and roadmapped into a hardened, production-grade software build

Retail and back-of-house brought under item-level controlUnlocks: the two categories with the most cost-savings headroom finally managed like injectables

Three-way match: ordered vs. received vs. paid Needs: inventory + revenue data liveUnlocks: catching billing errors and shrinkage automatically — a control that doesn't exist today

Purchasing gets predictiveReorders sized by demand forecasts, not just recent usage — accuracy tested against real history before it drives orders

Sourcing assistant pilotedFinds the best price, timing, and vendor for a given product and proposes the purchase — human approves; injectables move toward just-in-time

AI Scribe, Consultation Support & Pre-Charting

Purpose-built in-house — visit notes written by AI, providers' hours returned to patients

Measures: documentation hours returned per provider · consults becoming treatments · chart completeness

Custom AI Scribe piloted and running by Oct 1 CommittedVisit notes drafted automatically at pilot practices — providers review and sign instead of typing

Patient consent & privacy cleared before the first recorded visitRecording consent varies by state; agreements and workflows in place first — this protects the whole program

Standard across brands by DecemberFull rollout right behind the pilot — time saved and chart quality measured against baseline as it scales · Nov–Dec

Pre-visit prep live Needs: patient record liveUnlocks: providers walk in with the patient's full history and flags summarized — no chart-digging

Consultation support pilotedDuring the consult, the system suggests the treatments this patient is most likely to want next — same data that times the win-back campaigns

Consultation support expands with the rollout brandsSuggestions refined against real consult outcomes

Digital Marketing & SEO

Know which dollar produced which patient — measured in our own system, because ad platforms no longer allow health brands to track it in theirs

Measures: marketing spend vs. actual patient revenue · cost per new patient · share of leads from free/organic channels

Ad-account compliance check firstMeta now restricts health-related businesses from tracking bookings and building patient-based audiences. We confirm how each brand is classified and design around the rules — before building anything that would get shut off

Keep today's ad performance while HubSpot retiresNo regression during the transition — validated side by side

Website expansion: 10 more practices with World Tech Miami Budget decisionFast, modern sites built to rank in search, expanding on the pilot — investment up to ~$100K; budget alignment needed before the builds kick off

The step-change: spend tied to real patient revenue Needs: marketing funnel liveMeasured inside AMP's own data — where no ad-platform policy can take it away

Local search program per locationGoogle Business profiles, listings, and review flow — free patient demand matters more as paid targeting gets restricted

Budget follows evidenceMonthly reallocation toward the campaigns and markets producing actual patients, not clicks

New practice sites launch; AI-search visibility work begins

Spend allocated by predicted patient lifetime valueEvery paid dollar measured against patient revenue, internally

08 · Done by December

Reflections Medical Migration

The plastic surgery line joins the standard stack on AMP's proven Zenoti migration playbook — and every plastic surgery center standardizes on 22 Plastic Surgery's workflows

Measures: complete by December · every chart and consent accounted for · one workflow standard across plastic surgery

Migration kicks off on the standard Zenoti playbookThe same playbook AMP uses for every data migration — scope locked: which records live in Zenoti vs. Symplast, what moves, what connects · Sep–Oct

Workflow standard set: 22 Plastic SurgeryAll plastic surgery centers adopt 22 Plastics' Zenoti + Symplast workflows — one way of operating, not three

Migration complete by DecemberData moved, every chart and consent verified present, teams live on the standardized workflows

History joins the patient record; plastic surgery visible in cross-brand analytics$8–20K-per-case economics finally comparable beside the med spa business · Jan

Patient App — Wellness MVP

Purpose-built for AMP's business — booking, membership, and treatment plans in the patient's pocket. Only ~13% of appointments are booked or managed online today; the other ~87% land on the front desk, which is why front-of-house staffing runs as heavy as it does

Measures: active users · rebookings made in-app · membership sign-ups in-app

Purpose-built — committed; scope lockedA $2B platform owns its patient experience; no off-the-shelf app. First version: booking, membership management, treatment plan, wellness content — reading the same trusted patient record as everything else · Jan

Design and build begin, in-house dev team leadingEach practice's own brand on the front (per policy: the practice is credentialed, never the parent)

Build underway; beta and app-store submission nextApple/Google review planned into the schedule, not discovered at the end

10 · Cross-Cutting

Data Governance

The rules that keep the foundation trustworthy as it grows

Measures: every data area owned · every metric's formula published · brand-level separation verified

Every data area gets a named ownerSomeone accountable for its accuracy — not "the system"

Brand-level data separation and access logging live from the first loadEach brand sees only its own data; every access recorded

Every metric published with its formula and sourceAnyone can see exactly how a number is calculated — transparency by default

Every automated action logged, human sign-off where it mattersAs predictions start driving campaigns, schedules, and purchases, the audit trail keeps pace

Data quality monitored continuouslyFreshness and accuracy checks run on their own; problems surface before anyone reports a wrong number

Done / live In motion Waiting on a decision or dependency Watch closely — risk or legal groundwork

The Decisions That Keep Us on Schedule

Seven months leaves no room for open-ended debates. Each decision below has a date — and a default answer that goes into effect if the date passes.

1 · Heyday payment cards handed overRequested from the old vendor in week one. If not confirmed by Oct 10, the backup carries the date: members update their cards before the move.

Protects: the 45–60 day go-live

2 · Legacy terminations filedContract audit confirms notice periods by Sep 30 and the termination letters go out — locking the exit with nothing inherited.

Protects: zero costs carried into 2027

3 · Data platform confirmed Sep 21Build partner already selected; the agreement signs that week and the build clock starts.

Unlocks: everything in lane 01

4 · AesthetixCRM pilot exit check — late OctPilot live at Destination Aesthetics mid-Oct; speed-to-lead, routing, and campaign results verified there — passing it opens the brand-by-brand waves starting early November.

Protects: the CRM rollout quality

5 · One official utilization numberThree versions circulate and compensation rides on one. Leadership picks by Oct 15.

Unlocks: a labor savings figure everyone trusts

6 · Patient app scope locked — mid-JanPurpose-built is decided; freezing what's in and out of the first version keeps the build on rails.

Protects: the app build from scope creep

7 · Ad-account compliance review doneBy Oct 1 — confirms what tracking each brand is allowed before we build the marketing measurement around it.

Protects: lane 07 from being switched off

8 · Scribe pilot live Oct 1Purpose-built, in-house team leading — committed, with consent workflows cleared first.

Unlocks: the clinical workflow lane

9 · Website expansion budgetUp to ~$100K for 10 practice sites with World Tech Miami. Align by early Oct so builds start right behind the pilot.

Unlocks: the search-ranking foundation for 10 more practices

Why This Order

Heyday is the sprint inside the sprint: live in 45–60 days. Payment-card request and termination letters in week one; two full dress rehearsals in October; cutover by early November; support month and final reconciliation right behind it. Old contracts dead by Dec 31 at the latest.
Reflections runs on its own track, done by December. It uses AMP's standard Zenoti migration playbook — the same one used for every migration — and carries the workflow decision with it: every plastic surgery center standardizes on 22 Plastic Surgery's way of operating.
Everything runs in parallel — nothing waits on the foundation. The CRM, scribe, labor tools, AMPly, and website all launch now on the data available today, and each one gets sharper as the foundation delivers underneath them.
AesthetixCRM earns its rollout. Pilot live at Destination Aesthetics mid-October, waves begin early November, full network through the fall — and messaging graduates from batch blasts to personal, relational outreach timed to each patient's actual treatments.
Labor proves its savings before it predicts. Connect payroll, agree on one number, publish a verified figure — then forecast.
The scribe scales on evidence, fast. Piloted Oct 1, standard across brands by December — a purpose-built tool doesn't need a six-month rollout.
The patient app starts when its foundation exists. Scope locks in January against a live patient record — so the app is born connected, not another silo to clean up later.
Marketing measurement moves in-house by design. Ad platforms now restrict what health brands can track; our answer is measuring patient revenue in our own system and giving the platforms only what the rules allow.

What It Takes to Hit April 1

Parallel everything. Ten lanes moving at once, two to three build tracks inside the foundation alone — the same total work, spent faster.
Decisions in five business days. Every open question above has a date and a default. Schedules die from pending approvals, not from engineering.
Capacity through partnership, not recruiting. Engineering, data, and AI delivery flexes with the roadmap — no seats sitting open on the critical path.
Scope discipline. First versions stay first versions — the app ships its locked scope; AI assistants act only with human sign-off; wave-two items stay in wave two.
Migration quality is not negotiable. Heyday's 45–60 day pace comes from starting rehearsals immediately — never from skipping the second rehearsal or the reconciliation checks. Reflections holds the same bar on the standard playbook.
One weekly executive review across all ten lanes. A slip anywhere surfaces in days, not at quarter-end.

Where We Stand on April 1, 2027

Foundation: all patient, scheduling, labor, and inventory data in one trusted place; duplicates under 2%; every report on the same numbers

Heyday: live on AMP systems since early November, reconciled to the penny, old contracts terminated with nothing inherited

Reflections: migrated by December on the standard playbook; all plastic surgery centers on 22 Plastics' workflows; in cross-brand analytics

Patient journey: every AMP-brand location on AesthetixCRM; HubSpot gone; every message personal, relational, and timed to the patient's actual treatments

Labor: one official utilization number; a verified savings figure published; forecasts guiding schedules in pilot regions

Procurement: all three spend categories under item-level control in AMPly; three-way match live; sourcing assistant in pilot

Clinical: purpose-built AI scribe standard across brands since December; pre-visit prep live; consultation suggestions expanding

Marketing: every paid dollar measured against real patient revenue; local search running at every location

Patient app: purpose-built Wellness app in active build against locked scope, beta and app-store review scheduled

Beyond April — The AMPly Innovation Pipeline

AMPly is already the network's ordering brain. As the data foundation matures, these are the next capabilities queued behind it — each one compounding the procurement engine.

AMPly owns the inventory count

Real-time on-hand quantities maintained by AMPly itself — anchored by cycle counts and updated with every order, transfer, and treatment — instead of waiting on the booking system's snapshots. The groundwork is already running in shadow mode.

Forecast-driven ordering, proven then trusted

Reorders sized by predicted demand with guardrails (a floor so nothing runs out, a ceiling so nothing piles up). A testing harness already measures forecast accuracy against real history — the engine earns autonomy with evidence.

Pharmaceutical chain-of-custody (DSCSA)

Lot-level tracking from vendor to patient for injectables — the compliance posture a $2B healthcare platform is expected to hold, scoped and ready to build when prioritized.

Agentic sourcing at full autonomy

The Sprint 4 sourcing assistant graduates: monitoring vendor pricing and promotions continuously, timing purchases, and routing orders to the best source — with human sign-off shrinking as trust is earned.

Shipment-to-shelf visibility

Vendor shipment feeds already flow into AMPly; next: automatic matching to purchase orders, "shipped but never received" alerts, and delivery-date intelligence feeding the reorder engine.

Manufacturer program optimization

Training bundles, rebates, and loyalty programs (Allergan and peers) factored directly into order recommendations — so every order captures the incentive money currently left on the table.

Resourcing Model — Four Kinds of Work, One Deliberate Mix

Every technology organization does three jobs at once: building new things, automating recurring work, and keeping what's already shipped running. AMP has a permanent fourth — because corporate solutions is the growth engine, there is always an acquired business to migrate in and an inherited tech stack to deconstruct. Most companies never manage the mix — maintenance quietly eats the team. We manage it explicitly: one capacity pool, allocated on purpose, reviewed weekly.

How the one capacity pool is allocated across the four kinds of work. For context: Gartner benchmarks show a typical company spends roughly two-thirds of its technology budget just keeping existing systems running. Our model inverts that — build-heavy during the roadmap window, with maintenance capped at 15% and never allowed past ~30% at steady state. The cap is enforced by an engineering rule, not hope: nothing ships without monitoring and documentation, so it doesn't come back as break/fix.

Lane 1 · 40% now → 30% steady

Build & Design

Create what doesn't exist yet. The data foundation, the purpose-built patient app and AI scribe, internal apps (ordering, concierge, dashboards), and the website expansion.

Demand is roadmap-driven and spiky — it maps exactly to the sprint plan on the first tab and shrinks once the seven-month window closes.

Success looks like: milestones hit on their dates — and every deliverable handed over with documentation and monitoring, so its upkeep is prepaid.

Lane 2 · 20% now → 15% forever

Migrations & Deconstruction

Absorb what we acquire; tear down what it arrives with. Every corporate-solutions deal brings a business to migrate onto the standard stack and an inherited tech stack to deconstruct — data moved on the Zenoti playbook, contracts terminated, balances never inherited.

Heyday and Reflections are reps one and two. With 2–3 acquisitions expected annually, this lane never closes — it's a permanent part of the growth lifecycle, and every rep makes the playbook faster.

Success looks like: onboarding time falling with each rep; $0 in inherited terms; every migration reconciled to the penny.

Lane 3 · 25% now → 25% steady

Automation

Replace recurring corporate work with systems. Agentic and autonomous flows: the IT/HR service desk agents (already live), automated invoice coding (live), lead routing, campaign triggers, purchasing alerts, and report generation.

This is the lane that holds corporate headcount flat while locations multiply — a continuous backlog, ranked by one number: hours returned per month.

Success looks like: hours returned, tickets deflected, and corporate headcount per location falling every quarter.

Lane 4 · 15% now → capped ~30%

Maintenance & Reliability

Keep everything already shipped running. Break/fix on data feeds, failed webhooks and API connections, vendor changes (Zenoti updates), security patching, backups, and hardening.

The honest truth: this lane grows with every build that ships. Left unmanaged, it eats the team — which is exactly what happened to the last data platform. So it runs against a cap.

Success looks like: problems caught by monitoring before a user notices; fast time-to-restore; capacity share holding at or under the cap. Breaching the cap is the signal something shipped without hardening.

How the Pool Is Governed

One intake, one backlog. Every request — build, automation, or fix — lands in the same queue and gets ranked in the same weekly executive review already on the roadmap.
Build pays its own maintenance bill. No deliverable is "done" without monitoring, documentation, and a named owner. That's the mechanism that keeps Lane 3 capped.
Automation is ranked by hours returned. Corporate sprawl is attacked in order of payback — the flow that returns the most people-hours per month builds first.
Same people, rotating lanes. Engineers who carry the pager build things that don't page. Delivery capacity flexes through the partnership, so the mix can shift week to week without hiring or benching anyone.
Live is the starting line, not the finish. Tools already shipped — AMPly, utilization alerts, the service agents — earn their impact through continuous optimization. The automation lane's capacity is reserved for tuning what's live before building what's next.

Who Does the Work — Resources by Type of Work, per Window

Each cell lists only the resources supporting that type of work in that window. AMP's bench is from the actual roster; the vendor bench is per the signed NX Growth Partners SOW and the named implementation partners.

dedicated development resources across all four lanes

The development bench, counted:

AMP — 1: Jefferson Lich, Lead Systems Architect (platform reliability + select build/design)
NxGP core pod — 5: Principal Architect / Technical Lead (1) · Software Engineers (2) · Data Engineer (1) · QA Engineer (1)
Flexing, not counted: NxGP DevOps/cloud/security specialists + delivery management; expandable one engineer at a time under the SOW
Not development resources: the Zenoti admins and Help Desk & Systems teams — day-to-day ops support; on this table only where they support the Zenoti migrations

Window	Build & Design	Migrations & Deconstruction	Automation	Maintenance & Reliability
Roadmap window · Sep '26 – Mar '27
AMP Team	In-house development team — AMPly; product direction & acceptance on the NxGP builds Jefferson Lich — select build/design work Department product owners — requirements & acceptance Business Transformation — program coordination	IT — Zenoti admins + help desk — Heyday & Reflections cutover support: devices, access, go-live Finance & Accounting — reconciliation sign-off, parallel books Field & studio teams — counts, dry runs, cutover execution	In-house development team — builds and tunes the agentic flows Department owners — rank the queue by hours returned Inside Sales + field teams — pilots, UAT, training waves	Jefferson Lich — the one AMP engineer qualified for platform reliability
Vendor & Partner Teams	NxGP pod — AI scribe, patient app, workflow builds, and the warehouse redesign (Principal Architect, both software engineers, data engineer, QA) World Tech Miami — 10-practice website expansion	Zenoti onboarding/migration team — Heyday + Reflections on the standard playbook Sage Intacct & Centime implementation — the financials cutover NxGP data engineer — history loads into the foundation	AesthetixCRM onboarding team — campaign and journey automations, pilot through waves	NxGP DevOps/cloud/security specialists — monitoring, CI/CD, hardening
Steady state · after Apr 1, 2027
AMP Team	In-house development team — next-initiative builds (Pavilion Flow pilot, NuPlay, finance automation) Product ownership of everything shipped	Department owners + field teams — staff each acquisition's cutover IT — Zenoti admins — per-deal Zenoti migration support	In-house development team — owns the backlog; optimizing live tools first Department owners — rank by hours returned	Jefferson Lich — runs platform reliability against runbooks
Vendor & Partner Teams	NxGP pod — reserved capacity; scales up per initiative, back down between them	Migration vendors re-engage per acquisition — AMP's playbook, vendor throughput	Platform vendors on support contracts — Zenoti, AesthetixCRM, Intacct/Centime, data platform	NxGP pod — a natural backup to Jefferson

The Vendor Math — Same Roadmap, With vs. Without the Partner Bench

The honest capacity estimate. The roadmap is roughly 42 engineer-months of build, migration, and automation work (6 dedicated development resources × 7 months). AMP's own qualified engineering capacity is one person — who also carries platform reliability. The chart shows what that means for the calendar.

How the estimate is built: ~42 engineer-months of work ÷ effective capacity per scenario. AMP alone: one qualified engineer minus his reliability duties is ~0.5–0.7 effective engineers — lanes collapse from parallel to sequential, the Heyday cutover slips past the Dec 31 contract terminations (inheriting the renewals we're paying to escape), the CRM waves and website builds lose their implementation teams, and the scribe and patient app get cut to keep the lights on. Hiring instead — the scoped hire path: an ML engineer plus a data engineer, each seat estimated at 2–4 months from posting to in-seat — and the one req actually opened was paused mid-search in September. The warehouse still needs its own 4–5.5 person specialist pod for six months (~30 engineer-months) regardless of who else is hired. Leadership's read of this path lands the full scope ~10–15 months out: two seats can't run ten lanes in parallel, so builds queue behind each other — and the fixed payroll outlives the build spike, versus a vendor bench that flexes down when the window closes. The vendor bench is what makes April 1 arithmetic instead of ambition — and it flexes back down when the window closes. Estimates are directional, stated so they can be challenged.

Team Impact — Scale the Business, Not the Overhead

Grounded in the active roster (Sep 19, 2026): a corporate team of 49 supports 883 employees across 24 brands and ~100 locations — ~70 legacy plus Heyday's 29 studios (11 corporate-owned, 18 franchise). Corporate is just 5.5% of total headcount, and this roadmap is how it stays that lean at 140+ locations. Where a system removes work, people are redeployed to revenue-side work first; reductions happen through attrition and planned transitions, not cuts. That's already happening: one supply-chain seat is scheduled to sunset the moment AMPly auto-ordering covers retail and back bar across all vendors.

corporate team members — 5.5% of the 883-person company

17:1

field-to-corporate leverage today — the ratio this roadmap protects and extends

corporate hires required by this roadmap — capacity comes through partnership and automation

~17

corporate roles avoided by 140 locations if today's ratio were simply held (illustrative)

The team is re-evaluated as the tools arrive.
Every sprint that ships a new capability also resets the bar for the people working beside it. Through the development cycle, each corporate role is re-evaluated against four things: the skill set to use the new tools, the curiosity to push them further, the execution to turn their output into results, and the quality standard that advanced tooling makes possible. People who pair well with these systems become dramatically more valuable — and get first claim on the redeployed capacity. Where a role and the toolset can't be paired, the role gets redefined; training comes first, and changes run through the same attrition-and-redeployment approach as everything else on this page. The goal isn't a smaller team — it's a team where every seat is amplified by the platform underneath it.

Corporate headcount required as the location count grows — illustrative projection from the actual roster. Gray is AMP now: today's operating model, with its ratio (49 corporate for ~100 locations / 883 employees) held as the network scales. Navy is AMP future: the roadmap's automation lane absorbs most of the added load, with modest growth for genuinely new scope. At 140 locations the difference is roughly 17 corporate roles — avoided, not cut. Validated quarterly against the automation lane's actual hours returned.

Workstream by Workstream — Where the Hours Come From

Each row names the recurring work a system takes over, who carries that work today (real roster counts), and the honest headcount effect. Note that "live" is the starting line, not the finish — the tools already shipped (AMPly, utilization alerts, service agents) earn their headcount impact through continuous optimization, and the automation lane reserves capacity for exactly that.

Workstream	Recurring work the system takes over	Who carries it today	Headcount effect
On this roadmap
Data foundation	Manual report building, cross-system reconciliation, one-off data pulls, "whose number is right" debates	Finance & Accounting (5), corporate ops, brand managers	Avoid hires absorbs analyst load as brands scale — avoids 2–3 analytics/reporting hires
Patient journey (AesthetixCRM)	Manual speed-to-lead dialing and round-robin, win-back list pulls, no-show chasing, campaign assembly	Inside Sales (12) + ~164 PCC/concierge in the field	Hold flat lead volume grows; the 12-person inside sales team doesn't — leads handled per person rises instead
Labor optimization	Schedule auditing, utilization report assembly, staffing guesswork	Practice managers + corporate ops	Redeploy manager hours move from spreadsheets to coaching and patients
AI scribe & pre-charting	Visit documentation, chart preparation, after-hours note-writing	Providers (field)	Field hours returned documentation time becomes treatment time — the revenue side of the P&L
Procurement / AMPly	Order building, stock checks, transfer hunting, invoice coding into the ledger	Supply Chain (3) — AP touchpoints already automated out	Already real one supply-chain seat is planned to sunset once AMPly auto-ordering covers retail and back bar across all vendors — the first automation-driven transition on the books
Heyday & Reflections migrations	Two duplicate back-office stacks: separate ledgers, AP processes, reporting	Finance & operations	Avoid hires 20+ locations absorbed with zero added back-office headcount
IT/HR service agents (live)	Password resets, onboarding/offboarding tickets, routine HR questions	IT (7: Zenoti + Help Desk & Systems) & HR (7)	Hold flat ticket volume scales with 883 employees; a 7-person desk doesn't
Marketing & SEO in-house attribution	Manual channel reporting, agency reconciliation, spend allocation debates	Marketing + external agencies	Redeploy reporting labor becomes campaign work; agency spend trims as measurement moves in-house
Patient app	Booking, rescheduling, and membership management. Only ~13% of appointments are booked or managed online today — the other ~87% arrive as calls and walk-ins the front desk must absorb, which is what forces front-of-house staffing to stay heavy just to hold support quality	Front desk & call center (field)	Field hours returned every point of online-share gained is front-desk load removed — the app moves the 13% number directly, and staffing needs follow it down
Future workstreams — the next wave of leverage
Pavilion Flow (evaluate)	Check-in at a kiosk, checkout and rebooking on an iPad at the chair — runs natively on Zenoti, the stack we already operate	Front-of-house: ~164 PCC/concierge network-wide, plus Heyday's dedicated 47-person front desk and 22 studio attendants	Evaluate · pilot vendor reports −42% front-desk hours per site — see spotlight below
Voice AI — NuPlay by Nurix AI (demoed, not yet committed)	The reality of lead calling today: most outbound attempts land in voicemail, and when a lead connects and wants to move forward it's a 2–4 minute booking-mechanics call. Voice AI absorbs both — the voicemail loops and the routine booking calls — and routes only the calls that deserve a human: real questions about treatments, providers, or the practice	Inside Sales (12), call center & front desk	Re-engage post-CRM demo complete; wasn't implementation-ready because phone numbers and routing weren't in place — the AesthetixCRM rollout (per-brand number pools + routing) is the prerequisite. Target: the 12-person inside sales team shrinks into a smaller, elite consult-conversion group handling only the calls where expertise moves revenue
Finance close & three-way match (post-foundation)	Matching purchase orders to receipts to invoices — a gap today — plus close checklists and reconciliations	Accounting	Avoid hires the finance team absorbs acquisition volume without growing

Spotlight: Pavilion Flow — the Front Desk, Self-Served

A self-service kiosk handles check-in and walk-ins at the door; station iPads handle checkout and rebooking at the chair. It runs on Zenoti — so it rides the exact stack every AMP location already uses, with no new integration build. Vendor-reported results across their operator network:

−42%

front-desk hours per site

15s

average self-service check-in

+38%

walk-in conversion

+24%

rebook rate uplift

Why the front desk is this big in the first place: only ~13% of appointments are booked or managed online today — the app and Pavilion Flow attack that same dependency from both ends (self-service before the visit, self-service at the door). The directional math, from the actual roster: AMP's front-of-house is real, countable headcount — ~164 PCC/concierge across the brands, plus Heyday's dedicated 47-person front desk and 22 studio attendants. Applied to Heyday's model alone, −42% of front-desk hours is ~20 FTEs of capacity; extended across the network's front-of-house hours it scales toward 60–90 FTE-equivalents — redeployed toward membership sales and patient experience, or reduced through attrition. The recommendation: pilot at 2–3 locations in Q1 2027, after the CRM waves land, and validate the vendor's numbers inside our own four walls before counting a dollar. Vendor also claims payback inside the first quarter — the pilot proves or disproves it cheaply. Heyday, with the largest dedicated front desk in the network and already mid-migration onto Zenoti, is the natural pilot ground.

The Evidence Behind the Case

Published results from organizations that made these same bets — the reason each lever is on the roadmap.

Speed to lead is the highest-leverage minute in the funnelThe MIT / InsideSales lead-response study found contacting a lead within 5 minutes vs. 30 makes you ~21× more likely to qualify them and ~100× more likely to reach them; a Harvard Business Review audit found the average company takes 42 hours to respond.

Why Patient Journey Automation leads the roadmap

AI scribes return real clinical hoursKaiser Permanente's 7,260-physician deployment (published in NEJM Catalyst) logged 2.5M visits and returned ~15,800 documentation hours; 84% of physicians said it improved their connection with patients.

Why the scribe is worth owning, not renting

Unmanaged tech teams drown in upkeepGartner benchmarks show typical companies spend roughly two-thirds of their technology budget just running existing systems — leaving a third for everything new.

Why maintenance runs against a hard cap in our model

Self-service front desk is proven in this exact verticalPavilion Flow operators (salons, wellness, med spas on Zenoti) report −42% front-desk hours per site and payback inside the first quarter — vendor-reported, which is exactly why we pilot before we plan around it.

Why Pavilion Flow makes the future list