NEONDATA builds category-focused software and AI products for SMBs, and helps their owners run them through a modern lens. NEON and PUZL are the first two products, both live in the U.S. on one shared, model-agnostic engine.
Which channel drove last month’s bookings?
Paid search, with 41% of new bookings. Referrals were second at 27%.
Sources: booking records · ad accounts · PUZLMove Dana’s 2:00 tomorrow to Thursday and text her.
Software built deeply for one kind of business at a time, starting with aesthetics and wellness. NEON is the first.
Model Context Protocol products that connect the AI a team already uses to the business’s own data, limited by role. PUZL and NEON’s AI add-ons are the first.
Showing owners and operators, inside their own business, how to retire legacy systems and decide where AI fits, with any model they choose.
Most small businesses still run on software built before AI existed. The fix is not another vendor’s chatbot. It is modern systems, clean context and an owner who knows how to put any model to work. Our purpose
The AI was ready. The context was missing.
Co-founder Chris Castello co-owns and operates med spas. His teams asked AI questions about the business every day, and every answer depended on what someone remembered to paste into a chat window.
So his engineering team connected scheduling, patient records, marketing results, staff documents and daily operations, then put AI to work on top. That foundation is now the engine under both NEONDATA products.
Read the founding storyThe EMR and CRM for appointment-based aesthetics and wellness. Clinics run it from the Claude or ChatGPT account their team already uses, with every AI action limited by job role.
The context layer for small and midsize businesses. It connects a company’s systems, documents and permissions to the AI its team already uses, so answers are consistent and traceable.
The product began as one operator’s own team tooling, and it still runs inside his clinics.
One platform runs every NEONDATA product. The second product did not need a second stack.
Works with Claude and ChatGPT today and is not tied to any single model provider. The customer’s context stays with the customer.
Customers meet NEON and PUZL. NEONDATA appears to them only as an endorsement line. Both launched in the U.S. in 2026.
NEON replaces the patchwork most clinics run on — a separate EMR, booking tool, payment system, CRM, photo app and spreadsheets — with one record from first inquiry through charting, payment and follow-up.
Staff can work in NEON like any EMR, or connect their own AI account and do the same work by asking. The guardrails came first: the AI can draft a chart note, but it cannot sign one.
Med spas, aesthetic spas, medical weight loss, hormone therapy and men’s health. Charting, e-prescribing, injectable lot tracking and consent.
Day spas, salons, hair, nails, massage and tattoo. Same core record and approval model at a lighter tier, without clinical modules.
Connect the clinic’s Claude or ChatGPT to the EMR, then to marketing, revenue, operations and staff documents.
PUZL connects a business’s systems, documents and team permissions to Claude, ChatGPT and other AI tools. The same question from two employees draws on the same context and the same rules, and every answer traces back to its source.
It gives a small business the context foundation that enterprises build with dedicated data teams. PUZL is industry-agnostic and structured by location, so it fits owners adding sites and groups running several businesses.
NEON and PUZL serve different buyers. They share one data platform, one permission model, one set of AI connectors and one login. That is the part we expect to compound as we add products.
tools the NEON EMR exposes to a connected AI account.
of those an owner’s Claude or ChatGPT can carry out, scoped further for each job role.
reserved for a person and blocked for AI entirely, including signing charts, prescribing and refunds.
Anything that reaches a patient, moves money or changes the chart waits for a staff member to approve it. Every action is recorded.
Counts from the NEON product, October 2026.
U.S. small businesses alone carry nearly half of private-sector jobs, and more than 100,000 firms sit in the midsize band above them. Most have no data team, no integration budget and no AI staff. The models are ready. The context is missing.
Small means fewer than 500 employees, the SBA definition. Midsize, as we use it here, runs from 100 to 999 employees: the larger SBA-small firms plus those just above the line.
U.S. small businesses, including nonemployers. Together they account for about 46% of private-sector employment.
U.S. employer firms with 1 to 499 employees, 2023
firms with 100 to 499 employees, employing 18.3 million people, 2022
firms with 500 to 999 employees, 2022
Sources: SBA Office of Advocacy, 2025 Profiles · Census Bureau, Business Dynamics Statistics, 2023 · Census Bureau, Statistics of U.S. Businesses, 2022
Firms with 100 to 249 employees use AI at well over one and a half times the national rate. Even there, two in three do not. Both ends of the SMB market need the same things before AI can do real work: connected data, permissions by role and a person approving what matters.
Source: U.S. Census Bureau, BTOS, two weeks to May 3, 2026 · AI use in any business function · 250+ includes large enterprises
| The usual options | The problem | Our approach |
|---|---|---|
| AI wrapper startups | One model change away from irrelevance | A model-agnostic engine under two products |
| Vertical SaaS incumbents | Slow to become AI-native | Built new for bring-your-own-AI from day one |
| Vendor-owned AI assistants | Locked to one vendor’s branded assistant | The customer’s own Claude or ChatGPT, limited by role |
It is the business we run. It is also growing faster than the software behind it: the American Med Spa Association counted 10,488 U.S. med spas in 2023, up from 8,899, with average annual revenue of $1.40 million per location. Most are run by clinicians, and many are adding weight loss and hormone therapy that salon software was never built to chart.
Source: American Med Spa Association, 2024 State of the Industry recap (Nov 2024). Commercial datasets report a higher site count; we cite the conservative figure.
Where we go from hereOur aim is to be the AI infrastructure behind U.S. SMBs with 1–99 staff, starting with the category we know from the inside.
Context layer for U.S. SMBs, from $197 per month.
Med spas and wellness clinics in the U.S.
Day spas and salons on the same core record.
PUZL first. NEON once Canadian health-privacy requirements are worked out.
NEONDATA LLC is an Arizona company with three jobs: build category-focused software, build AI products on the Model Context Protocol, and help owners and operators run their businesses through a modern lens.
Deep software for one kind of business at a time, with the data model, workflows and guardrails that category needs. Aesthetics and wellness first, through NEON.
Model Context Protocol products that let a business connect Claude, ChatGPT or the next model to its own systems, limited by role. PUZL is the first.
Most owners have never seen AI work with full context. We show them, in their own business, what changes when they do.
Running a business through a modern lens means questioning the tools it inherited. These are the shifts we help owners make.
| The status quo | The modern lens |
|---|---|
| Separate tools for booking, records, payments and marketing | One record per business, from first inquiry to follow-up |
| Data locked inside legacy vendor software | Data the business controls and can bring to any AI |
| One vendor’s branded assistant | The AI the team already uses, chosen by the owner |
| Answers that depend on what someone pastes into a chat | Answers drawn from the business’s own context, traced to the source |
| AI as a novelty, or AI with no limits | AI with clear limits by role, and a person approving what matters |
MSN, APRN, FNP-C
Chris is a nurse practitioner who spent years in critical care before co-owning multiple med spas, where he still treats aesthetics and wellness patients. After years of disconnected software, repetitive documentation and thin reporting in his own clinics, he and his engineering team built the system they wanted to run them on.
That system became NEON. The context layer underneath it became PUZL.
“As a provider, I’m accountable for every chart my clinic signs. So the AI in NEON can draft a note, but it can’t sign one.”
New features run in our own clinics first. If they don’t save our front desk or providers time, they don’t ship.
Before we build a feature, we decide what AI may do alone and what waits for a person. Then we build to that line.
Real numbers with a source and a date. No claim we can’t show in the product.
We don’t depend on any single AI vendor, and neither do our customers. They bring the Claude or ChatGPT account they already pay for, and their business context stays with them as models change.
NEON Clinical debuts for med spas, aesthetic spas, weight loss and hormone therapy clinics, with every AI action limited by job role. NEON Salon is expected in Q1 2027.
Operator-built and AI-agnostic, PUZL connects a business’s systems, documents and permissions to Claude, ChatGPT and other AI tools.
NEONDATA LLC is an Arizona company that builds AI and data infrastructure for small and midsize businesses. Co-founded by Chris Castello, an owner and operator of multiple medical spa businesses, NEONDATA brings the tools his engineering team built for his own companies to SMB owners everywhere. Its brands are NEON and PUZL.
Write NEONDATA, NEON and PUZL in capitals. The web address stays lowercase: neondata.co.
We’re focused on building. Leave your email and we’ll send occasional notes on launches and progress. Press questions go to media@neondata.co.
Email us and tell us how you follow NEONDATA — investor, industry or press — along with your name, and we’ll add you to the list.
We use your email only to send NEONDATA updates. You can unsubscribe at any time. See our Privacy Policy.