Egma AI

Egma AI

AI-native marketing agency for elective care clinics

We're building the AI growth engine for elective healthcare clinics, starting with GEO (generative engine optimisation) for med spas. Med spas live and die by lead gen, but the way patients discover clinics is changing rapidly. They increasingly search through platforms like ChatGPT: "best Botox injector in SF" or "should I get a facelift or filler?" Most clinics don't know if they appear or how to get recommended. Egma helps medspa chains win local visibility inside AI search. Instead of treating a clinic brand as one generic entity, Egma optimizes each location against its own competitors, treatments, reviews, and patient questions. We monitor high-intent prompts, identify LLMs to cite. We make our GEO self-improving by integrating with the clinic's phone systems or deploying a 24/7 voice agent for them, because patient calls contain the highest-signal intent. E.g. with a pilot customer, we saw a caller ask about TRT pricing — first-hand local demand that can become a page like "TRT pellet therapy cost in San Mateo." We start with med spas because they spend heavily on marketing. Then we expand the same GEO agent to other elective categories like plastic surgery, dermatology, fertility, etc.

SF — and we'll stay in SF after YC. SF is the best place to build an AI company.

In four weeks, we went from customer discovery to $48K in pilots across 3 deals and 14 clinics — with one already deployed in production. The deployed pilot is Remedy Medical Aesthetics, run by Cameron Hemphill — ex-CEO of Growth99, the largest medspa marketing agency in the US (1,000+ medspas served). He's a category KOL and this deployment is our proof point for a Growth99 channel partnership that would put our agents in front of their entire client base. Five minutes after we went live on May 8th, Cameron called offering to intro us to another medspa. We have a signed contract worth $14,400 per customer (5-clinic system) under a B2B channel partnership with Supahealth, a healthcare RCM company. 4 more of their customer accounts are queued behind it. We are also piloting GEO for an 8-location PE-backed medspa called Revive MD worth $30K.

Four weeks. Both of us full-time.

Our stack is Bun + TS. We are building our agents on top of the OpenCode harness. We use Claude Code and (recently) Codex to build the product. We deploy on Cloudflare Workers and Durable Objects. For voice, we use Gemini 3 Flash + ElevenLabs TTS — deployed on LiveKit. We use Browserbase for browser infrastructure. The stack allows us to move extremely fast and deploy new GEO optimizations within minutes of identifying gaps. Our GTM runs entirely on Claude Code + Computer-use + Gooseworks (YC W23) + Custom Skills for tone and voice matching for personalized outbound to potential customers. BunTypeScriptOpenCodeClaude CodeCodexCloudflare WorkersDurable ObjectsGemini 3 FlashElevenLabsLiveKitBrowserbaseComputer-useGooseworks

Yes.

Three customers, all in pilot. None paying yet — billing on the first deal starts in 2 weeks. Remedy Medical Aesthetics (Minnesota) — we started with a front-desk voice agent as the wedge, deployed in production on May 8th and already handling live calls. This voice data will feed our GEO agent. Two-week trial, then $3.6K. Suncrest Clinic (Utah) — $14,400 signed contract across 5 clinics, via B2B partnership with Supahealth. Deployment starts June, gated on clinic bandwidth. Revive MD — 8-location PE-backed medspa. Currently doing a GEO audit across all 8 clinics.

Last time we applied with a different idea, and we kept at it — built the product, iterated on the hardware and the AI layer with real paid customers, launched and went viral on X with 380K views. We even raised $400,000 from South Park Commons for it. Ultimately decided to pivot because of increased regulatory risk around AI for kids. The GUARD Act just passed the Senate Judiciary banning under-18s from AI companion products. We learnt the importance of having a good read of the regulatory environment — especially with minors as your users.

We participated last year in South Park Commons Founder Fellowship in Summer 2025.

Nischal has been on GLP-1s for the last five months, and they had a life-changing impact on his health. That made us curious about the second-order effects of these drugs — and led us to med spas. They are a rapidly growing category that increasingly offer GLP-1s alongside aesthetics. We started going door to door across SF with personalized letters and chocolates, asking busy clinicians for 15 minutes to understand their biggest problems. We spoke with 28 med spa owners and operators. Overwhelmingly it became clear — they want more qualified patients, they do not know how discovery is changing with AI, and they miss high-intent calls when the front desk is busy or closed. We also spoke with industry KOLs & marketing agencies to understand how med spas currently acquire customers and how to sell into the category. Those conversations convinced us that med spas need an AI-native growth engine that helps them get discovered, capture demand, and turn it into booked appointments. We know people need this because we have strong demand in our pipeline.

Our competitors are the companies that help clinics get found online — legacy local-marketing companies like Yext ($450M ARR) & Birdeye ($150M ARR), and horizontal GEO platforms like Profound & AirOps. Legacy local-marketing companies have multi-location infrastructure, but they were built for local SEO and review management, not GEO. They can manage listings and reviews, but they are not built to automatically identify AI-search gaps and push location-specific improvements to the clinic's website agentically. Horizontal GEO platforms are mostly built for brand-level visibility. When you sign up, they ask for your website and competitors. But med spas are local businesses. A multi-location clinic does not compete as one brand; each location has a different local competitive set, different treatments, etc. A location in SF and a location in San Mateo need different GEO strategies. Existing GEO platforms would work for SaaS or e-commerce, but not for multi-location clinics. Another insight that differentiates us from everyone is our intent capture loop. We integrate with front desk calls that capture the real questions patients ask about treatments, objections, etc. Those questions feed back into our GEO system, so each clinic's AI-search presence improves from what real patients are asking in natural language, not static assumptions.

Price per location per month (X) = $3,000 (medspas currently spend $3–10K/mo on marketing) Number of targetable clinics today (Y) ≈ 200,000 TAM = X × Y × 12 = $6B / year in the US today. Our TAM is growing rapidly because of the underlying market's tailwinds. For example, the US med spa industry alone is growing 13–15% annually because of GLP-1s.

1. AI growth agents for Personal Injury and Immigration Lawyers 2. Open-Employee — an open-source alternative to Ramp Inspect / Stripe Minions but with OKRs as first-class citizens instead of tasks.

We've raised money from Tier 1 funds like Sequoia. We have raised from great accelerators like SPC. But we still think there's no place like YC. The acceleration in three months during the batch, the YC network — there's nothing like it. That's why we want to join YC. And we will keep applying until you don't let us in. No matter how much money we've already raised for the company. - Yes, friends have encouraged us to apply to YC. - No, we haven't been to any YC event.

PG Essays.