// Case Studies// Vertical AI · Real Estate

NYOA

A Claude plugin and SaaS platform built with a Nashville realtor as co-creator, designed to replace the stack of subscription tools a real estate agent stitches together every week.

// Nashville
ClientAnn-Riley CaldwellSimpliHOM · Nashville
EngagementCo-creator + Operator
SurfacePlugin · SaaS · Marketing site
StatusLive · nyoa.tech

What was getting in the way?

The average residential agent pays for somewhere between four and ten subscription tools — a CRM, a listing-copy generator, a content scheduler, a market-report builder, a transaction tracker, a separate AI assistant that promises to do all of the above. None of them know each other. Most carry a recurring monthly subscription. Most go unused after the first thirty days.

Ann-Riley Caldwell is a Nashville realtor with SimpliHOM. She's also a licensed California attorney with a financial planning degree — the rare agent who's comfortable evaluating tools on their actual merits instead of their marketing. We'd been working together on AEO content for her brand Not Your Ordinary Agent when she started sending over screenshots of competitors' AI tools and asking why they all sounded the same.

The answer was that they were all wrappers — generic prompts running on generic templates, producing prose that read like every other agent's prose. The thing missing wasn't AI. It was voice, context, and a workflow that didn't require an agent to learn yet another piece of software.

The decision

Anthropic had just shipped Claude plugins — bundles of skills and slash commands that install directly into an agent's Claude account. No app to maintain. No server bill. No seats to count. The marginal cost of running NYOA, once built, would be zero.

That made the pricing question easy: a one-time price, paid once, forever. Less than the referral fee on a single closed transaction. Cheaper than two months of any competitor. And structurally impossible for a SaaS-funded competitor to match without rewriting their business.

The harder question was what to put inside it. Most "AI for realtors" tools ship five generic prompts and call it a product. We wanted something that felt like Ann-Riley had hired a chief of staff who already knew her business.

Every realtor's AI sounds the same. The thing missing isn't AI — it's voice, context, and a workflow that doesn't require learning another piece of software.

Working note, NYOA discovery, Q1 2026

How we built it

The plugin shipped with fifteen Claude skills, each one mapped to an actual job an agent does in a week. Ann-Riley supplied the friction points: which emails she rewrites three times, which listings she audits before posting, which testimonials she keeps forgetting to ask for. We turned each of those into a slash command.

The voice work was the unlock. Most AI tools ask for "your tone" in a dropdown — friendly, professional, luxury. NYOA captures it from real samples: past emails, listing descriptions, testimonials. Output sounds like the agent, not like a press release. That single difference is what made early users stop comparing it to ChatGPT.

/listing-copy/listing-audit/listing-presentation/buyer-seller-comms/offer-analyzer/aeo/social-content/testimonial-engine/daily-ops
Claude SkillsPlugin runtime
Next.js 15 · App RouterMarketing + platform
SupabaseAuth · Postgres · RLS
Stripe CheckoutOne-time + free tier
VercelDeploy · preview · edge
Anthropic APIPlatform inference
ResendTransactional · digest
GitHub ActionsCI · PR previews

In the field

The plugin works beautifully for Claude-native agents — but the addressable market for "agents who already pay for Claude Pro and know how to install plugins" is small. To reach the other 99% of realtors, we built a SaaS platform that wraps the same skills in a purpose-built web UI.

Same brand. Same intelligence. Two tiers: plugin for technical users who want NYOA living in their Claude account, platform for everyone else. The platform is currently shipping at a rate of roughly one production deploy per day — agentic copilot sidebar, weekly digest emails, one-click actions on a "needs attention" dashboard, native voice calibration during onboarding.

Both share the same underlying skill library. A fix to the listing-audit skill ships to plugin and platform users simultaneously. That's the part no incumbent SaaS can copy — they'd have to throw out their stack.

15Skills shipped
30+Production PRs / month

How it unfolded

Jan 2026
DiscoveryAudit of Ann-Riley's existing workflow and tool stack. Identified fifteen recurring tasks worth automating and the brand voice samples needed to anchor output.
Feb 2026
Plugin buildSkill-by-skill development against real listings, real client emails, real offer analyses. Ann-Riley operated as live QA — every output measured against what she would have written herself.
Mar 2026
Marketing site + launchnyoa.tech goes live. Plugin published. First public posts hit the realtor AI corner of X and LinkedIn. Pricing locks in — one-time, no subscription, no per-seat.
Apr–May 2026
SaaS platformTwo-tier strategy formalized. nyoa-platform repo expands from marketing-only into full app: Supabase auth, Stripe checkout, agentic copilot, weekly digest, voice onboarding. Free Starter tier opens. Ann-Riley becomes first paying user on both surfaces.

The takeaways

The product distributes itselfAnn-Riley brings the credibility a builder can't manufacture — every realtor who hears about NYOA hears it from someone who actually uses it for their listings. The endorsement isn't a testimonial. It's a working agent.
The plugin model changes the mathZero marginal cost means a one-time price is structurally possible. Competitors charging recurring monthly subscriptions are defending a cost structure that no longer exists. That gap widens every quarter.
Voice is the only real differentiatorAnyone can wrap a prompt. Almost nobody is calibrating output against real samples of how a specific person writes. The agents who use NYOA stop sounding like every other agent on the MLS.
Plugin as wedge, SaaS as platformThe plugin reaches Claude-native power users at a low price point. The platform reaches everyone else at higher LTV. The same skill library powers both — meaning improvements compound across surfaces.

Got a vertical that's drowning in SaaS?