Notto / software company · est. 2026

Vertical software for service businesses.
Built deep.

Generic SaaS makes every small service business bend to the same template. We build the other way — one vertical-complete product at a time, with agents doing the domain-expert work that used to cost a year of engineering per niche.

Approach · Agent-native, human-in-loopFor · Service businessesStatus · First product in pilot
01 / THE GAP

Vertical SaaS used to be too expensive for niches. Not anymore.

Observation · 01

Enterprise SaaS doesn't scale down to small verticals.

ServiceTitan works because home services has 100,000+ US contractors. Most service niches number in the low thousands — too small to amortize per-customer config, too deep for horizontal tools. The middle is empty.

Observation · 02

Agents collapse the cost of going deep.

What used to take a specialist-consulting team months to configure per shop — playbooks, workflow logic, integration wiring — a domain-grounded agent now handles per session. The economics flip.

Horizontal SaaS · one shape for all

generic intake
config → (per-shop guess)
manual glue work

Vertical-complete · built for the trade

domain intake
agent handles the shape
team keeps how they work
02 / APPROACH

How we build a vertical in months, not years.

STEP · 01

Retrieve

Domain intake. We build the knowledge base with the trade — playbooks, routing logic, edge cases — so the agent speaks the work. Every answer is traceable to a source.

STEP · 02

Plan

Agent composition. The runtime picks the right workflow for the job in front of it and structures the output consistently, no matter how the work arrives — online, phone, or counter.

STEP · 03

Operate

The shop-floor app. Owners, teams, and customers each get the surface they actually need. Humans approve every customer-facing action. Nothing that requires IT.

03 / WHY IT WORKS

How the platform pays off.

01

One runtime, many verticals.

The agent orchestration, retrieval scaffold, and UI primitives are shared. Each new vertical reuses most of what we built before — which is where the economics come from.

02

Owner sees what was invisible.

Every Notto product ships with analytics from day one — where the money leaks, where the time goes, which parts of the workflow break. No separate BI stack.

03

Team keeps how they work.

We integrate read-only with whatever the shop already uses. No rip-and-replace. Software that joins the family, not replaces it.

04

Honest about AI.

Agents do the diagnostic and routing work. Humans approve every customer-facing action. Every decision is logged and reviewable. No autopilot on someone else's business.

04 / VERTICALS

Shapes we're mapping. One is in pilot, the rest are in scoping.

01Bars & cafésRES · POS · STOCK
02RestaurantsMENU · ORDER · SHIFT
03Local shopsINV · RESTOCK · CLIENT
04Salons & studiosBOOK · CRM · RECALL
05Fitness studiosCLASS · MEMBER · WAIT
PROFILE · BAR

A different agent graph per venue, not per template.

A bar build would ground retrieval in the venue's menu, supplier catalog, POS history and house rules. Agents would handle reservations, stock-outs, and shift hand-offs — assembled for this venue, not configured from a template.

Retrievalscoped / venue
Agentsreserve · stock · shift
UI surfacefloor console
Human in loopowner approval
05 / ARCHITECTURE

One runtime. Many verticals.

Every Notto product runs on the same three-layer stack. The adaptive shell renders what the current engagement needs. The agent layer plans, delegates, and recovers. The retrieval layer is scoped per tenant with every answer traceable to a source.

This is the shared asset. It's why we can ship a new vertical in months instead of years — and why the economics work at shop-family prices.

notto · runtime blueprint

UX · ADAPTIVE SHELLAGENTS · PLAN · DELEGATE · RECOVERRAG · INDEX · PROVENANCE · EVAL
Notto · 2026

Say hi early.

Our first product is in pilot now. Vertical #2 is in scoping. If you run a service business, think your industry might be next, or want to help build — we'd rather start talking early than polish a pitch.