Early Access · 2026

The scoping engine
service businesses
run on.

A client gets a link. They talk through the project. A structured scope lands in your dashboard, ready to review. No back-and-forth, no missed requirements, no lost deals.

Investment opportunity See the product
2wk
traditional scoping
manually, no tools
8hr
with generic AI tools
calls & back-and-forth
18min
with Monocular
structured engine
$3k
monthly time saved
per firm
417k
US service firms
in V1 TAM

Built & Live

Everything your firm needs.
Nothing it doesn't.

AI Intake: Live session
C
We need a mobile app (iOS & Android) plus a web admin dashboard. Targeting a Q4 launch.
M
Got it. Is this greenfield, or are you migrating an existing web product to mobile?
C
Greenfield. We have a SaaS platform but mobile is new territory. ~3 user roles.
M
Understood. What's the primary action each role takes, and which is the launch-critical one?
Scope generated · 16 deliverables · 4 milestones · 14 min
Avg Scope Time
18 min
takes 2+ weeks manually, 8 hours with generic AI
AI Cost Per Scope
$ 0.05
margins expand significantly
at volume
In production
Full-stack SaaS
Multi-tenant · Role-based access · Webhooks verified
Stripe Clerk Supabase Vercel
Powered by
Enterprise AI
Frontier models · intake, generation, live market-rate research & attachment parsing
Prompt caching · cost-optimized
No training on your data
Early access
Invite-only · Design partners first · Full product access
Branded PDF Export
Designed scope document, not a text dump · Structured, paginated, one click
Client Review
Approve or request changes · Revision loop built in

The Problem

Service firms lose deals in
the gap between interest
and proposal.

Discovery calls end. Notes sit in inboxes. A junior writes the scope from memory three days later. The client has already moved on.

"We lose 20–30% of qualified leads because our scoping process takes too long. By the time the proposal is ready, the client's already talking to someone else."

Design agency principal, 12-person firm, US

"The real cost isn't the proposal time. It's the six-figure projects that never got a chance because we couldn't respond fast enough."

Monocular user interview, April 2026

The Real Cost

Every day without a scope
is a deal the competitor
wins instead.

Service firms don't lose deals at delivery. They lose them in the gap between interest and proposal: a gap that averages five days and eight unbillable hours.

Without Monocular
Monday
Prospect emails. Discovery call needs to be scheduled: 3 back-and-forth emails to find a time.
Tuesday
60-minute discovery call. Notes in Notion. Junior starts writing the brief from memory at 5pm.
Wednesday
Senior rewrites the brief. Client gets 3 clarifying questions back. Competitor's proposal just landed in their inbox.
Thursday – Friday
Two revision threads. Revised pricing. Proposal finally out Friday. Deal is already half-lost.
5 days · 4–8 unbillable hours per scope · senior time on paperwork
With Monocular
Monday:30 seconds
Intake link sent. No scheduling, no friction. Client completes the guided chat at their own pace.
Monday:15 minutes later
AI conducts structured discovery. Scope draft appears: 16 deliverables, 4 milestones, risk flags, pricing range.
Monday afternoon: 20 minutes
Senior reviews, edits, approves. Proposal sent same afternoon. Client gets it before anyone else does.
Same day · Under 1 hour total · senior focused on the work that matters
71×
At 3 scopes/week (typical for a 10-person agency), Monocular recaptures 52 hours per month of senior time. At $150/hr that's $7,800/month returned to billable work, client relationships, and delivery. The Studio plan is $109/month. One saved scope session covers the next two months of subscription.

The Product

Intake to scope.
Automatically.

A complete AI-powered pipeline. No workflow changes. No client training. It works the way firms already work.

01 / Create

Generate a link

Pick a project template: web, mobile, brand, e-commerce, and more. Pre-fill client details. Live in seconds.

02 / Capture

Client chats, AI listens

No account, no app download. The client gets a link, opens it on any device, and has a 15-minute conversation. The AI probes for everything specific to that project type — budget, timeline, stack, risks.

03 / Generate

Scope appears

Deliverables, milestones, pricing range, risk flags, assumptions, recommended tech stack, out-of-scope — enriched with live market-rate research and any reference docs or screenshots the client uploaded. All structured and editable.

04 / Close

Edit, share, send

Edit inline. Share a client review link. Export PDF. Send directly. Track status from draft to won.

Why Monocular

Not a chatbot.
Not a form. Not a
prompt wrapper.

"Can't I just use a general AI tool for this?" Yes, and your team probably already tried it. Here's why the output ended up getting rewritten from scratch anyway.

Generic AI
General-purpose chat tools
Workflow Tools
Proposify · HoneyBook · Dubsado
Monocular
Purpose-built scoping engine
Multi-turn intake conversation (async, no scheduling) Manual
Domain-expert AI persona (probes scope-creep, project type, integrations)
Configured with your firm's rates, risk flags, and project templates Partial
Structured scope output: typed deliverables, milestones, risk flags, pricing Freeform Manual entry
Ingests client reference docs, screenshots & live market-rate data directly into the scope
Mandatory human review gate before any client-facing output
Full workflow: draft → review → edit → approve → PDF → send Post-manual
Per-run trace and audit log (every AI call recorded)
Vertical expansion via config: same engine, new domain
Data flywheel: output calibrates to your firm's history over time
The sophisticated objection

"What about an AI chat tool with a memory system? Couldn't I give it a knowledge graph of my firm?"

Memory is recall. Monocular is a system. An AI chat with memory can remember your rates and past scope docs. What it can't do: detect when a 15-minute conversation has collected enough to generate (that's a classification problem, not recall); enforce that a human reviews the output before it reaches a client (that's a workflow, not a prompt); isolate each firm's data from every other firm (that's multi-tenancy); or record what happened after: did pricing hold, how much did the reviewer edit, was it won? (that's the data flywheel).

You'd spend weeks stitching these together, and you'd end up with a Monocular-shaped system that still needs domain tuning, trust infrastructure, and PDF export. We built that system so firms don't have to.

One Engine · Every Vertical

Same orchestrator.
Any domain, any firm.

A vertical = persona prompt + extraction schema + risk library. Swap the config, not the code.

Live · V1

Web Development

CMS vs custom, auth, integrations, content migration, hosting, SEO, mobile, cookie compliance, scope-creep triggers

V2 · Planned

Legal

Matter type, jurisdiction, parties, deadlines, privilege concerns, billing structure, complexity flags

V2 · Planned

Healthcare

Specialty, visit type, EHR system, HIPAA posture, patient flow, referral sources, payer mix

V2 · Planned

Architecture

Project type, site conditions, regulatory requirements, design intent, phasing, CAD deliverables

V2 · Planned

IT Consulting

Infrastructure, integrations, security posture, compliance requirements, team size, migration risk

Frequently Asked

Every objection.
Answered honestly.

"I already use AI tools for scoping. Why do I need this?" +
General AI tools have no idea who you are. They don't know your rate range, your risk flags, your project types, or your assumptions. Every session starts from zero. The output is a blob of freeform text in a chat window: not in a dashboard, not editable inline, not one click from a PDF, not sendable to your client for review. Your team copies it to Word and rewrites it anyway.

Monocular is pre-configured with your firm's data. The AI persona runs your extraction schema, asks the questions a senior person from your firm would ask, and produces a structured scope document with typed deliverables, milestones, risk flags, and pricing within your configured rate range, landing in your review queue. The agency founder edits, not rewrites. That's the difference.
"Can't I just write a really good AI system prompt and get the same result?" +
To replicate Monocular with a system prompt you'd need to write: the intake persona (domain-expert with scope-creep detection and dynamic probing logic); the extraction schema for 10+ scoping areas; the risk library keyed by project type; the orchestrator that decides when intake has collected enough to generate; the generation step that produces typed deliverables, milestones, and assumptions; the pricing guardrails with clamping to your rate range; the human review gate with confidence scoring; the workflow from draft to approval to PDF to client send; the audit trail on every AI call; and the multi-tenancy that isolates your data from every other firm's.

That's not a prompt. That's Monocular. The system prompt is the starting point. The engine, trust infrastructure, and workflow are the product.
"Our discovery process already works. Why change it?" +
Good discovery processes work on the 30–40% of projects that are complex and novel, the ones that genuinely need a senior person in the room. Monocular handles the 60–70% that follow a pattern you've scoped a dozen times already. Shopify builds, SaaS dashboards, brand refresh and web presence. Your senior people have done these. The question is whether they need to be the ones doing intake for all of them.

The worst case of trying Monocular on one project: you get a draft you don't use and it cost you 30 seconds to send a link. The expected case: a draft your team spends 20 minutes reviewing instead of 8 hours writing, and your senior person focuses on the projects that actually need them.
"You're only in software development. How does this work for law, medicine, architecture?" +
Verticals are config, not code. The orchestrator, intake engine, human review gate, and data flywheel are domain-agnostic. What changes per vertical is the persona prompt, extraction schema, and risk library — stored as data, not compiled into the engine. A legal agent probes matter type, jurisdiction, and privilege concerns. A medical agent probes specialty, EHR system, and HIPAA posture. Same infrastructure. Different config file.

The ROI compounds in regulated verticals. Attorney time runs $400–800/hr. One recovered intake session covers months of subscription. We start with web dev because it has the fastest feedback loop. The engine was designed for all of them from day one.
"Won't Proposify or HoneyBook just add a chat intake feature?" +
Eventually, maybe. But their data models, UIs, and sales motions are built for what happens after requirements already exist. Proposify's AI polishes prose in a proposal editor. HoneyBook's AI summarizes notes after a call. Neither product has a domain-expert intake persona, an extraction schema, a human review gate, or an orchestrator that decides when intake is complete. That's not a feature they're adding. It's a different product with a different entry point and a different trust posture.

Their product cycles run 12–18 months. The defense isn't feature velocity. It's deeply configured paying customers who have risk flags, rate ranges, project types, and 12 months of scope history that would take months to reconstruct elsewhere. Configuration depth is the switching cost.
"Won't better AI models eventually make this obsolete?" +
Better models make Monocular better, not obsolete. The AI layer is the replaceable part. What isn't: the labeled dataset that accumulates as a byproduct of every review. (generated, reviewer-improved) preference pairs, confirmed and dismissed risk flags, win/loss outcomes with pricing deltas. A firm 12 months in has a training set calibrated to their project types, vocabulary, rates, and outcomes. No competitor can buy that data, and it doesn't reset when a new model comes out.

The architectural principle: "Built solely on model quality, you start over with every model generation. Built on proprietary data, the moat ports forward." The model is the ingredient. The firm's data history is the recipe.
"What if the AI generates a wrong scope or bad pricing?" +
Nothing reaches a client without explicit human review. The gate is structural, not a best practice, not a reminder. Export and send are blocked until every review flag has been acknowledged by a named reviewer. Confidence scoring surfaces where intake was thin. Pricing carries a mandatory disclaimer and is clamped to the firm's configured rate range. The reviewer's name and timestamp are recorded.

If a scope is wrong after human review, that's a review failure, and the system is designed to make review failures visible before they reach clients. The design-partner phase exists precisely to find gaps in question sets, risk libraries, and edge cases. Any pricing incident during pilot immediately removes pricing from V1. Trust is built gate by gate, not promised up front.
"Does Monocular's AI train on our intake data or scope documents?" +
No. The AI infrastructure we use does not train on customer inputs or outputs by default. A Data Processing Agreement is available confirming this. Client intake conversations and generated scope documents are stored under the firm's Monocular account; the agency is the data controller and Monocular is the processor. The firm can delete any scope from their dashboard at any time.

Regulated industries (HIPAA, attorney-client privilege) are V2 verticals, built to those standards with per-agency retention controls, DPA templates, and provider terms confirmation. Not retrofitted after the fact.

Where We Are

Built. Billing.
Not bragging yet.

We're early, intentionally. The foundation is solid. The next 90 days determine what this becomes.

Product

Agentic engine live.

Multi-tenant SaaS with a fully agentic intake engine: orchestrator loop, Zod-typed tool registry, live market-rate research, file-attachment parsing (PDFs, DOCX, images), AI-recommended tech stack, per-run trace tables with a live internal trace viewer. Branded PDF export, client review workflow, Stripe billing. Invite-only early access.

Revenue infrastructure

Billing complete and verified.

Solo $49 · Studio $109 · Agency $219/mo. Annual available. Stripe webhooks wired, DB synced. Feature gating enforced per plan.

AI Cost

~$0.05–0.09 per scope.

Frontier models with prompt caching. Cost is variable and scales with usage; margins expand significantly at volume.

Next 30 days

Admin console + first customers.

Full billing management UI, usage enforcement, marketing website launch. Design partners being qualified now. ICP: US digital agencies 5–50 people.

Market Opportunity

Every firm that sells
before it builds
needs this.

We start with digital agencies: fastest sales cycle, sharpest pain, most vocal community. Then we expand the engine to every professional service vertical.

$4.2B
V1 TAM: Digital Agencies

417K US digital agencies, studios, and consultants. Scoping is a weekly recurring pain at every firm.

$29B
V2 TAM: Professional Services

Law firms, physician practices, architecture, accounting. Same repeatable intake problem. Larger deal sizes.

$1M
ARR at 840 customers

0.2% of US digital agencies at $109 avg/mo. Studio LTV ~$2,100. Agency LTV ~$4,200 at 5% monthly churn.

Why Now

The model makes this possible.
The timing makes it inevitable.

This wasn't buildable two years ago. Long-context LLMs capable of structured multi-turn discovery at $0.09/scope didn't exist. They do now.

Monocular sits at the intersection of three converging trends: AI that can hold a real conversation, service businesses that are capacity-constrained, and a new generation of founders who expect to run a 20-person firm like a 5-person firm. That's the window.

Internal positioning memo, Q2 2026

Roadmap

From scoping tool
to scoping engine.

The product ships today. The platform takes 18 months. Phase 1 has a clear proof point before Phase 2 gets funded.

Now
Agentic Engine: Live
Multi-tenant platform with a full agentic intake engine: orchestrator loop, vertical configs (web-dev live), market-rate research + file-attachment parsing (PDF, DOCX, images), AI-recommended tech stack, tool-use trace tables with a live internal trace viewer, branded PDF export, client review flow, Stripe billing. Usage limits enforced per plan. First paying customers.
Live
Q2 2026
Admin Console
Full billing management UI (plan switching, invoice history), team + usage dashboard, marketing website, design partner program launch. Target: 10 paying orgs.
In Build
Q3 2026
Platform Hardening + Integrations
HubSpot, Slack, Zapier. Custom branding. Version history. Approval workflows. Target: 100 paying orgs, $120K ARR.
Next
Q4 2026
Public API: The Scoping Engine
REST API lets CRMs, proposal tools, and custom stacks embed Monocular's intelligence. Per-scope API billing on top of SaaS revenue.
Next
Q1 2027
Voice Intake
Voice intake via transcription and voice synthesis, plus live screen-share walkthroughs. (Document and image attachments — PDFs, screenshots, brand assets — are already live today, parsed directly into the generated scope.)
Planned
2027
Vertical Expansion
Law firms, physician practices, architecture & engineering. Same engine, vertical-specific intake schemas and compliance handling.
Planned
2027+
The Delivery Layer
Monocular moves past intake. The approved scope becomes a living contract: milestone tracking, scope drift detection, change order generation, client communication, and re-engagement triggers. The operating system for the full client engagement lifecycle, from first conversation to final invoice.
Vision

Business Model

Simple SaaS.
Predictable revenue.

Per-seat subscription with usage limits that create natural upgrade pressure. Annual plans at 2 months free. API billing in Phase 4 adds a second, higher-margin revenue stream.

PlanMonthlyAnnualSeatsScopes / moLTV @ 5% churn
Solo$49$490120~$941
Studio$109$1,090375~$2,092
Agency$219$2,190UnlimitedUnlimited~$4,202

Investment

Raising $2.5M to go
from engine to market.

The hard part is done. The intake engine is live, billing is wired, and the first vertical is shipping. We need capital and operators to close the first 100 firms.

We want your money.
We give you back your time.

So service firms can focus on the things that actually matter: the clients, the craft, the relationships. Not the paperwork between interest and proposal.

Use of Funds

  • 50% Go-to-market.GTM hire, design partner program, content, agency community presence
  • 30% Product.CTO + AI/ML engineer, admin console, integrations, API v1, vertical expansion
  • 20% Infrastructure & Ops.Production hardening, compliance groundwork, legal, visa sponsorship, AI cost optimization

Burn & Unit Economics

  • ~90% gross margin.AI inference and payment processing costs are minimal relative to subscription revenue, improving with scale
  • Lean infra model.Variable costs scale with usage, not headcount; margins expand significantly at volume
  • Break-even: ~$70K MRR.Achievable well inside the investment horizon at target growth rate
  • $2.5M: 18–24 months runway.Full founding team, built to reach Series A milestones before needing to raise again

Series A Triggers & Exit

  • Series A conditions.$1.5–3M ARR, 2–3× YoY growth, NRR 100–110%, payback under 12 months
  • Series A size.$5–8M defensible at those metrics (Proposify $11M rev; Qwilr $7.25M; Lawmatics $12.5M)
  • Acquirer profile.Salesforce, HubSpot, PandaDoc, Ignition: all processing high proposal volume, none with structured client-facing intake AI
  • Exit range.5–8× ARR at $3M = $15–24M floor. Platform exit at $10M ARR with data moat = $50–80M

The Team

One founder.
Building room for four more.

The full-stack product is built, billing is live, and the agentic engine is in production. We're not looking for contractors. We're looking for co-builders who see what this becomes.

PA

Prabhu Avula

Founder · CEO · Builder

Solo-built the entire Monocular stack from scratch: Next.js 16, a fully agentic AI intake engine with tool-use tracing, Supabase multi-tenant database, Clerk auth, Stripe billing with webhook sync, Inngest background jobs, PDF export, and a client review workflow. The product is live in production. Pre-team, post-build. Now focused on going from engine to market, looking for the founding team to do it with.

Next.js 16 TypeScript AI APIs Supabase Stripe Inngest Full-stack SaaS

Open Roles

We're hiring the founding team.

Founding Equity · Technical

CTO / Founding Engineer

Takes full ownership of the technical architecture as Monocular scales from MVP to platform. Deep in TypeScript, distributed systems, and AI/LLM APIs. You're not building features for the founder. You're building the platform the company runs on. Co-founder equity for the right person. The codebase is clean, the architecture is intentional, and the technical decisions are documented.

Founding Equity · AI/ML

AI / ML Engineer

Owns the data flywheel end-to-end: turning reviewer edits into preference pairs, risk flag outcomes into classifiers, win/loss data into pricing regression. The trace tables and labeled dataset are live today. You build the ML layer on top. Ideal background: DPO fine-tuning, structured extraction, RAG over firm history, or production-grade LLM evaluation. This is the role that builds the moat.

Senior · GTM

Head of Growth / Sales

Closes design partners, converts them to paid, and builds the pipeline to the first 100 firms. Agency or studio background strongly preferred. You know the ICP from the inside, not from a persona doc. Equity plus commission on a clear ramp. The product is ready, the pricing is proven, and the ICP is sharp. We need someone who can open doors and walk through them.

Advisory Equity · Vertical Expansion

Domain Expert Partners

Deep in legal, healthcare, architecture, or accounting. Not a full-time role yet. Advisory equity in exchange for co-building a vertical config (persona prompt, extraction schema, risk library for your domain), introducing the first 3 design partners in your network, and advising through V2 launch. This is category ownership at the ground floor of a platform that will serve your entire industry.

Get in Touch

A real company at a
real inflection point.

If you write early checks into AI-native vertical SaaS, or you're a co-founder who sees what this becomes, let's talk. We move fast.