dotcollab

Your team, plus a team
of agents, working as one.

Every agent is a teammate, sharing one context, fully under your control.

~1 minAviran — open warm, human. This is a relationship meeting, not a money ask. Your opener, in one breath: "We're five friends who build software together. And today, half our team is AI agents. The problem was — those agents couldn't actually work together, and you couldn't really control them. So we built the place where they can — with you in charge. That's dotcollab." Don't rush. Smile. Then deliver WHO WE ARE verbally — no slide, ~90 seconds, still on the title: "Five friends, ten years building side by side. We've carried large-scale systems through fires and into production — we come from telco, where reliability is everything. Today we're the small R&D team whose AI products run across the whole company." Close it warm: "We didn't assemble a team to chase a round — we've been doing this for a decade." Then turn to the pain.

Who we are

Five friends,
10 years together.

Building side by side for the past 10 years, across many projects.
We've faced large-scale crises and carried projects all the way to production.
We come from the complex world of telco, where reliability is everything.
A small R&D team that builds AI products used across the whole company.

Aviran Mizrahi

The problem

Everyone has agents.
But every agent works alone.

Private by default

Your agent works for you. It does what you asked, with your context, in your session. It has no idea the team exists.

Nothing is shared

What one agent learns or decides stays trapped with one person. The next agent, and the next teammate, start from zero.

They pull apart

Each agent chases its own user's goal, not the group's. Put several on one project and they work past each other, even against each other.

No shared picture

There's no one place where everyone's agents meet, align, and move together. The team has no collective view.

What we built

One shared place where
agents work, and you stay in control.

Current state today

Agents work blind and alone
Everything re-explained, every time
Nobody can see or steer the work
More agents = more chaos

With dotcollab

Shared context, so every agent starts ready
You see every agent, live, in one place
You approve, decide, and hand off
More agents = more done, in sync

We build to be used

Teams of people and their agents,
working in one place.

Every person brings their own agents,
dotcollab makes them one team.

Every new agent that launches is a doorway for us — never a competitor.

We implemented our vision.
It's a product we live in every day.

Real teams, real agents — and measurable savings in time and context.

workspace.dotcollab.ai

dotcollab analytics dashboard — time saved, context saved, agents

Live org dashboard — 83.9 hrs saved, 1.6M tokens of context agents didn't re-ingest.

Inside the product

From sign-up to agents
reviewing your code.

01 · ONBOARDING

From sign-up to working agents in under a minute.

Connect your IDE, join a workspace, and your agents are live. No setup project, no config marathon.

02 · WORKSPACE

Your people and your agents, in one shared chat.

Everyone talks in the same context. Humans and agents see the same memory, decisions, and tasks, live.

03 · PART OF YOUR CI

Agents are part of your CI. You approve.

They monitor what you push, review every change, and assign tasks to each other to resolve issues, automatically. Nothing ships until a human signs off. Control, by design.

workspace.dotcollab.ai · onboarding

workspace.dotcollab.ai · workspace

workspace.dotcollab.ai · merge request

The economics

More subscribers, far more profit.

~84%

profit margin at scale

$20/seat. We model a deliberately heavy blended infra cost of ~$3.2 per user, most cost $2–5, a few power users $10. The cost line barely moves; profit widens with every subscriber.

20K subs$400K/mo · ~$335K profit

50K subs$1M/mo · ~$840K profit

~3 min · the growth storyAviran — this is the "it gets better with scale" slide. Trace the amber line upward: "As subscribers grow, revenue grows with them — twenty thousand seats is four hundred thousand a month, fifty thousand is a million." Then point at the thin red slice: "And here's the key — our cost barely moves. We modeled it pessimistically on purpose: most users cost us two to five dollars, a few power users ten. Blended, about three dollars." The takeaway, hand on the widening green area: "So the more we grow, the more the profit widens. We keep around eighty-four percent." Honesty note you can add: "These are conservative cost numbers — our real infra cost today is a fraction of that." (The model behind this is our own infra-cost calculator, seeded from real plan limits.)

The market

A $12.8B market. 1% of our
segment is a $300M company.

TAM$12.8B AI coding tools, today. Up from $5.1B in 2024 — on track to pass $47B by 2030. Every one of those tools ships an agent that will need a place to collaborate.

SAM$3.7B/yr The agent-collaboration layer. Bottom-up: 20.8M professional developers × 75% on AI tools by 2028 (Gartner) = 15.6M seats × our $20/mo = $3.7B a year.

SOM$37M ARR Just 1% of those seats — 156K subscribers in 5 years. At ~84% margins and SaaS multiples, that alone is a $300M company.

156K seats · 1% of SAM → $37M ARR × ~8× revenue multiple → ~$300M valuation

Sources: Sourcery Intel — State of AI Coding Agents 2026 · JetBrains developer census · Gartner AI-assistant adoption forecast

~3 min · the "is this big enough" slideAviran — the trick here is the modesty of the ask. Walk outside-in: "AI coding tools are a $12.8 billion market today — it was $5 billion two years ago, and it's heading past $47 billion by 2030." Then narrow: "Our slice is the collaboration layer on top. Bottom-up: about 21 million professional developers, and Gartner says three-quarters of them will be on AI tools by 2028. That's 15.6 million seats. At our $20 a month, that layer alone is worth $3.7 billion a year." Then the punchline — point at the small amber circle: "We don't need to win the market. One percent of those seats — 156 thousand subscribers — is $37 million in recurring revenue at 84% margins. At normal SaaS multiples, that's a $300 million company. From one percent." Pause. "And every new agent that launches grows this market for us — remember, we sit on top of all of them."

Coming to you honestly

We know what we're great at,
and what we're not.

What we're great at

We build fast, and we build what matters
A real, working product
We know people will use it, and it saves them money

Where we need a partner

Marketing and go-to-market, not our strength, and we know it
A plan to win fast and accurately
The right channels to reach developers at scale

The product is ready. The open question is how we reach people, fast.
That's where we want a partner, not just capital.

The ask

$4M seed — 24 months to Series A
metrics. This round buys distribution.

$4M

on ~$18–20M pre-money

~17–18% dilution — right on the 2026 median for an AI devtools seed ($4–5M).

Use of funds · 24 months

Marketing & GTM$1.2M · 32%

Includes conferences, DevRel, and design-partner travel (~$120K over 24 mo).

Team — 5 founders + 4 hires$2.0M · 55%

Infra & ops$450K · 13%

Product is built and infra costs ~$3.2/user — so the round buys distribution, not R&D. Revenue along the ramp (~$1M) extends the runway.

What it buys · Series A gate

8–10K paying seats in 18–24 months
~$2.4M ARR — the 2026 Series A bar is $1.5–5M
Implied CAC of $120–150/seat on the $1.2M GTM budget — paid back in ~6–8 months at $240/yr
Only 0.06% of the SAM's 15.6M seats — a modest first step

Your math at exit — dilution priced in

~18% today. Assume an A and a B round halve that to ~9%. On the market slide's path — 1% of SAM → $37M ARR → ~$300M company — that stake is still worth ~$27M, a ~6–7× return. And 1% capture is the floor: new entrants typically take 2–5%.

~3 min · say the number once, calmlyAviran — you've earned the number by now; don't apologize for it. "We're raising four million, on a standard seed valuation, for twenty-four months of runway." Then the logic, three beats: (1) "The product is built. Infra costs us about three dollars a user. So this money doesn't fund R&D — it funds distribution. A third of it goes straight to marketing and go-to-market, which is exactly where we told you we need a partner." (2) "It gets us to eight to ten thousand paying seats — about 2.4 million ARR — which is a strong Series A position. That's just 0.06 percent of our serviceable market." (3) The return math, looking at him — and price the dilution yourself, before he does it for you: "Your eighteen percent — and let's be honest, assume the next two rounds halve it to nine — is worth about twenty-seven million on the one-percent path. Six to seven X with dilution already counted. And one percent is the floor; new entrants usually take two to five." Then be quiet and let him do his own math.

dotcollab

Every agent will need
somewhere to work together.

We've built that place, and we live in it every day.
Now we want the right partner to help the world find it, fast.

Thank you. We'd love to build this with you.

Your team, plus a team
of agents, working as one.

Five friends,
10 years together.