On the Brink of $1M per Year
Investor Report Oct 25
Actual Juma (ex Team-GPT) investor report, sent on 13 Oct 25.
Hey!
Our best month ever in terms of ARR, just shy of our most important milestone ever. Depending on how we count it and when you are reading this, we’ve hit $1M ARR. Yay!
Yet we don’t feel like a million bucks.
We’ve got so much potential and somehow we can’t seem to properly unlock it. Yet.
As of 1 Oct 2025:
ARR: ~$969,288 (+$137,155 vs last month, 16% MoM growth)
# Users: 67,606 (+7,264 vs last month)
Burn rate (approx): $145,424.05
Runway: 17 months
# People: 20 full time + 3 part-time (+ 1 → Simona is back!)
I. Built for everyone, optimized for no one
I think the biggest issue of Team-GPT from Day 1 has been that we sell to everybody, or better phrased: we don’t know who we sell best to.
This has undermined every decision we’ve made (because we never commit fully).
It’s the reason why some investors couldn’t build conviction around our vision.
It’s the reason why clients don’t buy our product (they don’t understand how we are useful to them).
It’s also why people say we are ‘just a GPT wrapper’, even though we’ve got lots of IP on top the LLMs.
We’ve been brutally aware that we are ‘built for everyone, optimized for no one’.
But as CEO, I never had conviction in the right focus for the company to just say ‘GO’.
I only had conviction in our team’s ability to make awesome things happen… fast.
Btw I’ve always been upfront about it: ‘We need 30-50 brilliant people working together to build an incredibly impactful B2B AI product’. N.B. We are at 24.
Together we’ve made lots of steps towards understanding product-market fit (PMF) and where it will happen for us. We’ve invested more than almost any other startup at our stage in analyzing our data and understanding it well.
II. Mission should never change
Jon Callaghan (co-founder of True Ventures ) said:
‘At Seed stage, your company and product will look nothing like it will at year 10. The only thing which will never change is your Mission.’
Our Mission never changed:
We are on a mission to help people adopt AI for their work.
nor the Vision*:
We are doing this by building a future-of-work B2B AI product, which puts humans at the center and promotes collaboration, privacy, and agentic workflows.
*the exact phrasing is work-in-progress.
III. Optimized for learning
To find the fastest route to fulfilling our mission, we’ve taken an extremely data-driven approach.
Our process:
We optimize for learning
The product is the playground where we experiment
Extreme PMF + global impact is the North star
$$$ value (ARR) and usage metrics are the ‘weights and biases’
It’s been almost 2.5 years since we started ‘learning’ and the results are becoming clearer and clearer.
IV. 1-million-dollar-weighted insights
I think reaching $1M ARR is the key here.
Not because it matters in the grand scheme of things. We are still making ‘peanuts’ and we still can’t support our 24 (brilliant) people without external investment (for which we are eternally grateful to our investors). If we were bootstrapping, as of today, we would be able to afford ~10 people full-time.
But $1M ARR is the key because we’ve gone from 0 to 1:
We’ve built something from nothing
We are the go-to software for a small group of people
If we stop existing today, thousands of people will actually suffer
We matter!
~1,000,000 USD per year are paid to our company bank account because of what we’ve built.
These dollars weigh.
They weigh on the data when we make decisions.
When analyzing our data through the lens of a million bucks, we’ve learned that:
1. Team-GPT main value prop -> Projects (aka assistants/agents)
First and foremost, our main value proportion are the Projects (60% of happy, dollar-weighted users use them).
Projects are own main organization unit + Context/knowledge unit.
Why? Because they solve the main issue with AI → explaining the same context again and again. This includes:
Project Knowledge - Upload docs, connect Google Drive/SharePoint
Context Library - Reusable context across the
Assistants (knowledge bots) - Project / department-specific assistants (HR, Sales, Legal).
Some users use Projects as folders, most use them like Assistants. From this month they can use them as Agents (see ‘automations’)
2. Multiple models (Claude, GPT, Gemini, etc.).
Biggest value prop in sales calls. Head-to-head with ChatGPT and Claude, this is our only benefit, which becomes clear straight away. In the context of our main value prop (projects), this means: “Before users underhand they actually care about Projects, they care about model availability and no vendor lock-in.”
3. Enterprise Control & Security
Our 2 biggest contracts ($50K+ ARR each) have been signed because of this. We’ve got many open conversations about it and it seems to be a place where we can position ourselves long-term.
These users, however, will NOT help us build a top notch product TODAY.
I think that’s the real reason why successful products like Slack, Notion, Miro, etc. move to enterprise last. We are forced to do the same. But we will try to close all deals that come our way before that.
P.S. These users care about:
Private instances
Compliance (SOC2, HIPPA, ISO)
Zero data retention options
Permissions and governance functionalities
4., 5., 6. And the rest are all secondary
We believe the combination of our other benefits is valuable. Yet, $-weighed, everything else about our product is much less significant than the above:
Pages → 9-20% of users that matter (depending on whether you care about $ weight, ICP weight, activation weight, etc.)
Prompt library → ~15% of users care
Tools → nobody cares (with the notable exception of Prompt builder)
Integrations → valuable but not enough (because we don’t have many integrations yet). We understand that to build a successful B2B product, we will need to connect to all other major software
We kind of know what features and benefits matter.
But who cares?
V. Approaching Marketers (our ICP) with conviction
Marketers care.
Experimentation + data showed us that we should narrow down to the marketing function.
We did this in June 2025.
Some called it a ‘pivot’. In reality we simply changed our messaging to marketing with 0 changes to the product. We showed commitment on the outside to validate if this is the right decision to make.
We’ve started making changes to the product but not by making it more flashy to marketers.
We are making the product more useful to our top ICP users, which in turn will make it more appealing to the right marketers.
While we are seeing growth along all categories, marketing seems to be our ‘bread and butter’.
Marketers care.
Looking at the data it is somewhat obvious, isn’t it?
We should approach Marketing ICP with conviction.
VI. On the brink of $1M ARR
We know who cares about us.
We know why they care about us.
But the name Team-GPT got us from 0 to 1… is now holding us back.
People pull out of deals because we are the ‘fake ChatGPT’. With this name we will always be in the shadow of OpenAI.
Worst of all… we don’t have our own identity.
Team-GPT’s time is up.
Long-live XXXX.
New name redacted, will be revealed in November 25.
⚠️ Name is a secret, please don’t spread yet ⚠️
We are rebranding in November and will launch XXXX in San Francisco.
We will build our new identity, which will be fully focused on marketing as ICP.
Now that we know who our ICP is, we will listen much more carefully to what they have to say.
VII. What brought you from 0 to 1, won’t bring you from 1 to 10
We’ve achieved what 96% of startups never will.
Yet I’m far from satisfied, because we’ve got much more to show.
Hard work is just beginning.
This time on a different scale.
I’m thrilled to embark on the journey to $10M ARR!
Time to unlock our true potential ⛽️
Best,
Iliya