Fraktal

Build your own GTM tools that always stay up to date.

GTM tools today have two big flaws

1

They are built for everyone

Different companies need different things from the same channel. The SEO tool a SaaS company needs is not the one a VC fund needs.

SaaS startupVC fundClinic
→
One general tool
→
Same outputSame outputSame output
2

They run on snapshots

A playbook from last quarter, an SEO tool's best practices, a LinkedIn template that worked in spring. They go stale while the channels change every week.

25%16 to 17 Oct
2.5%27 Oct

Eg. We asked AI engines the same 484 questions ten days apart. ChatGPT answers citing big news sites (WSJ, FT, Reuters and others) fell from 25% to 2.5%.

What would the ideal GTM tool do?

Imagine if your GTM tool…

…was built for your company, not for everyone

…ran on your own data, so it sounds like you

…read what works on every channel, every day, and remembered why

…knew the same day when a strategy stopped working

…showed what to do next, and did it for you

Your GTM toolUpdated today
✓Built for your company
✓Runs on your data
✓Knows what works now
✓Knows when a strategy stops
✓Does the work for you

Could founders stop switching tools?

Enter Fraktal

The brain: a labeled GTM dataset that watches 1,000+ live sources across every channel and keeps a running picture of what works: what changed, why it matters, and what to do next.

The toolkit: plugs into Claude Code, Codex or Cursor and interviews you about the tool you need. Your agent builds it on your own data, connected to the brain through our API and MCP.

Each tool is a full toolbox. Example: a LinkedIn content manager that

  • finds post ideas and researches them
  • writes hooks, bodies and CTAs in several variants
  • designs the visuals
  • recommends what to post next

When a post type stops working on LinkedIn, the tool knows the same day.

Fraktal CLI: npx @usefraktal/cli init, choosing LinkedIn as the channel
Fraktal CLI · preview

Why now? Building beats buying

What your own GTM tool gives you

  • It fits how your company sells, not an average of every company
  • It writes from your calls, your CRM and your customers' words, so it sounds like you
  • It does the work, not just shows a dashboard
  • You own it: change it any day, with no vendor roadmap to wait for
  • It stays connected to the brain, so it never goes stale
In 2026, an AI agent can build a GTM tool made for your company faster than you can pick one off the shelf.

Luka built custom GTM tools for a VC-backed fintech startup with AI agents. Each one fit the company better than the tools on the market.

What we heard so far

Marketing team

"I won't click around in any tool if I can hook into yours and my content just gets better."

Growth engineer

"For me one really important thing is that the tool has both an API and an MCP."

Growth lead

"I tried a couple of tools, I didn't like them, they don't give good overviews."

Content lead, on writing from their own data

"I think it's mandatory… Anything else is, like, literally a disaster."

Investor and advisor

"I like the methodology. I really do."

Sales lead

"I think it's good that you're going in that infrastructural direction."

GTM and sales

"It has to be scalable, affordable, and yet easy to integrate."

Content and growth lead

They asked for a demo, and for an AI-writing check.

08

Traction so far (in the 1st week!)

17
pilot companies
12

from the portfolio of Suhit Gupta (Sparked Ventures), one of our advisors

3

from our own network

2

from the portfolio of Aron Starosta (UE Ventures), one of our advisors

Fraktal's objective function: maximize freshness

Fraktal shows which strategies still work today, and the results that prove it
Every strategy is a bet:
- if it still works, it stays
- if it stops working, every tool on Fraktal knows at the next daily check
Fraktal's definition
fresh·ness
/ˈfreʃ.nəs/
noun
  1. The quality of being new.
  2. Advice that still works today, checked against real results.

AI finds patterns in new results every day. GTM engineers and consultants who do this work daily vet each result: they tell noise from what really works, and decide what the brain learns

10

What we're building: GTM tools founders own kept up to date by one brain

AI agents can build almost any software

But they do not know what works in GTM. GTM experts do.

Fraktal gives every founder's agent what works now, on their own data, so founders own GTM tools that fit them and stay up to date
11

Who is building Fraktal?

Luka Čereković
Luka Čereković
CEO

8 years running GTM and growth for startups in several countries; raised millions in VC and added millions in ARR for them. At a VC-backed fintech in New York he built the whole GTM system with AI agents: $700K ARR in qualified pipeline in 4 months, $200K ARR closed from SEO and AEO.

Bojan Stefanović
Bojan Stefanović
CTO

Former Google and Reddit engineer, 10+ years building software. Built Google's infrastructure that audits and deploys all its iOS apps. Cut Reddit's post page load time by over 90%. Built evaluation systems for Anthropic and Mercor.

Jovan Jovanović
Jovan Jovanović
CCO

13 years researching how people communicate. Built 22 engineering and 7 marketing teams from zero; co-founded Friendly HR (100+ clients) and Match IT (Genesis Startup Academy, run by Meta and AWS).

All 3 full time. Incorporated 31 October 2026.

Advisors: Suhit Gupta (Sparked Ventures) and Aron Starosta (UE Ventures). Design partner: Bo Motlagh (UE Labs).

Why you shouldn't let us in YC?

  1. Unproven at scale: we do not know yet how many founders will build their own GTM tools instead of buying them.
  2. Very early: Fraktal works with 17 pilot companies, and we build the first tools by hand with them. We think the network effect starts at about 1,000 tools built on Fraktal.
  3. Pricing is new: $3K to $5K a month when we build and run the tool with the client, $799 a month self-serve. Not tested at scale yet.
12

Why you will want to let us in YC?

Over the next 5 to 10 years…

  1. Most companies will build their own GTM tools instead of renting them
  2. AI agents will run most GTM work (one startup we spoke with already runs 30 to 50 agents)
  3. What works on a channel will change in days, not years
  4. Text that sounds like AI will lose value. Tools that write from a company's own data will win
  5. Anyone can collect data. The moat is GTM experts who do the work every day, vetting the data and telling noise from what really works

The right GTM tool at the right time can make your company

13

Let's talk

luka@usefraktal.com

Luka Čereković, CEO · usefraktal.com