Sep 3, 2025 · 46m · mixergy
#2278 How to build a $15M/year AI company.
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Keboola co-founder Pavel Dolezal details how democratizing data access, pivoting to product-led growth, and integrating generative AI workflows enable bootstrapped startups to scale into multi-million-dollar enterprise software companies.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Andrew holds 18.1% of the talking time here. How this is scored →
Andrew as informed peer 4.7 Guest teaching 3.7 Guest disagreement 0.7 Andrew pushing back 1.5
speaking balance: gold is Andrew, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 34:49 Pavel corrects Andrew on company mission
Pavel politely rejects Andrew's summary that their goal is merely data accessibility, emphasizing that accessibility is just a prerequisite for full business process automation.
Hardest push from Andrew ▶ 8:34 Andrew questions persistence of data access problemsAndrew challenges Pavel's timeline, expressing surprise that obtaining business data was still severely difficult up through his NetMail days in the 2010s.
Biggest teaching moment ▶ 8:59 Pavel explains enterprise SaaS sprawl and Hadoop drawbacksPavel educates Andrew on the technical shift from Hadoop to multi-backend SQL and cites Gartner statistics showing up to 300 disconnected SaaS tools inside large enterprises.
Andrew holds their own ▶ 12:27 Andrew cites cross-interview insights on data accessibility vs agentsAndrew demonstrates deep industry command by referencing his recent interviews with other AI companies to validate the industry-wide bottleneck of data access over flashy autonomous agents.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Andrew as informed peer | Guest teaching | Guest disagreement | Andrew pushing back | Why |
|---|---|---|---|---|---|---|
| Democratizing Data: The MacPen Case Study | 4 | 3 | 0 | 0 | The conversation begins cordially as Andrew introduces Pavel and prompts him to share the MacPen case study. Andrew synthesizes the story well by explaining the unit-economics impact of daily up-to-the-minute data. | |
| Origins of Keboola and Enterprise Data Challenges | 5 | 5 | 1 | 3 | Andrew challenges Pavel on why data extraction was still problematic well into the 2010s after Atlas. Pavel schools Andrew on the shift from Hadoop complexity to fragmented cloud data warehouses and the explosion of hundreds of enterprise SaaS apps. | |
| Building Software Through Hands-On Consulting | 6 | 4 | 0 | 1 | Andrew demonstrates his active industry knowledge by comparing Pavel's approach to other AI founders he interviewed who prioritize data access over agents. Pavel elaborates with anecdotes of early hands-on consulting and co-designing with founders like Tomas Cupr. | |
| Navigating COVID-19 and the Shift to Product-Led Growth | 4 | 3 | 1 | 2 | Pavel describes helping the Czech government during COVID-19 and pivoting the company to a product-led growth model. Andrew presses gently on the lack of government gratitude and asks how PLG fit into an enterprise sales motion. | |
| Grassroots Growth, Hackathons, and Enterprise Adoption | 4 | 3 | 0 | 1 | Pavel shares their grassroots strategy of running community hackathons to land large regulated enterprise clients like Erste Bank. Andrew guides the chronology smoothly and extracts key revenue metrics ($15M ARR). | |
| Scaling with AI and End-to-End Workflow Automation | 5 | 4 | 2 | 2 | Pavel reframes Andrew's view by clarifying that data accessibility is merely a prerequisite for end-to-end business process automation with LLMs. He concludes with practical examples like GymBeam and the explosive rise of vibe coding. |
Statements from this episode (14)
Dolezal: Data-informed frontline upselling can double a retail business's profit margin
“And this is so, so, so huge impact, you know, because like you can literally double the profit margin just by these simple things.”
Dolezal: Frontline staff finding daily micro-insights creates more value than rare breakthroughs
“Either you go and you find a big, big nuggets. There's just a couple of them, or you have a river and everybody sits alongside the river and they just, you know, like do this and everybody finds one small gold flake and a day. But if you have a thousand people…”
DXC runs global sales and marketing data pipelines on Keboola
“DXC, right? Like, Sixteen billion dollars a year company, you know, the original IT company, and they use us, you know, to run all their sales and marketing data pipelines around the globe for last eight years.”
Czech PE group bought Groupon and redesigned its data using Keboola
“Groupon as an original Groupon is now run by Czechs as well. You know, some of our friends and client, and it's our client actually have bought in as a private equity. Yeah. Yeah. I have redesigned the whole coupon and data using Kabula.”
Dolezal: Keboola signed 21,000+ companies via PLG, growing 3x in 12 months
“Over 21,000 companies actually signed up for us. We grew it. We started to grow it, you know, and now like last 12 months, we grew it three times.”
Dolezal: Post-COVID software sales always require a technical user trial
“The world has changed, you know, after Corona, you know, I don't think that people are buying now just, you know, like by phone only or introductions or reference. Always, especially in our industry, there will be some, one technical who will want to try the s…”
Dolezal: Keboola's enterprise starting contracts range from $75K to $250K
“Our tickets are, you know, like mid, mid, mid tickets are 75, 250 K, you know, as a starting ticket we still have, you know, like hundreds and hundreds of companies that pay us by credit card, just like 20 dollars, 50 dollars, and they grow, right?”
Keboola data workshop has educated over 30,000 women
“We designed, you know, a specific workshop with them, and since then, you know, like, it's been taught to over 30,000 ladies, and it actually evolved. Into an Academy”
Hackathons should avoid cash prizes to deter mercenary hunters
“Our, also our guiding principle was we're not going to do prizes. Because what happens with hackathons when you do prizes, you have a professional hackathon hunters and they go and just swoop in, you know, they take the prize there and they go off.”
Keboola reached $4M in profitable revenue before raising its seed round
“So we raised a seed round in June, and by then we had a company, we had if I'm correct, something three, 3.54 million dollars you know, like until then we were running the company to be profitable because like, you know, like we were bootstrapping or cashflow …”
Dolezal: GymBeam cut support team from 50 to 3 using Keboola and LLMs
“What they did, they had a 50, 50, 50 people team. With the support user support, like content support, and their biggest problem, like people give them reviews in 16 languages everywhere, right? One of their biggest issues was that people would, you know, actu…”
Dolezal: Non-technical founders win on vertical knowledge, not coding MVPs
“And so, you know, like, I think the, if you don't know, if you don't have technical things, what is your actual secret is your vertical knowledge.
You know, and that's the thing, you know, coding at least to MVP, not production, but MVP is like now easy, and a…”
Dolezal: Solo founder built $10M ARR vibe-coding business in past year
“I have literally next to me, I, like, I have two friends who built businesses over last year.
One of them is now ten million ARR.
Ten million.
One person.
Vibe coding app, and the second one is, is, is doing three million, you know, and just like, time is now,…”