We Invented an Oil Company (On Purpose)

We invented an oil company.

It doesn't exist. No rigs. No wells. No employees. Just five fields, forty wells, five hundred people, ten rigs, a hundred vendors, and two years of drilling history. All of it made up, on purpose.

Here's why. In oil and gas, a wrong number isn't a typo. It's a budget built on bad math. A safety barrier nobody checked. A risk register that's folklore. And drilling data sits behind locked doors: confidential, wrapped in NDAs, shared only with partners an operator trusts. We weren't going to wait for a key. So we built a synthetic world, detailed enough to bleed. And then a real one walked in. Everything you're about to read, the invented company, the agent fleet, the patents, the real client, happened inside eight months.

· · ·

We called our invented company DSEC. Deepwater fields with names like Orion and Vega. Around 2,650 data files: daily drilling reports, cost estimates, vendor contracts, near-miss logs, even employees with expired training certificates. And all of it messy on purpose: realistic delays, optimistic estimates, documents that quietly argue with each other. Why messy? Because clean data teaches you nothing.

There was a second reason, and it was hungrier. Our patent-pending ideas, UXLens and VideoLens, eat data for breakfast. How do you prototype something that runs on data, with no data to run on? You don't. You manufacture the data, in three moves. The best deep-research agents on the market mapped the anatomy of every document. Then our own AI agents wrote the reports, drawing on ChatGPT, Perplexity, Gemini, and Claude, all of them. And old-school statistics kept it honest, tuning the trends and patterns until the made-up company behaved like a lived-in one. A human in the loop on every file, of course.

Twenty-plus years around this industry, and I never once saw drilling data out in the open. I used to check Kaggle like a lottery ticket. Now I can ask for a daily drilling report where the depth climbs, sensibly, for 84 straight days, and get it. Ten wells blended into one best-of composite well? Possible. What a world we live in.

And "human in the loop" isn't a slogan. This is a collaboration: SwarmLens brought the AI (https://swarmlens.com/); rp² (https://rp-squared.com/) , pioneers in oil-and-gas consulting, brought a deep bench of world-class drilling SMEs; together we built AssureLens (https://assurelens.tech/). Models that read everything, working beside engineers who've seen everything. Call it collective intelligence: the only kind a start up our age can afford.

The synthetic data flipped the room, too. Show a drilling expert a slide deck and you get politeness. Show them a working application and you get war stories. "Where's the data?" used to be the dead end. Now it's the second meeting.

· · ·

And the part I'm most excited about? I can't show you yet.

On this campaign we built something we believe is genuinely new: a different way to experience an archive altogether. Not a dashboard. Not a report. Not a chatbot. The patent applications are in flight, so the details stay with the lawyers for now. I'll say only this: once you've seen it, a dashboard feels like a filing cabinet. That story gets its own post, once the paperwork clears.

· · ·

So that's the arc. Invent a company. Build its departments out of agents. Make every number confess its source. Then point the whole thing at a real drilling campaign, and watch the archive wake up and start predicting the next well.

Remember the locked door? It opens the other way now. AssureLens is ready, and the trade is simple: you bring the data, we do the rest.

We're looking for a few design partners: operators with years of daily reports, end-of-well reports, and risk registers gathering dust, written by people who solved hard problems and moved on. Oil and gas first, but the machinery doesn't care what industry your documents come from, only that the stakes are high. If you've ever wondered what's buried in your archive: [e-mail]

(And if you've ever tried re-reading years of daily reports by hand, I'd genuinely love to hear how far you got.)

And one last thing. We invented an oil company because that's where our experts live. But this is a recipe, not a one-off. A bank. An airline. A pharma company. Same synthetic world, same agent departments, same receipts on every number.

So tell me: which company should we invent next?

https://www.linkedin.com/pulse/we-invented-oil-company-purpose-hari-menath-4tasc/