Lens: reading the buyer before the buyer reads you.

For many years I have walked into industrial sales meetings underprepared. Not because I did not care and not because I was lazy. Preparing properly simply takes time, and most operators quietly do not have enough of it. Reading a company well is a four hour minimum exercise if done seriously. You need to understand what they actually sell, how they describe themselves, what they want to become famous for, which regions carry budget, how their teams are structured, what industries they serve, and what changed in the last ninety days.

Most people never get that far. They skim the About page in the taxi. Open LinkedIn before the call. Search the annual report for revenue. Then walk into the room pretending they understand the company. Sometimes that works. Sometimes you realize fifteen minutes into the meeting that the company in front of you is completely different from the company you thought you were speaking to.

That problem kept bothering me, because I was sitting across increasingly technical businesses across Europe and realizing that most communication failure in B2B happens before the first sentence is even spoken. The seller arrives speaking their own language instead of the buyer's.

So over weekends, mostly late at night, I started building something for myself. Not a CRM. Not another sales intelligence platform pretending to automate relationships. More like a translator. A system that runs on the buyer before the buyer runs on me. I call it Lens.

Today Lens holds 208 industrial accounts across Europe. Germany dominates heavily, followed by Switzerland, Poland, Austria, Denmark and Sweden, with the rest spread across the European industrial belt. Only twenty five dossiers remain active at any given time and refresh nightly. The other 183 stay dormant until something interesting moves.

That dormant layer turned out to be more valuable than the active one. Because the real advantage is not the dossier you already have. It is the dossier you can produce on demand.

The pipeline itself is not especially magical once you open the hood. One agent reads news. Another crawls websites. Another summarizes. One maps corporate voice patterns. One analyzes regional team structures. Another tries to infer priorities and internal OKRs by persona. Everything writes to disk. Everything remains inspectable. Nothing disappears into black box AI theatre.

On top sits a very simple viewer. One HTML file. No framework. No startup performance art. Flags by country. Status badges. Active dossiers. Dormant stubs waiting to be activated. Paste in a company URL and go make coffee. By the time you return, the first draft of a usable company dossier already exists.

Three things that surprised me once it started working

Ⓐ The most valuable output is not the summary. It is the voice DNA. Most companies expose their internal language system publicly without realizing it. Websites, hiring pages, product descriptions, certifications, press releases. Once you can see how a company naturally describes itself, outreach stops sounding like outreach, because you stop writing in your own register and start writing in theirs.

Once you understand that voice properly, the economics of communication start behaving differently. The message lands inside their existing vocabulary instead of fighting it.

Ⓑ Team structure predicts deal structure more than most people realize. A company with fourteen marketing people in Germany and two in Poland behaves very differently from a company with two in Germany and fourteen in Poland. The presentation rarely changes. The opening line does.

Lens catches patterns I would have missed, because humans are surprisingly bad at interpreting organizational asymmetry at scale. You begin noticing where influence sits, where execution sits and where communication friction likely exists before the meeting even starts.

Ⓒ The dormant catalogue changes the economics of preparedness. I no longer prepare companies one by one. I maintain latent understanding across an entire industrial landscape, and the moment something moves I can activate a company in seconds and walk into the room with context that previously took half a day to build.

That changes the quality of questions you ask in the first five minutes. More importantly, it changes how quickly the other side realizes you actually understand their world.

Then it stopped behaving like account intelligence

Once enough companies exist inside the catalogue, the dossiers stop functioning as individual records and start behaving like a perception layer. I can upload a LinkedIn post, a draft email, a render, a slide deck or a positioning statement and ask questions that would normally require expensive research, agencies, or weeks of feedback loops.

How would German automation firms react to this message? Would Swiss laboratory companies trust this tone? Does this sound premium to Scandinavian manufacturers or simply expensive? Would a Polish engineering buyer find this language credible, or overly American?

The system does not replace real human judgement. But it creates something surprisingly close to organizational intuition at scale. A synthetic audience built from the public language systems companies expose every day.

Most messaging today is still created in isolation. A founder, a marketing team or an agency room produces communication based mostly on internal assumptions and then pushes it into the market hoping resonance appears afterward. Lens allows me to pressure test communication against hundreds of organizational identities before the message goes live. Not just for clarity. For resonance. For trust. For cultural alignment. For whether the message actually sounds like it belongs inside the world it is trying to enter.

At some point I realized this was quietly becoming less of a sales tool and more of a communication system. Probably even a PR system.

Where it goes next

Reading companies was always step one. The obvious next step is rendering them. This is where Lens begins converging with Cloudly. Once a dossier exists, the natural next artifact is no longer another PDF. It becomes a first visualization of the buyer's world. Their product, their materials, their factory environment, their visual language. All generated from the same intelligence layer that built the dossier itself.

The render survives the meeting. The render becomes the memory anchor. Lens becomes the brief. Cloudly becomes the visual translation layer. I think they were always heading toward the same place.

Eventually the system becomes orchestration. A Linux machine running quietly twenty four hours a day. A queue. A button inside the viewer that says "scrape this account." Twenty minutes later the dossier exists, the render queue starts, outreach drafts appear, and the print package is ready. One operator deep. No SDR layer. No sales team. No meeting before the meeting. Just compressed situational awareness.

A few disclaimers, because operators can smell omission instantly

Lens is not for sale. Cloudly builds real commercial systems with real engineering requirements. Lens is a private tool, built mostly on weekends because I got tired of walking into rooms half blind. And despite all the AI around it, I still rewrite most of the output myself.

Lens does not replace judgement. It replaces unprepared judgement.

That distinction matters. The machine saves the four hours I was never realistically going to spend anyway. But the more interesting shift is philosophical. Every operator selling something complicated eventually becomes a translator. Most of us spend our careers translating our products into the buyer's language. Lens began as an attempt to reverse the direction of that translation. To understand how a company sees itself before deciding how to speak to it.

The four hours it saves are useful. But that was never really the point. The point is arriving in the room already inside the other side's reality.