Three engagements, three very different businesses, names withheld because these are clients, not case studies. Each is shared with the client's knowledge. What an analysis found, and what got built - including where AI honestly was not the answer.
A residential builder, six years in, growing faster than its paperwork.
The owners knew their business was healthy. What they could not see was how much of it lived in one office manager's head: which sub gets paid when, which insurance certificate is about to lapse, what was promised to which client and where that promise was written down, if it was written down at all.
The analysis did not tell them anything about AI at first. It held up a mirror. In one working session, we mapped how the company actually runs, not how the org chart says it runs. The dependency showed up in black and white, and the room went quiet for a moment. Then one of the owners said we had nailed exactly where they are.
The honest part: AI will not fix an unwritten process. If the knowledge lives in one person's memory, the first job is capturing it on paper. Only then do the tools have something real to work with. That capture is where their engagement started, because that is where their risk was.
A media and marketing agency winning work on instinct and rebuilding the wheel each time.
The agency's work was good. That was never the question. The question the analysis surfaced was why every new project felt like the first one: proposals rebuilt from blank pages, processes re-invented per client, the owner's standards applied from memory because they were not written anywhere.
The mirror showed which parts of their work genuinely are custom, and which parts repeat every single time and could be carried forward instead of rebuilt. Seeing that split on paper changed how they thought about their own margins.
The honest part: AI was never asked to invent their standards - the standards worth having were already theirs. The build captured them in writing and applied them consistently, which is the part that had never happened before. For a creative business, that is the real unlock: not automating the creativity, but capturing the repeatable frame around it so the creative time goes where it earns.
A documentary director running two productions simultaneously, with the story living in notes, memory, and late-night worry.
A documentary is thousands of moving pieces: interviews, story beats, releases, footage, funding threads. Running one is hard. This director was running two at once, and the connective tissue lived in stacks of notes and a very good memory that was reaching its limit.
The analysis mapped what the work actually needs against what the tools out there promise, which are rarely the same thing. What emerged was not an app recommendation. It was a picture of where the productions genuinely lose threads, and a way to hold those threads in one place the director actually works from.
The honest part: nothing we built directs the film. The story sense, the taste, the call on what a film is actually about - that stayed the filmmaker's, and this build never tried to move it. What AI took over was everything around that call: holding the corkboard steady, keeping every thread visible, stopping the 2 a.m. fear that something important slipped. That is worth having - and the ceiling is not the tool. What AI does beyond this grows with the person operating it; every starting point is different, and this was one filmmaker's.
Details anonymized in every note, shared with each client's knowledge. That is what an analysis is for: not a software pitch, but a clear look at your own company - including the places where AI honestly is not the right next step yet.