Cero
Cultural intelligence for one strategist. Built because the tools that do this are priced for enterprises and answer the wrong question.
The brief
Nobody set this one. I kept losing the same week to the same loop: find the relevant data, work out the "so what", make it presentable enough to show a client. The tools built for that job, Brandwatch and its peers, cost more than a junior strategist and are driven by Boolean queries that are wrong more often than they're right.
The signal
The bottleneck was never access. Every platform will happily give you ten thousand mentions. The bottleneck is relevance, and a firehose of mentions isn't research, it's noise with a subscription attached.
Watching where the hours actually went, almost none were spent thinking. They went on building queries, discarding irrelevant results, and rebuilding the query because the last one caught the wrong sense of the word.
I was never short of data. I was short of the few posts that mattered, and the hours I kept spending to find them.
What I made
Cero is a set of modules over one shared signal store, benchmarked openly against Brandwatch and built to beat it where it costs a strategist the most time.
- No Boolean. A client is set up from a plain-language brief and a few example accounts. A semantic classifier handles relevance from there.
- It corrects itself. One thumbs-down stops equivalent noise for that client permanently.
- Everything traces. Every number links back to the posts that produced it; every insight cites the signals it came from.
- Many clients at once. Live accounts and pitch prospects share one workspace, because that's the real shape of a strategist's week.
Where it stands
Running live across five modules: pattern recognition, listening, market analysis, research and reports, all over one shared signal store.
What it saves. The loop it replaces, scraping a category by hand and filtering it down to what matters, took a day or two. It now runs in under thirty minutes. That gap is the whole reason the tool exists, and it is why I could run a listening pass on Montblanc for a speculative pitch nobody was paying me for.
Three client workspaces are open in it: two active accounts and one pitch prospect, which is the real shape of a strategist's week. The skincare workspace is holding 12.4k documents from six live sources, and surfaces its own source health, so a scraper that has gone stale is visible rather than silently wrong.
-
The front door. A quiet room for the creative strategist, and the closest thing the tool has to a thesis. Elegance was written into the spec as an acceptance criterion rather than left as polish to be added later.
-
Where to play. Six fillable gaps drawn from 32 signals across 3,103 documents, each opening the sources behind it.
-
Setting up a client is a plain-language brief. No Boolean queries, ever. The filter learns the rest.
-
The relevance queue, holding 18 borderline documents awaiting a call. Each call tunes that client's filter.







