Why We Changed How We Describe Trinetro Labs, and What Stayed the Same
For a while we described Trinetro Labs with the same word as everyone else in the market: AI. It was the wrong word, and it put us next to products that work in a way ours does not. This is a short account of what we changed on the site, why, and what you can rely on staying exactly as it was.
Trinetro Labs · Published 12 October 2026 · 3 min read
The problem with the word
In data tools, AI has come to mean a language model that reads your question and predicts a response. That is a useful technique and it is not what Trinetro Labs does. Our answers are worked out from the structure of your data: the grain of each table, how tables join, what each metric means. The same question on the same data gives the same answer.
Using the same label as the tools that guess meant buyers, reasonably, expected a guess. We spent a surprising amount of every first conversation explaining that the number would not change from one run to the next.
What we say now
We call it Automated Intelligence. The work of reasoning through your data (profiling the tables, finding how they connect, choosing the right level at which to total a number) is automated, and it is shown to you in a reasoning trace. We describe Trinetro Labs as the world's first Automated Intelligence platform for data analysis, meaning we defined this category and shipped it: deterministic reasoning plus a governed team-memory layer.
The longer version is in the Automated Intelligence manifesto, and the terms are defined in the glossary.
What stayed exactly the same
- What the product does. Connect MySQL, PostgreSQL, MongoDB, Excel, Google Sheets, Tally and more, and ask questions in plain English.
- How answers are produced. Deterministic, with a reasoning trace. Nothing about the engine changed because of a rewrite of the words.
- The language model. On paid plans, one is used to put an already-computed insight into plain English. It never computes an answer, and that was true before this change.
- Access and security. Default deny, granted per member, down to individual tables.
- Pricing. A free 30-day trial with no credit card, and paid plans in rupees.
What you may notice
- Plainer claims. We say what the product does, and we say what it will not do, such as returning an all-time figure under a "last quarter" heading.
- Less comparison with chat assistants. They are good at what they do. We would rather describe our own product than argue with theirs.
- More about memory. How your team defines its terms is one of the most useful parts of the product and one of the least visible. You will see it described more clearly.
If you are weighing us against other approaches, text-to-SQL tools vs Trinetro Labs sets out the differences fairly.
Frequently asked questions
Did the product change when the wording changed?
No. The engine, the connectors, the security model and the pricing are the same. Only the way we describe them changed.
Why call it Automated Intelligence?
Because the intelligence is the automation of careful reasoning about your data, shown step by step, rather than a prediction of a likely answer.
Does Trinetro Labs use a language model?
Not in the answer path. On paid plans one puts an already-computed insight into plain English.
What does "world's first" refer to?
To defining and shipping the Automated Intelligence category for data analysis: deterministic reasoning plus a governed team-memory layer.