Intelligence answers the question you have today
Dashboards answer the questions someone thought to build. A look at what plain-language questions change, and what makes the answers worth trusting.

A dashboard is a set of answers to questions somebody asked last year. It is useful as long as this year’s questions are the same ones.
They usually aren’t. An owner wants to know why one job’s margin slipped. A director wants to know which programs cost more per participant than they did in the spring. Neither is on the dashboard, so someone is asked to pull a report, and the answer arrives after the meeting it was for.
Why this became possible
Asking a database a question in plain English has been promised for decades. It works now for a specific reason. Language models are very good at translating a question into a precise query and explaining what comes back. They are much less reliable when asked to recall a number themselves, because a model produces the most plausible next words, and a plausible figure is easy to produce.
So the design that works keeps the model away from the arithmetic. The model writes the query. The figures are calculated from your data by ordinary software, the kind that has added columns correctly for fifty years.
Intelligence is built this way. It draws on what the other Extensar products have captured as the work happened, alongside the systems you already run, and each answer comes with its sources.
What people ask
The first questions are the old report requests. Within a few weeks they get more specific. What did we promise this client, and have we done it? Which grantees are waiting on us?
Cheap questions change behavior. People check an assumption before a meeting when checking takes a minute. A project manager who would never have asked the office to pull three months of labor on one job will ask Intelligence on the way to the site. It also changes who gets to ask. In most companies a handful of people know how to pull a report, and everyone else waits on them.
Standing numbers still matter. Intelligence shows performance against the goals you set and budget to actuals for every project or program, without anyone maintaining a report. It also flags the ones that need a decision from your team, with what you need to know to make it. On a Monday morning that means seeing which jobs slipped last week and what each one needs from you.
The limit worth knowing
An answer can only draw on what was captured. If change orders live in email and never reach a system, no question will find them. This is the practical reason Intelligence sits on top of products that capture the work as it happens. The field text and the call with a funder are already captured, so the question has something to find.
Where the data is thin, Intelligence says so plainly. That is useful to know in its own right, because it shows which part of the business still runs on memory.
The decision stays with the person who asked. Intelligence gets them to it with this week’s information and the sources in view.


