A dashboard answers the questions somebody anticipated six months ago. Most of the questions worth asking are the ones nobody anticipated, and they usually span more systems than any single report was built to reach.
Accounting, CRM, PSA, practice management, telematics, file storage, databases, your own internal apps. Read-only credentials, connected once. Nothing is migrated and nothing moves.
No query language, no report builder, no waiting for someone in finance to have a free afternoon. A question typed the way you would say it out loud.
Every lookup is named as it runs: which system, how many rows, how long it took. The working log is not decoration, it is how you know the answer came from your data rather than from a guess.
Figures, a table, a chart or a written recommendation, whichever the question actually deserves. Then the follow-ups it thinks you will ask next.
Every answer is built from the same four parts. The first two are what make the last two checkable.
Which customers are slowest to pay?
| Customer | Avg days | Outstanding |
|---|---|---|
| Bellamy Ltd | 71 | £84,200 |
| Harcourt & Co | 64 | £52,900 |
| Winslow Group | 58 | £41,600 |
| Petrie Civil | 47 | £28,400 |
| Oakmoor | 44 | £19,850 |
Five accounts carry £227,000 of the £310k outstanding. Bellamy averages 71 days against 30-day terms and has done for four quarters, which is a terms problem rather than a collections one.
No query language and no field picker. The phrasing you would use talking to a colleague is the phrasing that works.
Each lookup named as it runs, with the system, the number of records and the time it took. This is the part that makes the answer checkable rather than merely plausible, and it folds shut once the answer starts so it never gets in the way.
A table here, because the question asks for a ranking. A chart where the question is about movement over time. Two paragraphs and a recommendation where that is the honest answer. Not everything is a chart.
The last line is the part a person would have had to work out themselves: that a 71-day average sustained over four quarters is a terms problem, not a chasing problem.
Grouped by the question each one answers for the person evaluating it.
No migration, no data warehouse project.
The systems your company runs on stay exactly where they are. ASKIQ reads them in place, which is why connecting one takes an afternoon rather than a quarter.
Accounting, CRM, ERP, PSA, practice management, project and field tools. The systems that already hold the answer.
SharePoint, OneDrive, Google Drive, network shares. Contracts, drawings, handovers and reports, read as content rather than filenames.
SQL databases and in-house applications connected directly, including the bespoke system nobody else integrates with.
A connector asks for the least it needs. Write access is never required to answer a question, and is never granted unless you ask for it.
The question nobody can answer today.
Most questions worth asking span two or three systems, which is exactly why they go unanswered. One question here reaches all of them.
Join a figure in the ledger to a record in the CRM and a timesheet in the field app, in a single answer.
Speaks the business, not the database. No table names, no query language, no report builder to learn.
Ask a follow-up and it knows what "that division" and "the same period" mean, without restating the question.
Every answer offers the questions it expects next, so you can keep pulling the thread instead of starting over.
An answer you can defend in a meeting.
An assistant that cannot show where a number came from is a liability in any conversation that matters. Every answer here carries its provenance.
Each step names the system it read, how many rows it touched and how long it took, as it happens.
Where a structured query was run, it is shown, so a technical reader can check it.
Any number in an answer can be traced back to the system and the records it came from.
Not everything is a chart. Where the true answer is two paragraphs and a recommendation, that is what arrives.
Fast enough to ask the second question.
Speed is not a vanity metric here. A report that takes three days is a report you only request when you are already sure; an answer in seconds is one you can be curious with.
A question that used to mean a ticket, a spreadsheet and a wait now resolves while you watch.
A short answer and a long one take about the same time, which is what makes it feel fast rather than variable.
The person with the question asks it directly, instead of queueing behind whoever knows how to build the report.
Sources are read live rather than from last night’s export, so the answer reflects today.
Your data does not become somebody else’s training set.
An internal assistant only earns the "internal" part if the data it reads stays under your control. This is the part IT asks about first.
A person sees through ASKIQ only what they could already see in the underlying system. It does not widen access.
Your data is never used to train a model, yours or anyone else’s.
Hosted, in your own cloud tenancy, or fully self-hosted for organisations whose data cannot leave the building.
Who asked what, when, and which systems were read to answer it.
It should know what a "matter" is.
A generic assistant makes the reader translate. This one is taught your vocabulary, your entities and the shape of your own reporting.
Matters, jobs, drops, episodes, tenants, sites. It learns the words your business actually uses for its own things.
How your financial year runs, which entities consolidate, what counts as billable and what does not.
A starting set of questions built for your business, so the first day is useful rather than a blank prompt.
Name it, colour it and put your own mark on it. Internally it is your tool, not a vendor’s.
Six industries, fifty questions each, every answer with its working shown. No signup.