Your team already knows. Nobody asked them.
A StackUp Discovery Agent calls or messages your people on a schedule you set and asks the questions you never have time to ask. You get a summary, the three highest-value actions you could take, and every transcript and recording underneath.
What your team tells an agent is not what they put in a survey. Free to run your first agent.
Real screen · sample round
The problem
The people who know are not the people in the meeting.
The person using the unapproved tool, the one who knows the integration is held together by one colleague, the one with the idea — they are two or three layers down, and the mechanisms for hearing from them are a survey nobody fills in and a skip-level you keep postponing.
Surveys get the polite answer
A form with your name on it, going to your manager, is not where anyone admits they have been pasting customer data into something they found.
Interviews do not scale
Half-hour conversations with a whole department is a working week you do not have, and by the time you finish the picture has already moved.
It gets diluted on the way up
What starts as “we cannot do our job on this laptop” arrives three layers later as “some hardware refresh requests”. Every retelling rounds the edges off, and the edges were the point.
So you find out late
Usually when something breaks, someone resigns, or a tool nobody approved turns up in a security review with two years of history behind it.
How it runs
You set it up once. It does the asking.
Point an agent at the people you want to hear from
Name, role and contact details, or a CSV. Pick what you want asked — a tech health check, a specific project, shadow IT, ideas, or a set of questions you write yourself — and how often it should run.
Each person takes it their own way
The agent reaches out, and the recipient chooses a voice call or a chat. It is a conversation, not a form: it follows what they say and asks the obvious next question.
You read the findings, not the transcripts
A summary across everyone who took part, the three highest-value actions you could take next, and every session recorded underneath so you can read or replay the conversation behind any of them.
What to ask
Point it at what you actually need to know.
Tech health check-in
What is actually slowing people down, what keeps breaking, and which workaround has quietly become the process.
Project check-in
How a specific delivery is really going, asked of the people doing it rather than the person who writes the status report.
Idea generation
The improvement everyone mentions in the corridor and nobody ever writes down, collected from the whole team at once.
Shadow IT and AI discovery
Which tools your team has adopted without telling anyone, and what they are putting into them. This is the one that produces the surprises.
Or write your own
Build a custom agent with your own questions and save it as a template. Anything you would ask if you had the time to ring fifty people yourself.
How teams use it
You do not have to ask everyone at once.
The organisations getting the most out of it run small, regular rounds rather than one big survey nobody wants to repeat.
A slice of the organisation each month
Ten per cent of the team, rotating. Nobody is interrupted often, and you get a continuous read rather than an annual snapshot that is stale by the time it lands.
Before a quarterly planning session
Run it the fortnight before. You walk into planning already knowing what the teams think the problems are, instead of spending the first two hours finding out.
As standing quality control for IT
A monthly rhythm on a rotating sample tells you whether service is actually improving, from the people receiving it rather than from the ticket queue.
To get every brain in the game
Ideas rarely come from the people in the planning meeting. Asking everyone, properly and on the record, surfaces the ones that never make it up the chain.
What you get
The three highest-value actions, with the calls behind them.
These are the actual screens, filled with a sample round rather than a customer's.
Findings, not a pile of transcripts
A round of conversations is a lot of listening. The results page does the reading: a summary of what came up across everybody, then the three things most worth doing something about, ranked by urgency rather than by who said them loudest.
- A summary across everyone who took part
- The three highest-value actions, ranked by urgency
- Every session kept, with its transcript and recording
- Re-run it later and watch what changed
And the conversation it came from
Open any session and you get the whole thing: what the agent took from it, and the full transcript underneath. Voice calls can be played back, so you can hear how something was said rather than only how it was summarised.
That matters more than it sounds. A finding that is going to start an awkward conversation should be checkable by the person having it, and “here is the recording” ends an argument that “the system says so” would not.
Why it works
People tell it things they would not put in writing.
Not because it is clever. Because of who is not on the other end.
No colleague forming a view
Nobody is deciding what your answer says about you. That removes most of the reason people soften what they tell their own organisation.
It asks the next question
A form takes what you type and stops. A conversation notices that you said 'we worked around it' and asks how, which is where the useful part usually is.
Everyone gets asked the same thing
One question set, one round, everybody asked the same way. When four people independently name the same colleague or the same tool, that is a finding rather than an anecdote.
Who it is for
Anyone who needs to hear past the first layer.
CTOs and IT leaders
The shadow-IT picture you cannot get from a survey, and a standing check-in that does not consume your calendar or theirs.
Executives and CEOs
An unfiltered read from further down the organisation than you normally hear, gathered without anybody having to volunteer to be the bearer of bad news.
Fractional CTOs and advisers
Discovery across a client's team in days rather than a fortnight of interviews, with the transcripts to back up what you tell them.
What it costs
Run your first agent for nothing
Create an account, add your people, choose what you want asked, and send it. You get the summary, the ranked actions and every session record.
- No card, no sales call, no trial clock
- Voice or chat, whichever each person prefers
- Schedule it, or run it once and see
- Paid plans add more agents and the rest of the platform
Running discovery for a client? Advisers with portfolio access can point an agent at a client's team and share the findings with them.
Questions
The things people ask first
The rest of the platform
Five other things you can run
Each answers a different question. They all land in the same place.
Technology Assessment
The whole technology function, scored out of 100.
How it worksAI Maturity Assessment
Where you stand on the four questions boards ask about AI.
How it worksAI Leadership Skills
Whether your leaders could defend the AI strategy they set.
How it worksProject Assessment
A health score for one project, calculated not reported.
How it worksCrisis Tabletop Simulation
An hour deciding your way through a year of incidents.
How it worksAsk your team what you are not being told.
Point one agent at one team this week. The shadow-IT answer usually arrives in the first two minutes of the first call.