The five questions that surface 80% of opportunities.
The discovery framework we run on every engagement. Deceptively simple, which is the point.
The discovery framework we run on every engagement. Deceptively simple, which is the point.
We've run discovery for enough engagements now that we have a working framework. Five questions that, asked carefully, surface most of what's actually worth automating in a given operation. They look simple. They are simple. The work is in how you listen to the answers.
Here they are, with notes on what each one is really doing.
The point of this question is not to get a list of activities. It's to watch how the person describes their work. People who are deeply in the weeds will describe their week in terms of specific tasks and the rhythm those tasks fall into. People who are managing from above will describe outcomes and goals. Both are useful, but they tell you different things.
Listen for the tasks they describe with the most fluency. Those are the things they do most often. Listen for the tasks they describe with frustration in their voice, those are the candidates for automation. Listen for the tasks they describe with pride. Those are usually the things that should not be automated, because they represent the judgment, craft, or care that the team brings.
This is the most reliable single question for surfacing automation opportunities, because manual data movement is almost always a sign of misaligned tools, and the people doing it usually know exactly where it happens.
Watch for answers like "we copy from the CRM into a spreadsheet, then upload that spreadsheet into the reporting tool every Monday." Each transition is friction. Each transition is also typically automatable in hours, not weeks. These are the highest ROI builds you'll find.
"Manual data movement is almost always a sign of misaligned tools, and the people doing it usually know exactly where it happens."
Recurring reports are a goldmine, because they have all the qualities that make automation easy: defined inputs, defined outputs, predictable cadence. The interesting part is asking how long they take. People will often say "a few hours", and then, when you press, you'll find that "a few hours" actually means "a full day, but distributed across two people."
Total real time spent on recurring reports is almost always larger than the team's casual estimate. The first build for any team that produces recurring reports should usually be one of those reports.
This question surfaces a different category of opportunity, not time savings, but decision quality. There's almost always information that exists somewhere in the team's systems that would be useful for a recurring decision, but the cost of pulling it together makes the decision get made on intuition instead.
An internal knowledge agent or a custom dashboard often pays for itself just on the second category. Better decisions made faster, with the underlying data actually consulted.
This is the only question that asks for an opinion. It comes last on purpose. By the time you ask it, you've already heard the actual texture of the team's work. The answer they give you is interesting both for what they pick and for what they don't pick.
Often, the task they name is not the one with the highest ROI. That's useful information. It tells you something about morale, about what's most psychologically taxing versus what's actually most expensive in time. Both matter. The team's stated wishlist is data, not specification.
The order matters. Asking the opinion question first will anchor the rest of the conversation around what they think they want, rather than what their work actually shows. We always end with it, never start.
If you've asked these well, you'll have more candidates than you can build. That's the point. Discovery is supposed to produce abundance, so that prioritisation can happen against real options. Discovery that produces a single recommendation is usually discovery done backwards. The recommendation was already in the consultant's head, and the questions were ways of getting the team to confirm it.
From the abundance, you rank. By time savings, by error reduction, by decision impact, by build effort. Some opportunities will be obviously worth doing. Some will be obviously not. The interesting ones are in the middle, and that's where the conversation about priorities really happens, with both sides looking at the same map.
That's the framework. Five questions, used carefully, and the willingness to listen to what they actually surface rather than what you expected to hear.
The first call is free. Send a few sentences about what's slowing your team down, and we'll reply with honest thoughts on whether this is the kind of work we can help with.
Start a conversation