How AI shifts process discovery ahead of the workshop – and turns two days of assessment into two days of decision-making
In many companies, process workshops still begin according to the same pattern. Ten or twelve subject-matter experts and managers come together for one or two days, with a whiteboard or a digital process map in front of them, and then the joint reconstruction of the current state begins. Who actually does what today? Which information is needed? Which systems are involved? Where are approvals given? At which points do queries, waiting times or media discontinuities arise? Step by step, the wall fills with boxes, arrows and Post-its, and at some point a shared picture emerges of how the process works today.
There is nothing wrong with that at first. A shared understanding of the current state is a prerequisite for being able to talk about improvement. Even so, a simple economic question arises: Do ten highly qualified people really need to sit in one room for two days to achieve that?
If ten people take part in a two-day workshop, we invest 20 person-days of shared attention. This time is scarce and expensive. And yet a significant part of it is often used to gather information that could also be collected in a structured way beforehand. While one person explains their part of the process, the others listen, add, correct or wait for the moment when their own interface becomes relevant. This does create an overall picture step by step, but a large share of the joint working time is used for information gathering.
This is exactly where we come in with AIS [ai]works Process Intelligence. Our position is not that workshops become unnecessary. On the contrary: good workshops become even more important. But they should be used for the tasks for which we actually need people together in one room. These are not primarily the collection of information or the drawing of an as-is process. They are resolving conflicting objectives, assessing different options, prioritising changes and making joint decisions.
That is why our guiding principle is: The best workshop does not begin with an empty whiteboard.
The actual process discovery can take place beforehand
The decisive difference lies in decoupling process discovery from the workshop. With AIS aiWorks Process Intelligence, the current state is examined in advance through structured role interviews. The point is not to ask the interviewee for an abstract process description. We reconstruct specific, actually completed cases and use them to examine how decisions, roles, data, systems and interfaces interact in reality.
If, for example, we analyse a procurement, budgeting or maintenance process, what interests us first is not how the process is supposed to work according to the manual. We want to understand how the last relevant case actually unfolded. This also makes visible those parts of the organisation that are often missing from official process representations. Information may not come from the designated system, but from an Excel file. An approval may initially have been obtained informally and only documented later. An interface may function reliably only because two people coordinate well personally. An officially fully digital process may in reality contain several manual checks.
Each individual interview already provides a structured piece of process reality. But the greater value arises only because these conversations are not then considered in isolation. Statements from different roles are compared with one another, contradictions are made visible and open points are specifically cross-checked in later conversations. At the same time, the Elephant Finder looks for those deviations between the official and the lived organisation that are easily overlooked in traditional process discovery.
In this way, several conversations do not simply result in a collection of interview notes, but step by step in a consolidated as-is model. Roles and responsibilities become clearer, actual process chains can be reconstructed, decision paths become visible and systems and data sources can be assigned to the individual process steps. At the same time, it becomes apparent at which interfaces queries, waiting times or manual handovers occur particularly frequently and where different roles perceive the same case differently. This very consolidation is already embedded methodologically in the AIS Process Analysis Framework: the cross-analysis produces a prepared process picture, pre-sorted contradictions, a deficit heat map and known Elephant Findings before the joint workshop begins.
The workshop therefore starts at a completely different point
Once this preparatory work has been completed, the function of the workshop changes fundamentally. The participants no longer come together to start from scratch. They are presented with a prepared picture of today’s organisation, based on real
cases and the perspectives of several roles. This picture is explicitly not yet a final judgement. It is a robust working hypothesis that is validated together and corrected where the analysis is not yet complete.
The decisive difference is that the group no longer needs several hours simply to build a shared understanding of the current state. It can move much more quickly to the points where decisions are actually required. If different functions have described the same interface inconsistently, that inconsistency is already on the table. If a particular process step regularly leads to long waiting times, this is known. If an informal workaround appears in several interviews, it can be discussed directly whether it is a sensible special route or an expression of a structural problem.
This also changes the quality of the discussion. In traditional workshops, conflicts often only arise in the room. One function describes a process, another contradicts it, a third introduces a special case, and suddenly the moderation has to understand at the same time what is happening factually, why perceptions differ and how the discussion can be brought back to a decision level.
If these differences are already known from the interviews, the moderation can prepare differently. Critical interfaces, differing perceptions and potentially conflict-prone questions are known before those involved come together. This makes it possible to deliberately integrate the relevant topics into the agenda, instead of hoping they will somehow come up by chance during the workshop.
This is more than more efficient preparation. It changes the role of the workshop from the ground up. The workshop shifts from being a place of assessment to being a place of decision.
This shift is one of the central ideas of our methodology. The process picture already exists and is validated instead of being created jointly from scratch. The relevant contradictions are known, the agenda can be developed from the deficits actually identified and quantified friction points help focus the discussion on the topics that truly create impact for the organisation.
The shared time is used for the difficult questions
This is precisely where we see the real economic lever. The valuable shared time of managers and specialists is not saved by replacing people with AI. It is used more effectively.
A system can structure interviews, compare statements with one another, mark contradictions and condense large volumes of qualitative information. It can detect that several roles describe the same case differently or that
a particular problem appears repeatedly. But the difficult question only begins after that: What does this insight mean for the future organisation?
Should a previously decentralised responsibility be centralised? Is a process variant professionally necessary or merely historically evolved? Can an approval step be removed or does it serve an important governance function? Does a system need to be replaced or is the real problem unclear responsibilities? Which change brings measurable benefit and which would merely make a process picture look better?
These are not questions that should be delegated to an interview agent. They are management and design questions. That is exactly what we need a workshop for.
The logic of AIS [ai]works Process Intelligence is therefore not “AI instead of workshop”, but AI before workshop. The machine takes over where structured assessment, comparison and consolidation scale. The joint space is reserved for the points at which different interests, experiences and responsibilities need to be brought together.
This distinction is important because the discussion about AI in consulting is often reduced too much to the question of which human work can be automated. For us, the more interesting question is which human work becomes more valuable through better preparation.
From pain points to priorities
Another advantage arises from the fact that the problems are not only collected during the workshop. In many process projects, at some point a wall fills up with pain points. Some of them are relevant, some merely annoying, and others occur only very rarely. Even so, at first they stand side by side on an equal footing.
The upstream analysis makes it possible, even before the workshop, to distinguish which deficits actually have a greater impact. Where possible, relevant friction points are therefore classified in terms of frequency, additional processing effort, waiting time, number of people involved, risk or economic impact. The methodology is deliberately designed not to quantify every small problem, but to identify the few high-impact deficits.
This also changes the discussion in the workshop. A participant no longer has to try to convince the group that a particular coordination loop is especially frustrating for them. If it is visible that the same loop occurs regularly, ties up several people and causes two days of additional waiting time, we have a different basis for prioritisation.
A subjective annoyance becomes an analysable deficit. A deficit can become a priority. And only then does the question arise of what the future process should look like.
This is exactly why it is important not to mix as-is assessment and to-be design. Anyone who discusses a solution immediately during the first description of a problem runs the risk of optimising symptoms whose cause has not yet been understood at all. Here too, AIS aiWorks Process Intelligence follows a clear sequence: first reconstruct reality, then understand causes and impact, and only afterwards design the future process.
The best workshop does not produce process pictures, but decisions
Of course, process pictures will still be created in a workshop prepared in this way. Target processes need to be visualised, responsibilities represented and new interfaces made understandable. So the whiteboard does not disappear.
But it is given a different task.
It no longer serves primarily to jointly find out what happened yesterday. It serves to make visible what should work differently tomorrow.
For me, that is the decisive change. We do not have to make workshops shorter just because AI can take over more tasks. We should make them more productive. If ten people work together for two days, then those two days should be used as completely as possible for the questions for which their shared perspective is actually necessary.
Individual process reality can be captured beforehand. Statements can be compared beforehand. Contradictions can be identified beforehand. Relevant elephants can be visible beforehand. And the most important deficits can already be prepared to such an extent that the group no longer has to discuss whether a problem exists at all.
Then the workshop no longer begins with the request to first draw the as-is process together. It begins with an already robust picture and a much more demanding question:
What do we want to change now?
This is exactly where we see the difference between traditional process discovery and AIS [ai]works Process Intelligence. We shift information gathering to where it can be done more efficiently, and reserve the shared time for the work for which organisations actually need to bring people together: assessment, design and decision-making.
That is why the best workshop does not begin with an empty whiteboard.
It begins where the real work only starts to become interesting.