The Process-Selection Workshop, Before You Automate the Wrong Thing You have a shortlist of automation candidates: invoicing, repetitive customer emails, re-typing contract data into the CRM. Every department insists their process needs it most. This workshop settles the question of what to automate first — based on numbers, not on who argued loudest.
Ready to publish without client-provided data. --- # The Process-Selection Workshop, Before You Automate the Wrong Thing
You have a shortlist of automation candidates: invoicing, repetitive customer emails, re-typing contract data into the CRM. Every department insists their process needs it most. This workshop settles the question of what to automate first — based on numbers, not on who argued loudest.
Leadership has decided to try AI automation — but where to start? Finance points to manual invoicing, support points to repetitive email threads, HR wants first-pass CV screening automated. Everyone has an argument; nobody has numbers.
Without a shared way to compare candidates, the decision comes down to whoever spoke loudest in the meeting. A few months later it turns out the chosen process runs too rarely to pay off, or the input data is too inconsistent to automate cleanly. That is an expensive lesson, and two hours of structured conversation can save you from learning it the hard way.
We run a structured workshop where every submitted process is scored against the same criteria — not whoever pitched it best. You leave with one decision backed by a score, not a wish list.
We talk to the people who actually run the process day to day, not only their managers — they know how long a task really takes and where the undocumented exceptions live. Every process is scored on three dimensions: time spent, how standardised the input data is, and the risk of error once automated.
What we don't do: we never close a workshop with a ready-made AI implementation pitch regardless of the outcome. If none of the submitted processes qualify, we say so plainly instead of pushing the next paid stage.
This is a new service line, and we don't yet have published workshop results to show — we say that plainly rather than inventing references. Our first workshops run at a preferential rate in exchange for the right to describe the outcome, without disclosing client data, as a reference case.
The workshop runs for two hours, online or on-site at your offices. It includes our facilitator and — critically — the people who actually run the processes under review; without them, the scoring reflects an idea of the process, not how it really works.
Price is set once we see the list of processes — it depends on how many there are, how many departments are involved, and whether a confidentiality agreement is needed before discussing sensitive data. You know the price before we confirm the date, not after.
A written recommendation. Every workshop ends with a document scoring each process individually — not a verbal opinion you can't show anyone.
No pressure toward a pilot. If none of the submitted processes qualify for automation, we say so directly in the recommendation instead of bending the result to sell the next stage.
Data confidentiality. When sensitive data is involved, we sign an NDA and agree on anonymisation rules before any document leaves your company.
The same small team that runs the workshops later runs the measurement and pilot stages, so we take on a limited number of workshops per month — so the recommendation is never written between other projects.
Send us the processes you suspect could be automated, along with dates that work for you — we'll reply within a few business days with a proposed date and a firm price.
Every month without this decision is a month your team keeps doing manually a process that could already be automated — while a competitor who decided sooner pulls a step ahead.
Two-hour workshop, online or on-site, with the people who run the process day to day · every process scored on time spent, data standardisation and risk · written recommendation with reasoning · no pressure toward a pilot if nothing qualifies · you know the price before confirming the date.
Usually three to five — enough to score each one properly within two hours, rather than a shallow pass over ten.
No. It ends with a recommendation on which process to automate first. A quote for measurement, a pilot, or full implementation is prepared separately, once you decide to move forward.
Yes — at least one person who runs the process day to day. Otherwise the scoring reflects an idea of the process, not how it actually works.
We sign an NDA before the workshop and agree on anonymisation rules — client names and personal data never appear in the analysis or the recommendation.
Yes — it works equally well online or on-site; the choice depends on where it's easier to gather the people who run the processes under review.