See your own process in two weeks.
One process, your real event data, a working process model and a quantified list of where it stalls — in ten working days from the day the data arrives. No platform to buy first, no programme to sign up to. The pilot ends with a decision, not a proposal.
- Starts with
- An event-log export
- Shape
- Fixed fee, fixed scope
- Duration
- 10 working days from data access
- Then
- Scale it — or stop
What arrives on day ten
The process you actually run
Every case, every activity, every hand-off — discovered from the systems that recorded them, not from interviews.
Where it stalls, in numbers
Waiting time, rework loops and deviations from the designed process, quantified by product, team and case type.
A short list worth acting on
The handful of interventions the evidence supports, with the size of the prize on each — and an honest view of whether scaling is worth it.
Three fields and one hour.
Process intelligence works from data your systems already write. For each step a case passes through, we need to know which case, what happened, and when.
Field 1
Case ID
The order number, application reference, claim number or ticket ID — whatever ties one case together across systems.
Field 2
Activity
What happened at that step: "credit check completed", "document requested", "approved". Status changes are usually enough.
Field 3
Timestamp
When it happened. Mixed formats and time zones are fine; harmonising them is part of the first two days.
An export, or read access
A CSV or table extract of the last three to six months from each system the process runs through: workflow, ERP, CRM, core or case-management. Default connectors cover the common platforms.
One hour with the process owner
To agree what counts as a case, where it starts and ends, and which outcome matters — on time, first-time-right, within policy.
Is your process ready for a pilot?
Five questions, one minute, and an honest answer — including what to export if the answer is yes.
1.How many cases does the process handle a month?
2.How many systems does a case pass through?
3.Do those systems record status changes with a time?
4.Does one identifier follow the case across systems?
5.Is there an owner who can act on what is found?
Your result
Answer all five and the result appears here — with what to do about any gaps.
Send me the export checklist for my answers.
We reply with the fields to export from each system and what to ask IT for. Nothing else.
Sent. The checklist is on its way, and a senior team member will follow up within one business day.
How the ten days run.
- 1Days 1–2
Connect and define
Extract the event data, agree the case definition and the activities that matter, and settle what a good outcome looks like.
- 2Days 3–5
Harmonise and discover
Merge the logs into one case timeline, clean timestamps and identifiers, and produce the first discovered process model. You see it on day five.
- 3Days 6–8
Analyse
Bottlenecks, rework loops and conformance against the designed process, quantified and cut by product, team, channel and case type. Where the data supports it, a first model of which cases will breach.
- 4Days 9–10
Decide
A findings review with the process owner and sponsor: what we found, what it is worth, what to do first — and whether to scale, extend to a second process, or stop here.
High volume, several systems, timestamps that exist.
- Hundreds of cases a month or more — enough for the patterns to be real, not anecdotes.
- It crosses at least two systems or teams, so no one currently sees it end to end.
- The systems record status changes with a time — most workflow, ERP, CRM and case-management tools do.
- Someone owns the outcome and can act on what we find.
Processes we have run this on, or that fit
- Customer onboarding
- Order-to-cash
- Payment exceptions
- Claims handling
- Procure-to-pay
- Incident & service requests
- Loan origination
- Returns & refunds
Fixed fee. Your findings. Your call.
Agreed before day one
A fixed fee for a fixed scope: one process, ten working days. No change requests, no metre running.
Nothing to buy first
We are vendor-agnostic and sell no licences. The pilot runs on your existing process-mining platform where you have one, or on an evaluation instance under the platform vendor’s trial terms. Choosing a platform is a decision for after the pilot, if the results justify it.
The findings are yours
The process model, the analysis and the roadmap stay with you whether or not we continue — and the second process onboards far faster than the first.
Questions we're asked about the pilot
They usually are. Different formats, time zones, and case IDs that change name between systems are normal; harmonising them is what days one to five are for. What we cannot work without is some record of when each step happened.
No. If you already run one, we use it. If not, the pilot runs on an evaluation instance under the platform vendor’s trial terms, and the findings do not depend on which platform it was. Choosing one is a decision for after the pilot, if the results justify it.
On the day the data arrives. In most organisations, getting the export approved is the step that takes longest, and it is outside our control — so we help you write the request so it goes through first time, and day one is the day we can open the files.
A process owner for one hour at the start and two at the end, and someone who can export data or grant read access. We do not need a project team.
A demo shows you someone else’s process on sample data. The pilot shows you your own, from your own logs, with findings you can check against what your teams already suspect — and some they do not.
One of three things: scale across the full estate and monitor continuously; take the roadmap and act on it yourselves; or stop, with the findings in hand. All three are fine outcomes.
Ready to see your own process? Start a Conversation and tell us which one.