Process Intelligence Consulting: What a Partner Can Deliver — article illustration

Put real evidence behind one client process

Process Mining

Process Intelligence Consulting: What a Partner Can Deliver

A partner's guide to process intelligence consulting: the roles you can take, how a client's data becomes evidence, and how licensing and margin work.

Process intelligence consulting turns a client’s own event data into evidence the client can act on. A partner can deliver all of it on ProcessMind, or only the part the client cannot staff: the data engineering, the analysis, the dashboards, the simulation, or the monitoring that keeps the measures alive after the project ends.

This page is written for consultancies, system integrators and managed service providers. If you are a client who found it: nothing here is hidden from you. We publish our prices, and we publish how partner delivery works too.

What can a partner do with process intelligence?

A partner can do everything the client needs done with the data, from getting it in to explaining what it shows. In practice we see five roles, and most partners start with the analysis.

Role you take What it looks like Who it suits
Data engineer Extracts the event log, agrees the case notion, maps activity names and gets one usable dataset Clients whose systems were never designed for process analysis
Analyst Discovers the real flow, variants, rework and waiting, and compares it with the agreed model The most common first engagement
Dashboard builder Turns the measures into dashboards the client’s owner can read and defend Clients who want the numbers in their own operations rhythm
Simulation specialist Tests a proposed change on a modeled representation before the client commits resources Clients deciding between improvement options
Monitoring partner Runs the datasets, dashboards and health checks on a cadence and reports on them Managed service providers, and clients without an analytics team

Two rules hold across all five. First, the analysis attaches to a process someone already agreed on and owns; the getting started guide covers how that picture is built. Second, the output is evidence for a decision, not the decision itself. A dashboard nobody owns changes nothing.

Process intelligence views showing process flow, variants and performance measures
Figure 1 — The deliverable is evidence the client's owner can act on: what happens, where it waits, and what a change would do.

When is a client ready for process intelligence?

A client is ready when three things exist, and not before:

  • An agreed process with a named owner. If the process is not agreed, the dashboard answers a question nobody owns. The process architecture partner guide covers the step that comes before this one.
  • A case identifier and timestamps in the source systems. One order, one ticket, one invoice: the log needs something to follow through the process.
  • Access to the data owner. Extraction is rarely blocked by technology; it is blocked by a person who has not been asked yet.

If the client has an event log but no agreed process, run the architecture step first. If the process is agreed but there is no usable data, say so early: the preparation guide is a better conversation to have before you quote an analysis.

What does process intelligence consulting look like for a partner?

The engagement shape is repeatable, and it is shorter than most clients expect to the first useful answer.

  1. Assess the data with the question. Start from the question the owner wants answered, then check whether the systems can answer it. That order prevents an expensive extraction that answers nothing.
  2. Build one dataset. Agree the case notion, map the activities and fix the time range. The dataset mapping is where most of the effort sits, and it is honest work to bill.
  3. Discover and compare. Show the real flow next to the model the owner approved: variants, rework loops, waiting time and the steps the documentation does not mention.
  4. Put the measures where the owner works. A dashboard that the owner opens is worth more than an analysis deck. Dashboard viewers read it without a paid seat.
  5. Test the change before it is made. Use simulation to compare the options the owner is actually choosing between, with the assumptions visible.
  6. Hand over the monitoring. Set up process health and conformance checking so the client sees drift without waiting for the next project.

Steps 3 to 6 are where a partner earns the follow-on work. Steps 1 and 2 are where a partner loses money by guessing.

How does partner licensing work for intelligence work?

Partner licensing has two parts: a fee per client environment, and a pool of seats that moves with your consultants.

Per client. Every client tenant is an isolated environment with its own data, datasets, dashboards and users. A client tenant includes 100 portal viewers and 3 dashboard viewers, and the per-tenant fee steps down as the number of client tenants in your contract grows. For a managed service, that is the environment your client’s people log into; the environments and tenants guide explains the isolation.

Per seat. Intelligence work needs Process Intelligence seats, because that is where mining, dashboards, simulation and conformance live; Foundation and Architecture seats do not include event-log upload. The seats come from one pool, so a consultant who moves from one client’s analysis to another’s keeps the same paid seat.

Per user. Client analysts can hold seats and work in their own environment, and that is often the best outcome for a long engagement: the client’s team learns the analysis while you review it. Readers never need a paid seat, because portal and dashboard viewers are included.

Usage. The client pays for usage rather than for a process count: AI credits, data processing, simulations, dashboards, exports and API calls are metered and visible, and a credit pack covers an unusually heavy month. The usage and credits guide explains how the allowance works.

Price. Partner pricing starts at 20% off the list. Seats take the better of that partner rate and the annual volume ladder, never the sum, and the tenant fees and packs get the partner rate. The pricing calculator shows the whole scenario, on the same list a client sees on the pricing page.

The calculator opens on a scenario when you follow these links, and you can change every number afterwards:

Price a monitoring service before you sell it

What does the commercial model look like for data work?

You buy at the partner rate and sell at your own, and intelligence work has one advantage over architecture work: it repeats.

  • Bill the data work as consulting. Extraction, case notion and mapping are the effort a client cannot estimate. Price them as work, not as a licence discount.
  • Sell monitoring as the recurring service. A dashboard that is refreshed, checked and explained every month is a subscription the client understands. Process health gives that service a measure.
  • Resell or bundle the licences. Quote the platform next to your analysis with your own markup, or fold it into a managed service where the client sees one fee.
  • Let volume improve the margin. Volume discounts apply across all your paid seats, so the seats you already sell get cheaper as your commitment grows, and a seat pool shared across clients keeps the bench small.

Keep the exit honest. Datasets, models and dashboards belong to the client’s workspace, and the exports exist so their knowledge is never trapped. A client who knows they can leave is a client who renews.

We have seen what happens when mining is sold as the product: a beautiful dashboard, a workshop, and then nothing, because no one at the client owned the process the dashboard described. The partners who build a practice do it the other way round: they earn the right to the data by helping the client agree on the process first, and then they keep coming back with evidence. That is a longer sale and a much better business.

Christiaan Esmeijer
Christiaan Esmeijer Co-founder and CEO

How do you run the first intelligence engagement?

The first dataset decides whether the client trusts the next one. Run it as a measurement project with a named consumer, not as a tool demo.

  1. Start from one owner's question

    “Why do orders wait before credit approval?” is a better first question than “what does our process look like?”. One owner, one question, one answer they can use.
  2. Check the data before you promise an answer

    Confirm the case identifier, the timestamps and the systems that hold them. If the case notion is ambiguous, agree it with the owner before extraction.
  3. Build one dataset and document its limits

    Fix the time range and the activity mapping, then write down what the data cannot show. A dataset with known limits is trusted; one presented as complete is not.
  4. Show the flow against the agreed model

    Discovery is most convincing as a comparison: this is what you said happens, this is what the log shows. The process mining guide covers the views that make the difference visible.
  5. Put one measure in front of the owner

    One or two measures they will actually open beat a wall of charts. Give the owner and their readers dashboard access and check that they use it.
  6. Set the monitoring and the next question

    Agree the cadence, the health thresholds and the question for the next cycle. That cadence is the recurring service, and it starts on day one of the handover.

What do partners get wrong in intelligence work?

  • Mining an unowned process. The analysis is correct and nothing happens, because nobody is accountable for the result.
  • Promising a clean dataset. The first extraction usually has gaps; say so before you start, not after the dashboard looks odd.
  • Ignoring the case notion. Cases that do not match across systems produce a flow that is plausible and wrong.
  • Building dashboards nobody asked for. A measure the owner did not request rarely survives the second month.
  • Selling a one-off analysis as a programme. Without monitoring, the client’s understanding decays from the day you leave.

What should you not promise?

Do not promise a percentage saving from the tool. Mining shows what happens; the client’s owner decides what to change, and the change still has to be implemented and measured. Do not promise that the data will be clean, or that the log will tell the client what to do with it. And do not promise that one dashboard will change behavior: the cadence and the owner are what do that.

What you can promise is a defensible answer to a real question, built on the client’s own data, with the assumptions and the limits written down. That is what makes the next question easy to sell.

If the client’s processes are not yet agreed, start with the process architecture partner guide, which covers the landscape, the owners and the published version this analysis attaches to. For the pricing logic, including why we price per seat and per tenant rather than per process, the pricing explainer goes through it.

Where to Go From Here

You have a client with a process worth measuring, an owner who wants an answer, and data that can be extracted.

Frequently Asked Questions

A partner turns a client's event data into evidence the client can act on: extracting and preparing the data, agreeing the case notion, building datasets and dashboards, testing improvement scenarios in simulation, and setting up the monitoring that keeps the measures visible after the project.

When the process is already agreed and has a named owner, and the source systems hold a case identifier and timestamps. Without an agreed process, a dashboard answers a question nobody owns; without usable event data, mining has nothing to read.

Seats sit in one pool that you move between client tenants, and each client gets an isolated environment for a per-tenant fee that steps down as the number of clients grows. Partner pricing starts at 20 percent off the list, seats take the better of that rate and the annual volume ladder, and the tenant fees and packs get the partner rate.

Mining, dashboards, simulation, conformance checking and process health need a Process Intelligence seat. Each Intelligence seat includes 10 portal viewers and 3 dashboard viewers, and each client tenant includes 100 portal viewers and 3 dashboard viewers of its own. Client analysts can hold seats; readers do not need one.

Usage is metered, not the number of processes: AI credits, data processing, simulations, dashboards, exports and API calls are visible in the workspace, and credit packs can be added when a heavy dataset month exceeds the allowance.

Do not promise a percentage saving, a clean first dataset, or that the log will decide what to change. Mining shows what happens; the owner still chooses what to change, and the improvement still has to be tested and implemented.

Related Blog Posts

Receive expert insights on process mining and workflow optimization in your inbox
BPO Process Outsourcing: Measure Before You Outsource

Process Mining

BPO Process Outsourcing: Measure Before You Outsource

Business process outsourcing moves a process to an external provider. Measure how it runs, redesign what you can, and give both sides one shared dashboard.

Process Modeling and Process Mining: Better Together

Process Modeling

Process Modeling and Process Mining: Better Together

Learn what process modeling and process mining each show, where they differ, and how conformance checking connects them.

ETL Tools for Process Mining: What You Need

Data

ETL Tools for Process Mining: What You Need

When process mining needs ETL, what the extract has to contain, and why ProcessMind loads whatever your existing pipeline produces.

How Process Intelligence for Sustainability Makes Claims

Digital Transformation

How Process Intelligence for Sustainability Makes Claims

Trace sustainability figures from reports to the process events behind them.

Design better processes. Build a connected architecture. Stay in control.

Get instant access with no credit card and no waiting. Turn the way your organization works into clear, connected process designs.

Build your process architecture, define ownership and controls, and align roles and responsibilities across every level.

Start your free trial and create one reliable foundation for governing, managing, and continuously improving your processes.