Free / 30 minutes / one workflow

Name the
repeat offender.

You do not need an AI strategy deck. You need to know where the work gets stuck.

  1. 01

    Bring one real example.

    A recent enquiry, handover, report, or approval. Describe the process without sharing sensitive records.

  2. 02

    Trace the friction.

    Who touches it? Where does it wait? What must happen before it can move?

  3. 03

    Choose the next move.

    A focused pilot, a simpler manual process, or no project at all. The answer should earn its place.

No obligation to start a project. Detailed technical design is scoped separately.

Tell us where to look

What happens next

A conversation.
Then an honest decision.

We read the workflow.

We check whether the task is specific enough for a useful conversation and whether the data or risk needs special handling.

We ask for one example.

The current process matters more than an idealised brief. We trace what actually happened.

We recommend a next step.

That may be a scoped pilot, a simpler process change, more discovery, or no automation project.

What a workflow audit examines

Thirty minutes should produce
better questions.

The free workflow audit is a focused first conversation about one recurring process. It is not a disguised sales deck and it is not a complete technical specification.

The purpose is to identify whether the problem is specific enough, valuable enough, and safe enough to investigate further. If it is not ready, the useful next step may be better data, a clearer manual process, or no automation project.

01

Trigger and volume

What event starts the work, how often does it happen, and does the team recognise it consistently? A vague trigger creates duplicate runs, missed records, and arguments about whether the automation behaved correctly.

02

Inputs and systems

What information is needed, where does it come from, which system is authoritative, and how reliable is access? We look for missing identifiers, inconsistent formats, manual exports, and information that should not enter an AI service.

03

Decisions and exceptions

Which steps follow an explicit rule, which require interpretation, and which must stay with a person? We identify common exceptions, high-consequence cases, approvals, opt-outs, and the manual fallback.

04

Outcome and measurement

What should be faster, more complete, more consistent, or more visible? We separate activity measures such as messages sent from business measures such as qualified appointments, completed handovers, or reporting time.

What to bring

Bring one recent example, the names of the tools involved, approximate frequency, the current owner, and the failure that matters. Screenshots with confidential information removed can be discussed later if needed. A useful example shows the normal path and at least one exception, because exceptions usually reveal the real implementation difficulty.

What happens afterwards

You receive an honest direction: a possible pilot, a paid discovery need, a simpler process recommendation, a data-readiness task, or a clear reason not to automate the workflow. If a pilot makes sense, the next step defines scope, success measures, required access, human checkpoints, and who will own the workflow after launch.