Home / The Operations X-ray

The Operations X-ray

In three weeks, see exactly where your value is leaking — and what it is worth.

A fixed-fee X-ray of one high-value process, run on your own operational data. Not a benchmark. Not a workshop. Your actual process, reconstructed from the record your systems already keep.

You finish with a ranked list of what is costing you money, what it is worth in dollars, and what to do about it first.

  • One process
  • Three weeks
  • Fixed fee
  • Read-only data

Your systems already recorded what happened. We read it back.

Every purchase order, work order, approval and service ticket leaves a timestamped trail in your ERP, EAM and service tools. Process intelligence reads that trail and rebuilds how the work actually ran — every case, every step, every detour, every time it went round again.

That matters because the process on the wall chart and the process in the system are rarely the same thing. Workshops capture what people believe happens. The event log captures what did happen, across every case, with no recall bias and no sampling.

The X-ray takes that reconstruction and answers five questions a board or an executive team can act on.

What we look for

One process. Five questions.

Same three weeks, same single process. We read it through five lenses, because the same event log answers all five. Every answer comes back with a number attached and a named owner.

01

Where value is leaking

Money lost inside the process itself — not in the budget line, in the running of the work.

What we measure
Cost and time per case, rework loops, waiting between steps, manual touches, duplicate activity, penalties and missed terms.
What you get
The top three to five leaks, ranked, each with an indicative dollar value and the step where it starts.

"Eleven per cent of invoices were paid without a matching goods receipt, and 0.6% were paid twice — because the receipt was posted after the invoice cleared."Example of a finding, procure-to-pay

02

Whether work follows the process you approved

Conformance. The designed process is compared against every real case, and we count the departures.

What we measure
Approval steps skipped or performed after the fact, delegation thresholds worked around, mandatory fields left empty, sequence reversed, cases closed without evidence.
What you get
A conformance rate for each control, the exact points where it breaks, how often, and by which route. Evidence an auditor will accept.

"Capital requests were split into two below the $250k delegation limit in 34 cases, which moved approval two levels down."Example of a finding, capital approval

03

What you can safely automate or hand to AI

Automation applied to a broken process makes the mess faster and harder to see. This lens separates what is ready from what is not.

What we measure
How many distinct paths the work actually takes, the volume sitting on each, how rule-driven each step is, and whether the data an agent would need is complete and reliable.
What you get
A shortlist of automation candidates with real volumes behind them, and an explicit "not yet" list with the reason for each. Redesign first, automate second.

"Sixty-eight per cent of service requests followed four paths. The remaining 32% followed 190. Automate the four; redesign before touching the rest."Example of a finding, service request to resolution

04

Whether your ERP is delivering what the business case promised

Most ERP value is lost after go-live, quietly, as people build workarounds. This lens finds them.

What we measure
Standard functionality switched off or bypassed, steps that have moved to spreadsheets and email, custom code doing work the platform already does, licensed modules with no usage.
What you get
A clean-core view: what is custom that need not be, where licence spend is not being used, and what to fix before your next upgrade rather than after it.

"Three-way match was disabled for 40% of vendors. Approvals for those vendors were happening in email and re-keyed the next day."Example of a finding, procure-to-pay

05

Whether your benefits are real and measurable

Benefits management fails for a boring reason: nobody set a baseline, so nobody can prove the change worked.

What we measure
Each committed benefit traced back to the process metric that would have to move for it to be true, and whether that metric was ever baselined.
What you get
A measured baseline you did not have before, and one tracking line per benefit claim — so the next report is evidence rather than opinion.

"The business case claimed a 15% cycle-time reduction. No baseline was ever captured, so the benefit could be neither proved nor disproved."Example of a finding, benefits review

What it looks like

The process you approved, and the process you have

The chart on the left is the process as designed. The chart on the right is the same process as the event log records it, over roughly nine thousand cases.

Designed process compared with the actual process The designed process runs as five steps in a straight line. The actual process has the same five steps but adds a rework loop back from approval to request, a path that skips approval entirely, and a path that skips goods receipt so that the invoice is paid unmatched. AS DESIGNED One path. Five steps. 100% conforming. Request Approve Order Receipt Pay AS IT ACTUALLY RUNS Same five steps, plus three routes nobody approved. Request Approve Order Receipt Pay rework 1,420 cases approval skipped 612 cases · $2.1m committed paid without receipt 980 cases · no price control Approved path Departure from the approved path

Scroll the diagram sideways to see all three departures.

IllustrativeConformance view of a procure-to-pay process. Three departures account for most of the value at risk. Each one is countable, and each one has an owner.

Where the cost of a process actually goes

Finance sees one number for what a process costs. The event log splits that number into the part that produced something and the part that did not.

Breakdown of total process cost Of the total cost to run the process, 38 per cent is value-adding work, 34 per cent is waiting and handoffs between teams, and 28 per cent is rework and failure. 62 per cent of the cost does not produce anything. 38% Value-adding work 34% Waiting and handoffs 28% Rework and failure 62% of the cost produced nothing This is the part that is addressable — and the part no dashboard reports separately.
IllustrativeThe X-ray reports this split for your process, in your currency, with the biggest contributors named.

Start to finish

What three weeks actually produces

A worked example on procure-to-pay, from the data we take to the decision it supports.

Procure-to-pay, 8,900 cases, twelve months

Illustrative — figures are realistic placeholders, not a client result.

What we took
Four columns from the ERP: case ID, activity name, timestamp, and the role that performed it. No vendor names, no prices, no contract terms, no personal data.
What we found
Invoices were being paid before goods receipt in 980 cases, which removed price control at the point it matters. A further 612 purchase orders were raised after the invoice had already arrived — the approval was retrospective. Average time from request to payment was 34 days against a designed 11.
Why it was happening
Three-way match had been switched off for a group of vendors during go-live to clear a backlog, and never switched back on. The workaround became the process.
What it was worth
Duplicate and unmatched payments, lost early-settlement discounts and the labour cost of manual matching came to an indicative $2.4m a year, with $1.6m of it in a single vendor group.
What was decided
Re-enable three-way match for that vendor group in the current quarter — a configuration change, not a project. Two of the four remaining leaks went into the redesign backlog. One was accepted and closed.

Lens 03 in practice

Automate the pattern, not the mess

Before you put an agent or a bot on a process, you need to know how many different ways that process is actually performed. Where the work is concentrated on a few paths, automation is straightforward and the payback is quick. Where it fragments into hundreds of one-off routes, automation just encodes the chaos.

The X-ray gives you that split as a number, so the automation decision stops being a matter of opinion.

Concentration of cases across process paths Sixty-eight per cent of cases follow just four repeatable paths and are ready to automate. Thirty-two per cent of cases are spread across 190 different paths and need redesign before automation. 4 paths 68% of cases · ready to automate 68% 190 paths 32% of cases · redesign first 32% Four paths carry two-thirds of the volume. Automating them is a contained, measurable change.
IllustrativePath concentration, service request to resolution.

Lens 04 in practice

Where ERP benefits quietly stop arriving

ERP business cases assume the benefit curve keeps climbing after go-live. In practice it flattens, because the organisation reverts to workarounds that were never meant to be permanent — a match switched off, an approval moved to email, a step done in a spreadsheet.

Because those workarounds live in the event log, the gap can be measured rather than argued about, and each contributing cause can be named.

ERP benefit realisation, business case compared with actual The business case projected benefits rising steadily for eight quarters after go-live. Actual benefits rose for the first two quarters, then flattened well below the projection, leaving a widening shortfall. 100% 0 Go-live Q8 Business case Actual The gap is not a mystery. It is a list of workarounds. Each one is visible in the log, countable, and assignable.
IllustrativeBenefit realisation after an ERP go-live.

Lens 05 in practice

A benefit you cannot measure is a benefit you cannot claim

Most benefit registers break in the same place. A financial target is set, a project is delivered, and nobody can connect the two, because the operational metric in between was never measured before the change.

The X-ray fixes that by working backwards. Each committed benefit is traced to the process metric that would have to move for it to be true, and that metric is baselined from the event log — including for the period before the change, because the historical record is already there.

The benefits chain from process change to financial outcome Four linked stages: a process change produces a measurable process metric, which drives an operational outcome, which produces a financial result. Each stage names an owner and a measure. THE CHANGE Re-enable three-way match for vendor group B PROCESS METRIC Unmatched payments 980 → target 0 OPERATIONAL OUTCOME Manual matching effort down 4.2 FTE-equivalent FINANCIAL RESULT $1.6m annualised, owned by the CFO Every link is measured from the same event log, so the claim can be audited end to end.

Scroll the diagram sideways to see the full chain.

IllustrativeOne benefit, traced from the change that causes it to the number the CFO reports.

Scope, tight by design

One process, chosen with you from a short menu:

  • Work and maintenance management
  • Capital approval
  • Procure-to-pay
  • Outage and shutdown execution
  • Service request to resolution

Everything else is deliberately out of scope. One process, one X-ray — that is what makes three weeks credible. If a second process matters, it is a second X-ray, not a longer one.

What you get

Six deliverables. All of them short, all of them evidenced, none of them a slide deck about methodology.

  • The reconstructed as-is flow — how the work actually runs, built from the record rather than from interviews.
  • The conformance picture — where real cases depart from the approved process, how often, and by which route.
  • Where time and cost go — the split between value-adding work, waiting and rework, in dollars.
  • The top three to five leaks, ranked — by value at stake, asset criticality and regulatory exposure.
  • An automation readiness call — what is safe to automate now, and what needs redesign first, with the reason.
  • A one-page executive readout — presented in a findings session, with a measured baseline you can track against.

How it works

01

Scoping call

We agree the process, the data path and the fixed fee. Sixty minutes, no obligation.

02

Performance discovery

The minimised event log is prepared, in your environment or under NDA. The three-week clock starts here.

03

The X-ray

We reconstruct the real flow, measure conformance against the approved process, and rank the leaks by value at stake.

04

Readout

A one-page executive readout and a findings session with the people who can act on it. Three weeks, fixed.

Where the data lives

In your environment — nothing leaves your boundary. This is the default for defence and SOCI-regulated operators. Alternatively: a minimised, de-identified extract under NDA, held and processed in Australia, destroyed on completion.

What we take

A minimised event log — when things happened, not what was in them. Typically four columns: case ID, activity, timestamp, role. No personal data, no commercial terms.

How it connects

Read-only. Never connected to live or OT systems. Nothing is written back, and no agent or automation is deployed during the X-ray.

Commercials

A single fixed fee, agreed on the scoping call and confirmed in writing. If you proceed to Value Diagnostic & Delivery within 90 days, the X-ray fee is credited in full.

No variations, no time and materials, no scope creep — the scope is one process and the duration is three weeks.

Book the scoping call

Sixty minutes to agree the process, the data path and the fee. If the data is not there to support an X-ray, we will tell you on that call rather than after you have paid for one.