Skip to content

Selected work

Results we're accountable for.

A look at what we've advised on, built and run for clients, the problem, the approach, and the measurable outcome.

Catching batch quality issues before they ship

A speciality chemicals manufacturer

Catching batch quality issues before they ship

Quality problems surfaced in the lab, after the batch was made — sometimes after it shipped. We put the signals in front of operators as the batch ran, so the save happened on the line, not in a recall.

Off-spec batches
−22%
Time to detect a deviation
Hours → live
Rework and scrap
−17%
Turning a document backlog into minutes, not weeks

An enterprise services firm

Turning a document backlog into minutes, not weeks

A team of specialists spent their days keying data out of contracts and forms. We built an AI pipeline that reads the documents and drafts the entries — with a human confirming, not transcribing.

Processing time per document
Days → minutes
Manual keying removed
~70%
Exceptions caught earlier
+3x
One control tower for a utility's field operations

A regional energy & utilities operator

One control tower for a utility's field operations

Crews, faults and jobs lived across radios, spreadsheets and three systems that didn't talk. We built a single live picture of the field — and response times, and the guessing, both came down.

Fault response time
−31%
Jobs completed first visit
+18%
Manual dispatch effort
−40%
An AI radar that turns one sentence into a meeting-ready target list

A global management consultancy

An AI radar that turns one sentence into a meeting-ready target list

Business development ran on weeks of manual research for a single target list. We built an AI lead-generation platform: describe the client you want in plain English, and in about a minute it returns the right companies, the real decision-makers, verified emails and a briefing pack.

Research to target list
Weeks → minutes
Decision-makers found
231
Contacts verified real
90%
An AI knowledge platform that answers from approved enterprise content

A 28-division enterprise

An AI knowledge platform that answers from approved enterprise content

Years of institutional knowledge were scattered across SharePoint sites nobody could search. We built a governed, AI-powered knowledge platform on Microsoft 365 and Copilot, so people get trusted answers in seconds, grounded only in approved documentation.

Departments unified
28
Enterprise data indexed
150+ GB
Document governance
Automated
A modern, AI-ready SharePoint workplace for a global manufacturer

A multinational manufacturing and supply-chain organisation

A modern, AI-ready SharePoint workplace for a global manufacturer

A legacy SharePoint estate held years of SOPs, quality manuals and audit records nobody could find. We modernised it into a secure, governed digital workplace for 500+ users across 31+ sites, and laid the information architecture that makes enterprise AI possible.

Facilities unified
31+
Enterprise users
500+
AI readiness
Enterprise-wide
A bespoke field-operations platform, from strategy to rollout

An industrial field-services company

A bespoke field-operations platform, from strategy to rollout

Off-the-shelf tools couldn't fit the way this business worked. We advised on the approach, built a bespoke platform, and rolled it out to hundreds of field crews across sites.

Job admin time
−60%
Field adoption
94%
Strategy to launch
5 months
A live production view for a mining operation

A mining and minerals operator

A live production view for a mining operation

Production data lived in a dozen disconnected systems, so problems only surfaced in the next day's report. We built one live view, and unplanned downtime started falling.

Unplanned downtime
−28%
Throughput
+12%
Reporting lag
Days → live
Joining berth, yard and gate into one flow for a container terminal

A container terminal operator

Joining berth, yard and gate into one flow for a container terminal

The quay, the yard and the gate each optimised for themselves, so congestion moved around rather than clearing. We joined them into one live picture, and the whole terminal started to flow.

Truck turnaround
−26%
Quay throughput
+11%
Yard dwell
−15%
Line-side visibility and full traceability for an auto components supplier

A tier-1 automotive components supplier

Line-side visibility and full traceability for an auto components supplier

Shortages showed up at the line, not before it, and a recall query meant days of paper. We built inbound and line-side visibility and part-level traceability, so shortages are seen early and a batch can be traced in seconds.

Line stoppages
−41%
Premium freight
−33%
Batch traceability
Hours → seconds
Predictive maintenance that cut unplanned downtime for an oil & gas operator

An oil & gas operator

Predictive maintenance that cut unplanned downtime for an oil & gas operator

Critical equipment was failing between scheduled inspections, each failure expensive and dangerous. We built models that see failures coming, and a team that acts on them.

Unplanned downtime
−35%
Maintenance cost
−18%
Early warning
Days ahead
One live view of cost, schedule and progress for an EPC contractor

A large EPC and infrastructure contractor

One live view of cost, schedule and progress for an EPC contractor

Progress lived in site diaries and month-end spreadsheets, so overruns were discovered late and reported slowly. We built field progress capture and a single live view of cost, schedule and materials, and the surprises got smaller.

Schedule slippage
−19%
Rework
−24%
Monthly reporting
Days → hours
Cutting a logistics platform's cloud bill by 41%, while it grew

A national logistics operator

Cutting a logistics platform's cloud bill by 41%, while it grew

A fast-growing platform's cloud cost was rising faster than its revenue. We reshaped the architecture so cost tracked demand, then locked in the savings.

Cloud spend
−41%
Order volume
+30%
P95 latency
−35%
Lifting yield and cutting off-grade at an integrated steel plant

An integrated steel producer

Lifting yield and cutting off-grade at an integrated steel plant

Yield and quality data sat in the historian and the Level-2 systems but never came together, so off-grade was explained after the heat, not prevented. We joined the data and gave the plant models that flag risk while there is still time to act.

Prime yield
+3.4 pts
Off-grade and scrap
−22%
Energy per tonne
−7%
A shop-floor system that lifted OEE for a manufacturer

A discrete manufacturer

A shop-floor system that lifted OEE for a manufacturer

Shop-floor performance was tracked on paper and guessed at in meetings. We digitised it, and gave the plant the numbers to fix what was actually holding it back.

OEE
+14 pts
Scrap rate
−22%
Manual reporting
Eliminated