Offerings

One partner, from the first question to your team running it alone.

Six things we do. Run together, they become an AI program. Underneath sits one accountable person.

Book a consultation
The offering, at a glance

We help you work out where AI is worth real money, build the first cases, get your people using them, and hand the whole thing to your team. Most firms sell one part of that. We run the full line, or step in at a single point.

Advise
The lightest way in.
Strategy & Program
The plan, and the engine that runs it.
Data & Platform
The foundation your use cases run on.
Build & Deploy
From proof of concept to something that runs.
Adopt, Operate & Hand Over
Where most AI stalls, and where we stay.
The person underneath it all
Fractional Head of Data & AI
One person across the whole journey, or any single part.
The pillars, in detail

Five delivery pillars, one operator

01

Advise

The lightest way in.

The lightest way to work with us. We help you decide where to invest, whether to build or buy, which vendor holds up, and how to bring the board along. Senior, independent, fast. It is the on-ramp to everything else, or a standalone answer to one hard question.

What sits inside
  • Vendor and build-versus-buy evaluation. Market scan, shortlist, feasibility and cost, a recommendation the executive team can act on.
  • Executive and board advisory. Bringing leadership up to speed, with a consistent story across the room.
  • Ad hoc guidance. Senior input on the pressing questions as they land.
02

Strategy & Program

The plan, and the engine that runs it.

We set the AI strategy, choose the use cases worth real money, and put the governance and operating model around them. The result is a central engine that gives you visibility and scale across sites, plants and functions, while your own teams keep ownership of their use cases. The point is to enable the business, not to gatekeep it.

What sits inside
  • Strategy and use-case selection. Where AI moves EBITDA, and where it does not.
  • Governance and responsible AI. Decision rights, guardrails, and the central-versus-local balance.
  • Operating model and org design. Who owns AI, and how the program supports the business.
  • Portfolio and investment management. Prioritised on value and reuse, so spend follows proven cases.
  • Risk and assurance. Monitoring, model governance, vendor risk.
03

Data & Platform

The foundation your use cases run on.

We build the platform and set the data strategy: ownership, access, and ingestion from the telematics, SCADA, historian and ERP/EAM systems your operation already runs. It is built so use cases scale and stay governable from day one, across the business rather than stuck in one site or function. No multi-year data clean-up before anything ships.

What sits inside
  • Data strategy. Ownership, access, governance, roadmap.
  • Platform build. Lakehouse, pipelines, AI-ready architecture.
  • Ingestion and integration. OEM telematics, SCADA and historians, ERP and EAM.
04

Build & Deploy

From proof of concept to something that runs.

The delivery engine. We take a single high-value use case from a bounded proof of concept, through a production pilot, into production, and keep it running a year on: monitored, maintained, still earning. ML, forecasting, computer vision, LLMs. The secure enterprise AI assistant is often the fast first step, the quick win that shows value while the bigger cases build.

Every proof of concept has a number it has to hit, a fixed budget, and a clock. It converts or it stops. We cap how many run at once, and tech and data spend follows the cases that prove out.

What sits inside
  • PoC to Pilot to Production. Fixed scope, fixed budget, a target it has to hit. It converts or it stops.
  • Secure enterprise AI assistant. The fast, low-risk wedge that stops people pasting company data into public tools.
  • The full technical toolkit. ML, LLMs, forecasting, computer vision.
Where to start, by industry
One or two proven cases, then what is possible next.
Mining
Predictive maintenance that fuses telematics, oil-sample and EAM data, taken from PoC to production pilot, catching breakdowns before they happen.
Possible next
Grade / ore-quality predictionHaulage & fleet optimisationComponent & tyre life predictionMill throughput from the geological model
Cement, Aggregates & Building Materials
Kiln optimisation using real-time process data to hold clinker quality while cutting fuel and alternative-fuel spend across the burn line.
Possible next
Kiln thermal / energy optimisationClinker quality predictionAlternative-fuel rate optimisationPredictive maintenance on kiln & mill
Energy & Power
Turbine and boiler failure prediction that fuses vibration, temperature and load data to flag bearing and blade degradation weeks before an outage.
Possible next
Turbine / boiler failure predictionDemand forecastingHeat-rate optimisationLoss detection
Oil & Gas
Rotating-equipment condition monitoring across pumps, compressors and turbines, cutting unplanned shutdowns and closing integrity risk.
Possible next
Asset-integrity predictionProcess optimisationPredictive maintenance on rotating equipment
Industrial Manufacturing
Vision-based in-line defect detection running at line speed, replacing sample-based QA and pulling scrap rates down without slowing throughput.
Possible next
In-line quality / defect detectionSupply-chain forecastingPredictive maintenanceThroughput optimisation
Machinery & Equipment (OEM)
Predictive maintenance turned into a service line: OEM telematics powering a recurring revenue stream customers pay for, not a support cost.
Possible next
Predictive maintenance as a serviceUsage analytics feeding product designRemote diagnosticsOffline scan verification in the field
Construction & Infrastructure
Heavy-plant utilisation analytics that turn fleet telematics into uptime, plus cross-project cost learning that stops the same overruns twice.
Possible next
Asset & fleet utilisationCross-project cost learningPredictive maintenance on heavy plant
Cross-functional (any industry)
A secure internal AI assistant that replaces public tools, and finance forecasting.
Possible next
Document intelligenceReporting automationCommercial / contract analyticsAsk-the-manuals assistant
05

Adopt, Operate & Hand Over

Where most AI stalls, and where we stay.

This is the work that gets people using what was built: super users, an AI community, role-specific enablement, a Center of Excellence, KPI tracking tied to business outcomes rather than technical vanity metrics. And it is the ongoing operation: monitoring, iteration, and keeping the models honest as conditions change on the ground.

What sits inside
  • Change and adoption. Super users, AI community, role-specific enablement, Center of Excellence.
  • KPI tracking. Value tied to business outcomes, not technical vanity metrics.
  • Ongoing operation. Monitoring, iteration, keeping models honest as conditions change.
Hand Over
Your team ends up owning it.

A good program is one you no longer need us to run. We transfer the platform, the operating model, the use-case pipeline and the adoption engine to your own people. Your team ends up owning the models and the roadmap, able to run the next use cases themselves, with a maintenance and governance model that holds after we step back. We stay as far as you want us to, and no further.

  • Capability transfer. Your team owns the platform, the models and the operating model.
  • Future use-case pipeline. A way for your people to find and run the next ones themselves.
  • Maintenance and governance model. What keeps the platform and models healthy after we step back.
  • A clean exit, or a lighter ongoing role. Your call, not ours.
The delivery pillars, run together

This is an AI Program.

The AI Program is not a separate offering. It is what the delivery pillars become when you commit past a single project: strategy, platform, use cases, adoption and hand-over, run as one coordinated effort with shared governance, one portfolio and one operating model. It is the difference between a pile of initiatives and a program that compounds.

On the model above, it is the bracket over the delivery pillars. Advise sits outside it, as the entry point before you commit.

The resource beneath all of it

Fractional Head of Data & AI

Beneath every pillar sits one person: a fractional Head of Data & AI who owns your data and AI agenda the way a permanent leader would, without the cost or commitment of a full-time executive. Bring us in for as much or as little as you need, from every pillar to a single proof of concept, an adoption push, or a vendor decision. The role flexes to what you need now, and grows or shrinks as that changes.

Three ways to engage the same team
01
Option one
As single projects
Buy one pillar on its own: a vendor decision, a PoC, an adoption push, a platform build. A bounded first project with a clear outcome.
02
Option two
As an AI Program
Have us run the delivery pillars together, with shared governance and one portfolio.
03
Option three
As your fractional Head of AI
Bring us in to lead the whole thing, accountable for the agenda, building the capability into your team, and handing it over.
The path through

How it fits together

01
Start anywhere

Advice, a single PoC, a secure assistant, an adoption push. Each pillar is a valid, low-commitment way in.

02
Expand into a program

A use case needs the platform. The platform invites the program. The delivery pillars cohere into one AI Program that compounds.

03
End in hand-over

The destination is your team owning it. Embedded while needed, then a clean transfer.

Start with one question, or the whole program.

Consultation

One partner, from the first question to your team running it alone.

Bring us the pillar that matters most right now: a vendor decision, a PoC, a platform build, or an adoption push. Or the whole program end to end. Either way, the conversation starts the same. Tell us where it hurts.

Book a meeting
Consultation