Services

Start where the uncertainty is.

You may need to choose a pilot, prove a workflow, put a prototype into real use, or keep a live AI service healthy. Graymatter offers a clear engagement for each point.

Every stage has an agreed fixed price, ends with evidence and a decision, and stands on its own. Continuing is not assumed.

01Opportunity MapFixed-price product 02Working PrototypeFixed-price product 03Pilot to PracticeFixed scope and stage price 04Ongoing AI PartnerFixed monthly service

A facilitated technology working session — DISCOVERY IN THE WORKSponsor + users + systems + controls

Built with the people who do the work

We do not discover a workflow from the conference room alone.

The sponsor sets the outcome. The people doing the work show us the exceptions, workarounds and judgement that a process map misses. IT and control owners shape the safe information and integration path.

That is how a fixed-price scope becomes credible: the boundary is understood before the build, and the people who will judge it help define what good looks like.

Fixed-price product · Typically 2 weeks

Opportunity Map

Choose a first move you can defend.

For leadership teams with pressure to act, a long list of ideas, and no shared way to decide which one deserves time and money.

What we do together

Executive intent and constraint session
Interviews with workflow owners and real users
Current AI use, systems and data snapshot
Candidate longlist and comparison workshop
Value, feasibility, data, risk and adoption scoring
Initial accountability and stakeholder map

What you leave with

Ranked pilot candidate board
Recommendation for the first test
Explicit not-yet list
Pilot brief and success measures
90-day decision roadmap

DECISION GATE

Proceed with one candidate, reshape it, do readiness work first, or stop.

Fixed-price product · Typically 4–6 weeks

Working Prototype

Replace opinions with something your team can use and test.

For a chosen workflow that needs to be made tangible before a production, platform or change commitment.

What we do together

Workflow boundary and prototype brief
Approved-data and tool decisions
Working build against representative tasks
Evaluation set and acceptance criteria
Difficult-case and failure-mode tests
User demonstrations and structured feedback

What you leave with

Working prototype
Source and data path
Quality and failure findings
Human review and escalation design
Pilot recommendation
Indicative production scope

DECISION GATE

Pilot it, revise the design, hold for a dependency, or stop.

Fixed scope and stage price · Typically 6–12 weeks

Pilot to Practice

Make the workflow useful on an ordinary Tuesday.

For teams with a promising prototype that now has to survive real users, permissions, edge cases, operating responsibilities and change.

What we do together

Bounded production build or configuration
Integration and access implementation
User trial and feedback cadence
Role-based training and working guides
Quality, adoption and cycle-time measurement
Monitoring, support and incident design

What you leave with

Operational AI service or workflow
Named owner and human control points
System record and test evidence
User guidance and handover
Benefits review
Next-wave recommendation

DECISION GATE

Adopt, extend carefully, maintain at the current boundary, or retire.

Fixed monthly service · Monthly cadence

Ongoing AI Partner

Keep working services healthy and the next decisions sensible.

For organisations that want senior AI continuity without immediately creating a full internal AI product and operations team.

What we do together

Monthly portfolio and backlog review
Service performance and exception review
Prompt, retrieval, workflow and evaluation improvements
Vendor and model change assessment
Short prototype cycles for new candidates
Coaching for internal owners and builders

What you leave with

Prioritised improvement backlog
Operating and management report
Updated tests and documentation
Risk and issue record
New decision-ready candidates
Growing internal capability

DECISION GATE

Continue, change the cadence, bring work in-house, or close the service cleanly.

Fixed price, properly bounded

You know what the phase costs before it starts.

Fixed price works when the job is clear. We do the work to make it clear: boundary, deliverables, inputs, assumptions, exclusions and the decision you need at the end.

No hourly meter

The commercial agreement is tied to the agreed outcome and phase, not the number of hours visible to you.

No automatic next phase

Each engagement is useful on its own. You can continue with Graymatter, take the evidence elsewhere or stop.

No hidden scope creep

If new evidence changes the job materially, we explain the choice before additional work or cost is committed.

No buried platform cost

Third-party licences, usage charges and specialist work are identified separately where they apply.

We do not publish a rate card because integrations, data access and risk can change the size of the same-looking workflow. After a short fit conversation, you receive one scope and one price for the recommended stage.

Before a price is quoted

The boundary is part of the product.

A fixed price is useful only when both sides can see what the stage must decide, what is included and what the client needs to provide.

Fit conversation

We clarify the workflow, reason it matters, sponsor, users, information, systems and consequence of error.

NO PITCH DECK REQUIRED

Smallest sensible stage

We recommend an Opportunity Map, direct prototype scope, a dependency to resolve first, or no engagement.

AI IS NOT ASSUMED

Written boundary

You see the deliverables, acceptance method, client inputs, assumptions, exclusions, third-party costs and end decision.

ONE SCOPE · ONE PRICE

Decision before work

You approve the stage and price before it is scheduled. The next phase is never bundled or presumed.

STOP / GO STAYS VISIBLE

What we need from you

Access to the work, not a perfect AI brief.

01A sponsor who can decide. Someone who can set the outcome, involve the right people and make the stop/go call.

02People who do the work. Their judgement, exceptions and workarounds are part of the product requirements.

03A safe way to inspect the workflow. Approved examples, representative data or a carefully constructed test set.

04Time for short feedback loops. Useful prototypes come from regular checks, not a large reveal at the end.

05Relevant control owners. IT, privacy, risk, security or legal involvement should match the use—not arrive after the build.

Independence matters

We will tell you when AI is not the next move.

A workflow may need a simpler process fix, better source information, cleaner permissions or a conventional automation first. A use case may also be too consequential, too poorly owned or too hard to test safely as an early pilot.

A clear “not yet” is a useful deliverable. It protects budget and trust for the opportunities that are ready.

A practical first conversation

Which uncertainty are you carrying?

If you already have a candidate, bring it. If you have too many, bring the list. We’ll help you identify the smallest sensible next decision.

Start with the workflow