Start a conversation

Bring us one workflow.

It might be repeated, slow, inconsistent, expensive or difficult to scale. It might already have three different AI experiments around it. Tell us what happens today and why it matters.

Direct contact

Phil Gray

Founder, Graymatter
Sydney · Australia-wide delivery

phil@graymatter.team (02) 5761 0386

Email is the simplest way to start. The local website does not submit or store form data.

Five useful details

You do not need a polished brief.

  1. 01

    What is the work?

    Describe what starts it, the main steps and the outcome.

  2. 02

    Who does it now?

    The team involved, the judgement they add and where hand-offs occur.

  3. 03

    What makes it worth improving?

    Time, capacity, quality, customer experience, growth, cost or risk.

  4. 04

    What information and systems are involved?

    A high-level description is enough. Do not email sensitive records.

  5. 05

    What have you tried?

    Include current tools, experiments, internal opinions and known constraints.

Open a ready-to-use email

A good first fit

You do not need to be “AI ready”.

You need enough access and ownership to test one real piece of work.

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

02The workflow matters and repeats. There is visible friction, delay, inconsistency, lost capacity or risk.

03The people who do it can take part. Their workarounds and judgement are part of the requirements.

04We can inspect safe examples. Approved records, representative data or a carefully constructed test set are available.

05A result can be judged in weeks. We can agree what better, acceptable and unsafe look like before building.

What happens next

A useful first call, not an AI pitch.

We will clarify the workflow, desired outcome, people, information and constraints. If it is a plausible fit, we recommend the smallest next step.

For a paid stage, you receive a written boundary, deliverables, timing and one fixed price before work starts. If AI is not the sensible next move, we say so.