How we work
“What is it like to work with you?”
We are management consultants with deep domain knowledge. Where it adds value, we work with our own AI platform. We like to show rather than talk: you see something tangible before you decide on the next step.
From your material to something you can see
We like to show rather than talk. Where it adds value, we build further, with your people and, if needed, a partner.
Your material
You share your question and the documents of one bounded process: for example the functional design, the work instruction and the policy behind it. No personal data is needed. We are happy to sign an NDA, and if sharing is difficult, we can start from empty templates or tables of contents.
Your decision What you share; an NDA comes first.
We build and present it back
Soon afterwards, we show you something tangible: a first model, a prototype or a point of view, built on your own documentation.
Your decision Is this worth taking further?
You decide
Continue, adjust or stop. Many engagements end with a point of view or a prototype; that is a result in itself.
Your decision The next step, or none.
How do you make sure it is right?
The people who do the work know the domain. Every step passes five phases, each closed by an approval gate, and the person who does the work never signs it off.
Research We read your material and talk to the people who do the work.
Every phase closes with an approval gate: your people approve before the work moves on.
Choose a phase to see what it is about.
Research
We read your material and talk to the people who do the work. Documents, exports and interviews go into the model from day one; nothing is assumed that can be checked.
Who decides Our consultants with your experts; AI reads and drafts.
Approval gate The question is framed and the facts behind it are marked: verified, assumed or still open.
When a gate is not passed, the work goes back to where the problem lies. Every claim is marked verified, assumed or unverified, and partners work inside the same gates.
What does it ask of your people?
A process owner and the experts who know the work. We record the sessions, so nobody has to explain things twice, and our platform does the reading and drafting. Your experts check and approve what we capture, item by item, so the result is theirs, not ours.
Is my data safe with you?
We work with confidential knowledge about how your organisation runs, and we treat it that way. Select a card to read more.
What you take away
Every engagement leaves a result you can act on: a decision, a plan, and where it helps, an application your people use. Examples for each service:
Strategy & Perspectives
Strategic and business value, side by side
Breadth across the operation or depth in one process: the same grid, three options, and what each brings.
Design & Improvement Potential
Current versus target state
Per service where it sits today and where it belongs, with the benefit case alongside.
Set-up & Build
Your operating model as an application
Vita, the first application on Mira: one activity through seven layers, every layer one click away, used and corrected by your people.
Supplier Selection & Contracting
The model as the scope
The service catalogue, with an owner and service level for each service: the scope of the tender, taken from the model.
Transition & Transformation
Tested in a transition
Knowledge captured and approved while the work moves, so the receiving team starts from a validated model.
Governance & Optimisation
Maturity and a growth path
Twelve dimensions in four areas, from Lagging to Visionary: where you stand and the path to your target.
Questions we are asked
Do we always end up with a platform?
No. We are management consultants first. Often the result is a point of view, a plan or a prototype; we build further only where it adds value for you.
What do we need to share?
Only the documents of one bounded process, such as a functional design, a work instruction and the related policy. No personal or participant data. An NDA comes first; templates or tables of contents work too.
Do you work with partners?
Where a specific capability requires it, yes: inside the same plan, approval gates and agreements on data, and we stay responsible for the result. Every partner that touches your data does so under a signed data processing agreement.
Can we not do this with a good prompt and our own LLM?
An assistant on a document store gives answers that sound right, but you cannot see which source or control it used, or when it guesses. We first build one governed model in which every item is traced to its source and checked. Mira then answers from that approved set, always with a reference.
Where is it AI, and where is it rules?
AI reads the sources, drafts, and answers questions on the approved set. Storing, versioning and checking follow fixed rules, and people do the approving. A control is always a fixed rule, never a judgement by the AI.
Where is the human in the loop?
Everywhere. Your experts approve every item before it counts. On the route to automation, the steps an engine may take are chosen one by one, with controls around them; approvals and exceptions stay with people.
Who owns the model?
You do. What is ours is the method and the platform we build it with.
How long before we see something?
Soon after you share your material, you see something built on it. A typical assessment takes five to six weeks.
Start with your own material
Share your question and the documents of one process; no personal data needed, and an NDA comes first. We come back with something you can see.
