How mature is your architecture?

A five-to-seven-minute, vendor-neutral self-check. Two questions about what your business actually needs, then eight dimensions spanning data, decisioning and the operating model around them. You are scored against your own need, not against a vendor roadmap.

First, what does your business actually need?

Two of the eight dimensions below describe a design choice rather than a quality. These two answers set the level you are scored against on those, so an organisation that has deliberately chosen scheduled, human-approved operations is not marked down for it.

? Decision speed How quickly does a customer decision actually have to be made in your business? Not how fast you would like to be. How fast the customer, the product and the moment genuinely require.
? Decision autonomy How much should the system be allowed to decide without a person? What your policy, your regulator and your risk appetite actually permit, rather than the ambition.
1 Data model & profile How unified and governed is your customer profile?
2 Identity resolution How well do you resolve people across devices and channels?
3 Real-time decisioning How real-time is your decisioning?
4 Channel orchestration How coordinated is engagement across channels?
5 Governance & consent How well is consent and governance built into the stack?
6 Measurement & feedback How well do outcomes feed back into decisioning?
7 Agentic readiness How ready is your stack for AI-driven, autonomous engagement?
8 Operating model & enablement How well do people, ownership, and ways of working support the stack?
How this is scored, and what it does not tell you

Eight dimensions, each rated one to four, for a maximum of 32. The bands sit at 27 for Optimised, 21 for Established, 14 for Developing and 8 for Foundational.

Three of the eight are treated as prerequisites: the data model and profile, identity resolution, and governance and consent. Rating any of them at level one caps the verdict at Developing, and rating any at level two caps it at Established, however strong the rest of the answers are. Strength elsewhere does not compensate for a foundation that is not there, and a total that averaged one away would tell you the opposite of what your own stack is about to teach you.

Six of the dimensions are qualities, where level four is better for everyone: nobody deliberately chooses ad hoc consent or siloed ownership. Two of them, real-time decisioning and agentic readiness, describe a design choice instead. Continuous in-session decisioning is a genuine requirement for one business and an expensive irrelevance for another, and the same is true of letting a system act without a person.

So those two are scored against the target you set in the two questions at the top rather than against level four. Meeting your stated need scores full marks, and only the gap counts against you. An insurer that has deliberately chosen scheduled, human-approved operations can therefore reach the top band on the strength of its foundations, which is the honest answer and the one the earlier version of this tool refused to give.

The dashed outline on the profile chart shows the level you said you need. The reading path is ranked on the gap to that target, not on the raw level, so it will not send you to fix something you have deliberately decided not to do.

This is a self-reported discussion aid built to start an internal conversation, not a validated instrument and not an assessment of operational readiness. Two people in the same organisation will often answer it differently, and that disagreement is usually the most useful output.