What I do
Problems do not arrive sorted by category. An AI programme stalls on undocumented process; a platform decision turns out to be a question about who owns the data.
Capability maps, architecture layers and BPMN process descriptions: clear, consistent and maintainable.
More Data and information architectureOne unified and agreed picture of the organisation's core information.
More AI and agent readinessAn honest assessment of where you can safely apply AI agents, and what has to be in place first.
More Solution architecture and assuranceThe right platform decision, and an independent read on a design or a proposal before it becomes irreversible.
MoreAn ongoing arrangement
Fractional chief architect
Some engagements end in a decision. Others need someone who holds the context between them. A defined monthly commitment, participation in forums where decisions are made, and mentoring for your own people. The intent is to become unnecessary.
MoreWho this is for
Transformation, system selection, and processes that need pinning down before the next system is bought.
Independent advice that outlives the procurement, capability-based specifications, and architecture the next team can inherit.
Product architecture, roadmap and ALM from an architect who has carried all of it himself.
Why the order of work matters
Transformation programmes rarely run aground because the wrong platform was chosen. The root cause is typically that nobody established what the organisation actually does, how it does it and what information it depends on, before the platform was selected. Architecture is then reverse engineered from the implementation, and every decision after that inherits that error.
That was survivable while the only consumers were people, because people compensate; they know the diagram is two years old and they adjust. AI agents do not compensate. They work with the vocabulary they are given and resolve inconsistency by guessing, which can quickly become a problem.
The seemingly optional work of pinning down processes and agreeing what a term means was always load-bearing. AI removed the tolerance for pretending otherwise.
A big decision ahead, or a programme that is not going to plan?
Does any of this sound familiar? An ERP or CRM decision to be made where the analysis so far has come from the bidders. An implementation that is not delivering what was promised. You are about to start something significant and want the foundation right from the beginning. Data in three systems that will not reconcile. A board asking what you are doing about AI.
Our first conversation costs nothing and usually helps clarify the question.
Get in touch