Approach. From clarity to useful change.

Ask what needs to change.

Define the outcome and the decision someone must make. I use interviews, workshops and process discovery to understand where people wait, repeat work or lack reliable information. Then I agree the scope, owner and measure before proposing a tool.

In requisition fulfilment, I mapped hand-offs and approval status before exploring prediction as a separate experiment. A business case then defined target benefits; I keep proposed improvements distinct from measures verified after delivery.

See the whole flow.

Map current and intended journeys, roles, rules, exceptions, data sources and permissions. A shared picture lets operational colleagues, architects and delivery teams challenge the same assumptions and see where service design, integration or guidance might help.

I established an eight-gate service catalogue and a process-modelling approach from high-level value chains to working procedures. For learner engagement, service management, assets and workplace operations, I then tailored data and delivery questions to each project rather than imposing one generic template.

Make the source and its limits clear.

Check definitions, ownership, quality, missing values and comparability before modelling. I ask where data originates, who can use it, what must be tested in migration and how a decision will be reviewed. Real restricted data can support a bounded analysis; curated synthetic data can rehearse an approach; a research blueprint can establish design questions. I label these separately.

For three annual feedback cycles, harmonising questions came before reading a trend. For a synthetic purchase-order demonstrator, model scores showed behaviour in constructed examples and required a separate plan for real-world validation.

Build trust into delivery.

Set a human decision point beside a signal: who reviews it, what further context they need, when they can override it and how errors are reported. I connect acceptance criteria, privacy, security, explainability, support and feedback to pilot-readiness decisions.

For meeting agents, I helped turn roles, platform constraints, test identities, private-calendar behaviour and no-permission fallbacks into user guidance and pilot-readiness questions. Controlled technical checks were milestones, while wider access and support remained separate gates.

Grow the capability around the change.

Guidance, coaching, communities of practice and role-relevant learning help teams use evidence and AI with confidence. I led the BA Capability, a team of 10 analysts, and designed a six-course AI and data pathway with applied discussions and proposed protected learning time. The team pathway is distinct from my own MSc, and its design is not a claim that every colleague completed it.

After delivery, the question remains: did people adopt the change, did the intended measure move, and what should be revised?

Give a team a service offer and room to judge.

My consultancy model defined what analysts could offer, when to engage them and how to make their advice traceable. I frame value as equivalent capacity and cost, time saved in a workflow, or methods that can be reused across projects. Each needs an agreed baseline and follow-up measure before a target becomes an outcome. I produced a catalogue of more than 100 activities, a project-first process standard and persona governance guidance, then developed a Compass board prototype for recording work and value. Some components were working methods; others were documented proposals for adoption.

A later alignment business case recommended connecting project-level models to a wider process catalogue and a readiness check for automation. That was a recommendation with a proposed pilot, not an institution-wide implementation claim. See the BA Capability story ↗

Join the analysis to an owned decision.

Established data teams bring source knowledge, definitions and trusted reporting. I contribute problem framing, Python-based investigation, model evaluation and an explanation that a service owner can act on. We agree the hand-offs: who validates the data, who challenges the analysis, and who owns the intervention and its follow-up.

This collaboration model grew through my MSc learning and workplace reflection. It connects data ownership, decision intelligence and adoption measures beyond accuracy.

Analysis, leadership and applied AI.

My experience spans more than 15 years in business analysis and transformation. During my 2022–2026 tenure at Queen Mary University of London (QMUL), I led a team of 10 analysts and supported a £16m annual digital transformation portfolio, with projects valued from £50k to £2.5m. I completed an MSc in AI & Data Science at Northeastern University London in April 2026. I also hold an MSc in Government, Policy & Politics, a BSc in Computing and professional qualifications in business analysis, product ownership, service management and project delivery.