AI enablement & governance · UK financial services

From AI pilots
to sustainable, governed value

I help regulated businesses turn AI licences into everyday use, with the governance and measurement to prove it's working.

Working with Microsoft 365 Copilot since launch, from both sides: at Microsoft, and now inside an FCA-regulated insurer.

1,500+colleagues enabled on Copilot Chat
150 → 400licensed users in under 6 months
90%+of licensed users active every month

Get Copilot used, not just licensed

Readiness-first rollouts and adoption measurement.

Govern AI agents without slowing teams down

A tiered model, self-serve routes and less shadow AI.

Prove AI is paying off

ROI tracking, cost showback and executive reporting.

Where I've worked
  • UK insurer now
  • Microsoft
  • Verizon
  • Accenture
  • CSC
  • IBM

Selected work

What I've delivered

Rolling out Copilot, proving its value, and governing AI agents.

Outcome

A readiness-first Copilot rollout

Colleagues learn what Copilot can do before they request a licence, so they know how to use it from day one.

300 → 1,000Copilot licences for one Microsoft customer, using the same approach
Senior stakeholders

Use-case discovery with business leaders

Discovery sessions with senior leaders that turn team problems into AI and automation solutions.

Governance & ComplianceOperational ExcellenceHR & PeopleCustomer OutcomesChange
Value & cost

AI value and cost transparency

  • Agent ROI tracker: savings in pounds, net of cost
  • Cost model for agent running costs
  • Cost-centre showback of GitHub Copilot usage to business budgets
Measurement

Adoption reporting for leaders

Monthly KPI reporting that separates one-off, repeat and habitual use, plus briefings on licensing value and Agent 365 readiness.

Enablement

Role-based enablement

  • Workshops for risk, finance, fraud, BAs, PMs and PA/EA teams
  • AI Foundations pathway with L&D
  • Clinics and Tech Bites sessions
Tier 1Personal agents
Tier 2Team agents, named owner
Tier 3Enterprise agents, exec sponsor

Tiered Governance Model for AI agents

A risk-based model, aligned to AI policies, that sets which agents teams can build themselves and which need a central solution. Oversight scales with the tier, giving teams a governed route to build and helping prevent shadow AI.

Capabilities

Get a return on your AI investment

AI governance & risk

Tiered, risk-based approaches to approving AI agents, with working knowledge of the UK regulatory context.

Tiered agent governanceAI Governance ForumFCA Consumer DutySM&CR

Copilot adoption & value

Encouraging regular use of Copilot and reporting on adoption to senior stakeholders.

M365 CopilotAdoption KPIsLicence strategyValue realisationExecutive reporting

Agentic AI

Matching use cases to the right Microsoft tool, making AI costs visible to the business areas that use them, and the governance tooling around agents.

Copilot StudioAgent BuilderAgent 365PurviewEntra IDCopilot CreditsGitHub CopilotFinOps showback

Change & enablement

Readiness-first rollouts, role-based training and communications, using structured change management.

Prosci ADKARRole-based curriculaInternal comms

Executive & customer leadership

Working with senior leaders and executives to turn business problems into AI and automation use cases, building on years of C-suite work on strategic accounts.

Use-case discoveryC-suite engagementSuccess planningQBRsCross-functional orchestration

Service & operational governance

Service transitions, SLAs, RACI and process design for global clients.

ITILPRINCE2Service transitionVendor management

Proof of work

Frameworks, agents and training I've built

Practical tools for governing AI and helping people use it well.

Tool

Agent ROI tracker

Measures the real return on personal productivity agents built in Microsoft 365 Copilot and Copilot Studio, use case by use case.

  1. BeforeRecords how the workflow is done today and the effort it takes.
  2. AfterCaptures effort, quality, time and actual usage once the agent is live.
  3. ReconcileConverts hours saved into pounds at fully loaded rates, adds strategic gains like faster decisions, better customer outcomes and accuracy, then nets off the agent's cost.

Reviewed quarterly with each agent's owner, as part of agent lifecycle governance.

Framework

Agent Lifecycle Governance

The controls to observe, manage and control the agent estate, so the business can scale agents safely. It also gives agent owners clear guidance on their accountability from idea to retirement.

Agent

Use-case intake agent

Colleagues submit AI use cases and see how complex the solution is, and whether they can self-serve. Where they need support, it raises a solution build request in Azure DevOps.

Training

Copilot training for insurance

My own Microsoft 365 Copilot training, built around the insurance industry and its real use cases.

Speaking & outreach

Making AI practical for every audience

Amita broke down big tech into practical, easy-to-use tools we can start using today. Her no-nonsense, approachable style made tech feel a lot less scary.
Lara HarrisonManaging Director, Superstar PAs Ltd, after Amita's guest expert session on AI productivity and prompting
  • Panellist: Neurodiversity and AIHow Copilot and AI tools can support neurodivergent colleagues, at a UK insurer
  • Guest expert sessionsAI productivity and prompting for PA and business support communities
  • Host and speakerCustomer Success UK events (600+ audience), International Women's Day, Women in Tech
  • OutreachAI and cyber sessions for young people through Cyber Girls First, 100+ pupils per event

Career

How I got here

I've spent my career in the gap between new technology and the people expected to use it. I started in 1998 as a network engineer in IBM's global network operations. At CSC and Accenture, I brought new processes and tools into large organisations and trained the teams who would run them. At Verizon, leading a team of 40+ and governing global client services, I learned that clear governance and strong relationships are what keep complex change working.

At Microsoft, I saw the same pattern with Copilot from launch: licences only create value when people know what to do with them. Now I'm applying all of it inside a regulated insurer, taking AI from pilots to adoption that's sustainable, governed and measurable.

  1. 2025 – now

    AI in a regulated business

    AI and Copilot enablement at an FCA-regulated UK insurer.

    Tiered agent governance · agent ROI tracker
  2. 2022 – 2025

    Driving adoption at scale

    Customer success for enterprise retail at Microsoft, working with Copilot from launch.

    50,000 seats in 6 months, half the usual time
  3. 2011 – 2022

    Governing services, leading people

    Global client governance at Verizon, then leading a team of 40+ in financial service management.

    £35m client portfolios · 7.94% credit reduction
  4. 1998 – 2011

    Making new technology land

    Network engineering at IBM Global Network Services (1998 – 2002), then service transitions at CSC and Accenture, bringing new processes and tools into large organisations.

    Client portfolios up to $1.9bn

Full career history on LinkedIn

Credentials

Qualifications and recognition

Certifications

  • Microsoft Azure AI Fundamentals
  • Microsoft Security, Compliance & Identity Fundamentals
  • Microsoft Azure Fundamentals
  • Prosci Change Management for Managers
  • ITIL IntermediateService Strategy · Service Transition
  • PRINCE2 Practitioner

Recognition

  • CSU AwardSocial impact
  • Leadership AwardVendor partner transition
  • Customer Excellence Award
  • Personal Commitment Award

Education & languages

  • BA (Hons) Information Technology in SocietyManchester Metropolitan University
  • LanguagesEnglish, Gujarati (fluent) · Hindi (basic)

Let's talk

Building AI capability in a regulated business?

I'm happy to talk about AI adoption, agent governance and roles in financial services.

Connect with me on LinkedIn