Fractional CDO / Interim Data & AI Leadership / PE and Board Adviser

You cannot AI your way out of data you don’t trust.

I build the data a business can trust, then put AI to work on it, with a named owner for every agent, value agreed before the work starts, and accountability measured to drive performance.

I’ve built the first version in a £10bn regulated business: the same foundations helped prevent more than £250M of fraud and took data agents into production. The next five years will be decided by who can trust AI enough to let it do real work, with people still accountable. For a PE sponsor, a founder or a CEO buying senior cover, a reset or a value-creation plan, that is the work I do in the first months, in the same order, with the value measured the same way.

Experience
Virgin Media O2, Salesforce and MuleSoft.
Value, signed off
circa £500M of value enabled, £23M+ of it signed off by Finance.
Public proof
Tableau, Google Cloud Next 2026, Anaplan Connect London 2026, Interface Magazine and British Data Awards 2026.
Now
Independent Executive Advisor, Data & AI. Delivery, value creation, executive and strategic advisory on data and AI for CEOs and executive committees.
Availability
Taking a small number of advisory engagements. My next role is a permanent Chief Data Officer, Chief Data & AI Officer or Chief AI Officer seat.

Three conversations

Which conversation are we having?

  • Fractional or interim leadership

    Senior cover, a reset, or CDO-level judgement before or during a permanent hire: the foundations, the owners and the value gate set in the first months.

  • Diagnostic or 90-day reset

    A board-ready read of the data estate, AI readiness and the value on the table, then a plan the team can run, in the order I run it: foundations first, agents last.

  • Data & AI value creation for PE and boards

    For sponsors and leadership teams where the brief is to mature data and AI and convert it into measurable commercial outcomes.

The record

The record

  • circa £500M

    of value enabled; £23M+ of it directly attributed and signed off by Finance. Everything we did had a number attached.

  • 13,000

    Gemini Enterprise licences approved for the following year, and Google asked me to speak on data agents at Cloud Next 2026.

  • £13M+

    a year of run cost out: 9 BI and data platforms consolidated into one governed stack, with nothing switched off until its replacement was live and accepted.

  • 11,000+

    colleagues self-serving on 250+ certified data products; 8,000+ trained by a Data & AI University I founded and ran.

  • Data Team of the Year

    British Data Awards 2026, won with a 300+ data and AI organisation built from one pillar to seven, each led by a director reporting to me.

01 / Engagement Shapes

Three ways to bring the method into your business.

Every engagement runs the same order: foundations, adoption, value, agents. What changes is how much of it you need and how fast: a diagnostic, a 90-day reset, interim cover, or an advisory rhythm beside the board.

  • 4 to 6 week diagnostic

    The data estate, who owns what, the platform footprint, AI readiness and the value on the table, read against the four stages.

    A board-ready view of what is working, what is blocked and what to do first.

  • 90-day reset or interim cover

    The first three months of the playbook: certified business layers with one named owner per domain, a value case before anything is built, and adoption measured with the business.

    An operating rhythm the team can run, a prioritised roadmap and the decisions that unblock delivery.

  • Board, PE and role-design advisory

    Decisions on the CDO or CDAO seat, AI readiness, platform simplification and value creation, made with someone who has sat in the seat and owned the data milestones at the board.

    A safer brief, better sponsorship and a clearer decision path.

Engagements are usually shaped around a clear problem, sponsor and time horizon: a 4 to 6 week diagnostic, a 90-day reset, interim cover, or a focused advisory rhythm. Exact structure depends on the scope and decision path.

02 / Case Proof

Proof the method pays off.

The pattern in my work is turning fragmented systems, unclear ownership and low-trust data into something the business uses and Finance signs off.

  1. Customer Trust Indicator, data turned into the top customer measure

    Problem: the business needed a real measure of how well it serves customers, beyond NPS and revenue.

    Intervention: I built the measurement behind the Customer Trust Indicator, the company’s top customer measure, with the Consumer Strategy team: hundreds of signals scored by AI into a trust score and an action.

    Result: it changed priorities across the business. The principle, measuring customers by their own outcome, came with me from Salesforce, where my team first built it on the business layers.

  2. From fragmented reporting to one governed stack

    Problem: fragmented reporting, weak ownership, manual workarounds and inconsistent definitions across 9 BI and data platforms.

    Intervention: business layers that model how each domain operates, one owner per definition, rule and data product, and nothing live without certification; the 9 platforms consolidated into one governed stack, with nothing switched off until its replacement was live and accepted.

    Result: £13M+ a year of run cost out; 11,000+ colleagues self-serving on 250+ certified data products; 4,500+ users and 17,000 contact-centre agents moved off legacy tools; £23M+ directly attributed and signed off by Finance.

  3. Gemini Enterprise and data agents

    Problem: business teams needed a trusted, governed way to put AI to work on the data.

    Intervention: stood up a shared agent platform with the technology team, security and controls built in, in a few months, working with Google’s engineering and product teams; agents ran only on certified business layers, with documentation, semantics and owners in place.

    Result: a new agent reached production in hours, with 30 more approved in testing; 10+ agents live, answering around 500 questions a day; all 3,000 available Gemini Enterprise licences deployed, and 13,000 approved for the following year.

If this sounds close to the problem you are trying to solve, it is worth a conversation.Discuss an engagement

The method

The playbook I wrote

I didn’t follow a transformation playbook. I wrote one in 2023, had ten senior colleagues review it, and rewrote it in 2025 from two years of running it. Four stages, in the order I run them, and the four shifts I set on arrival. The operating model ran on data mesh principles: each domain had an accountable senior leader, a process owner and a business data owner, partnered with a data owner from my team, and work was prioritised fortnightly under a service agreement co-designed with the business and a value gate.

  1. 01

    Foundations

    Certified business layers with one named owner per domain, and governance that covers ownership, quality, metadata and lifecycle.

  2. 02

    Adoption

    Self-service on data people trust, with the training, support and communities that keep it in use inside the business.

  3. 03

    Value

    Run as a value centre, not a cost centre. Nothing starts without a case the business has signed off, and realised value is tracked afterwards.

  4. 04

    Agents

    AI built only on certified foundations, governed, human in the loop, and owned by a named person in the business.

  • Service to partnership

    A signed partnership per business domain: what my team commits to in one column, what the business commits to in the other.

  • Central data to business domains

    A dedicated squad per domain builds that domain’s business layer before anything else.

  • Renters to owners

    One owner per definition, rule and data product, in the business; disagreements between domains go to senior leaders.

  • Centre of excellence to centre for enablement

    Nothing live without certification, and a Data & AI University so people came to us.

03 / Selected Public Evidence

Public sources you can check for your due diligence.

These links show the work from outside the CV: customer stories, interviews, speaking profiles and coverage from the data and cloud industry.

Video

The 3-minute story

The career in three minutes: the engineering, the enterprise data scale, AI in production, and the work I do now.

Watch the 3-minute story (opens in a new tab)

  • Engineering foundation

    Computer Systems Engineer, graduated top of the generation; hands-on work for over 20 years before the executive seats.

  • Enterprise transformation

    Operating-model change at Peñoles, MuleSoft, Salesforce and Virgin Media O2, so the business owns the outcome.

  • Data and AI leadership

    Trusted data first, then AI doing real work on it, with the value agreed before the work starts and measured after.

04 / How I Work

Clear scope, senior access and enough time to make it practical.

This page is for fractional, interim and advisory conversations where the problem is important, the sponsorship is senior, and the work needs to become practical quickly. I take a board, an executive committee, the business and the engineers to the same answer.

Best-fit situations

  • PE portfolio companies that need data and AI maturity ahead of exit or after an acquisition
  • Founders and scale-ups building a data operating model before a permanent CDO hire
  • Mid-market companies with fragmented data and weak ownership
  • Companies that need interim CDO cover during a permanent hire
  • Boards and executive committees that want clear data and AI advice
  • Teams that want measurable value from data, with the number signed off

Good-fit engagements

  • Fractional or interim CDO or CDAO role
  • Data and AI strategy reset or diagnostic
  • Operating model on data mesh principles, designed and adopted
  • AI adoption and data-agent readiness, with an owner behind every agent
  • Governance, trust and the data-product operating model
  • PE value creation or pre-exit data maturity acceleration

Working rhythm

Typically 1 to 3 days per week per client. UK-based. Remote-first as the default, with travel for board sessions, key workshops or leadership alignment where it adds value. A diagnostic or strategy sprint usually takes 4 to 6 weeks: clear assessment, board-ready readout and prioritised roadmap.

“Most importantly, Mauro gets things done. He moves the needle, challenges the status quo and doesn’t settle for the way things have always been done.”

Tom Brown, Founder & Managing Director, The Information Lab, on LinkedIn.

I do my best work where data and AI leadership comes with real scope: a business that wants judgement, a better operating model, adoption the business measures and accountable progress. If you are not sure whether your situation fits, share it anyway. I would rather hear it and give you an honest view than have you hold it back. And if it is not right for me, there is a good chance I know someone strong who it is right for.

How it works

Four steps,
no theatre.

The process is deliberately short. Most of the value in the first month comes from somebody senior asking the questions nobody internal is positioned to ask.

  1. 01

    Discovery call

    Thirty minutes. You describe the position, I ask the awkward questions, and we both decide whether there is something here worth doing.

  2. 02

    Proposal within 48 hours

    One page: what I would do, in what order, for what price, and what you should expect to see by when. No deck, no discovery phase to sell you a discovery phase.

  3. 03

    The work

    Fixed scope with a named deliverable. You get an executive who has carried the accountability, not a slide-writer with a framework.

  4. 04

    The next decision

    Every engagement ends with a clear recommendation on whether to continue, hand over, or stop. Stopping is a legitimate answer, and I will say so.

Questions

Before you
book the call.

Why fractional rather than hiring a full-time chief data officer?

A permanent hire at that level is a six-figure commitment plus equity, and it usually takes months to land the right person. A fractional arrangement gives you the judgement now, costs a fraction of that, and can stop or scale when your needs change. If what you actually need is a full-time CDO, I will tell you, and I will help you write the brief.

Can an interim turn into the permanent appointment?

Sometimes, and I will say so early rather than late if I think it should. What I will not do is take an interim brief, quietly angle for the permanent job and leave you without a shortlist. Part of the interim work is writing the brief for whoever comes next, and that stays true whether or not the answer turns out to be me.

What makes this different from a consultancy?

A consultancy sells you a team. I sell you one accountable executive who has actually run the thing at scale, built the operating model, and rewritten it once it met reality. You get the judgement, not the pyramid.

Do you do the delivery as well?

I own direction, sequencing and the operating model, and I coach your leads to deliver it.

What can you not help with?

Anything that needs a large delivery team on the ground. If it is not a fit I will tell you on the call rather than take the engagement and hope.

How quickly can you start?

A diagnostic usually starts within two to three weeks of the call. Advisory can start sooner because the commitment is lighter.

Start a conversation.

A short note is enough: the business or the portfolio company, where the data or AI work is stuck, and what would need to be true in 90 days for it to have been worth it. I read every message personally and reply within 24 hours.