CFO Advisory #1: Unlock AI value in finance

Operating model redesign 

Start from value, then enable with AI through data, governance, and skills 

When you talk about AI, start with value, not technology. Identify where finance creates value by examining areas such as margin drivers, working capital levers, and the dynamics of customers and markets. From here, determine use cases where AI could meaningfully influence outcomes. For each use case, be clear about what the business must provide: ownership of the data, quality standards, system access, compliance needs, and the people and skills to support the solution. For material decisions, require three things: explainability (“why”), evidence/lineage (“based on what”), and human approval (“who signed off”). 

Data is what makes AI scalable. In finance, the biggest limiter is rarely the model but whether you have trusted numbers with clear definitions, traceable sources, and the right access controls. If “revenue,” “margin,” or “working capital” mean different things across teams, AI will amplify confusion faster than it creates insight. Treat key finance datasets as data products with an owner, quality rules, and documented meaning, so they can be reused across multiple use cases.

How to get value in weeks, not years 

Rapid pilots and reusable data foundation 

1. Build learning capability, not one-off solutions 

Start with your learning journey. AI is moving fast. The tools and models you choose today may be outdated within a year. The real long‑term advantage is not one clever use case but a team and an organisation that knows how to learn, adapt, and pivot quickly. Organisations that win will be the ones that can change direction in weeks instead of years. 

2. Anchor use cases in value and competitiveness 

Link this to your value drivers and the three modes of thought. You need to:

• Involve and educate people through personal productivity use cases
• Keep up with competitors by improving key processes
• Get ahead of competitors with selected big bets that transform how you work

3. Use rapid pilots to learn fast and reduce risk 

Fail fast and learn fast. Use rapid prototyping to explore what is possible. Your AI journey will not be a straight line. You will try things that do not work as expected. That is normal and helpful if you use it well. 

Extract the most value from each setback. Document what you learned, what the limiting factors were, and share these lessons and concrete stories across the organisation. When you do this, your return on failure goes up, and every experiment, even the failed ones, moves you forward.

4. Scale what works through shared data, governance, and product ownership 

The task at hand is to establish a clear path from prototype to production, so lessons from wins and failures accumulate. Once you know what works, build to scale and harden solutions with IT. AI solutions decay unless they are continuously managed. Value compounds only if learning is institutionalised – otherwise it stays with the pilot team. CFOs are increasingly expecting product-like ownership of digital capabilities in finance, so make AI a “flywheel”, not a project. Treat AI in finance as a product that you continuously improve: measure performance, prioritise the next improvements, release changes safely, and standardise what works so benefits compound across teams and processes. This is how you stay ahead while maintaining trust, governance, and control. 

And don’t forget the users. A good rule of thumb is that your investment in change management should be close to your investment in the technology itself. 

Adoption is the main event 

Digital solutions do not deliver their expected value unless people in your organisation trust and use them. This is true for any new technology. AI-enhanced forecasting models may be impressive, but they only help if people believe that, on average, the model performs better than they do on their own. 

To establish trust and relieve uncertainty, leaders need to listen to and nurture what is already growing in the organisation. The most transformative ideas often appear in unexpected places, so make it easy for people on the ground to suggest and test ideas. Where ambiguity or risk is material, it is important to keep humans in the loop. 

What sets AI apart is its speed: it is evolving so fast that your people must keep up with the change. This requires an organisation that is comfortable with exploring this uncertainty.