AI can build a pilot in minutes. Turning it into business value is the hard part

The speed at which organisations can now create things using AI is remarkable. Dashboards that previously took days or weeks to develop can potentially be created in hours, minutes or even seconds. Teams can test ideas, build proofs of concept and experiment with new ways of working much faster than before.

There has understandably been a rapid embrace of AI as a result. But more recently, there has also been a degree of pause and reflection as organisations consider what comes next. Creating something quickly is one thing. Getting it into the hands of users, making it sustainable and demonstrating the business value it delivers is a much bigger challenge.

That is where the focus now needs to be. Organisations have spent a lot of time experimenting with AI and understanding what technology can do. The next stage is about taking an enterprise view and connecting that experimentation with business value, governance, data, security and adoption.

Start with the business problem

AI makes it possible to generate lots of ideas very quickly. That does not necessarily mean they are all good ideas.

The business challenge should drive the use of AI. Organisations need to understand what problem they are trying to solve, how it aligns with their strategy and what value it is expected to create. They also need to define how that value will be measured and continue tracking the benefits, risks and adoption once the technology is being used.

In many ways, AI is not dissimilar to other forms of business change and transformation. The technology may be moving incredibly quickly, but organisations still need to understand the outcome they are trying to achieve and measure whether the change is delivering it.

There is also value in looking beyond individual tasks. Organisations should look at their processes and identify high-friction workflows where AI could make a bigger difference. That means considering the controls, handovers and different stages within a process and understanding how those could change.

Some elements may become automated while the human element becomes more important elsewhere. Looking at the whole process gives organisations a better opportunity to understand where AI can make a meaningful difference rather than simply finding individual tasks where the technology can be applied.

Treat AI as a strategic asset

AI needs to be treated as a strategic asset because of both the opportunity it creates and the impact it can have across an organisation. That starts with clear executive accountability and visibility of what is happening.

An AI governance board with executive accountability can help provide that oversight. Organisations should also have a clear view of their AI use cases, including the value, risk and ownership attached to them, alongside defined approaches to approvals, escalation and monitoring.

This becomes increasingly important because AI can move at its own pace, and that pace is very fast. Different teams can start experimenting with different tools and use cases, making it difficult to maintain a clear view of what is happening across the organisation.

Governance should therefore work alongside adoption. The aim is to retain the speed that AI provides while creating the trust, control and confidence needed to use it across the enterprise.

Data is a major part of that. Organisations do not need to get all of their data right before they start using AI, but the data being used for a particular application needs the appropriate quality controls and governance around it. Extra attention is needed where that data is connected to customers, financial information or operational activity.

Security should also be considered from the start. Cyber and security teams need to be part of the AI organisation, with the appropriate controls and operational resilience built into the approach.

Build AI literacy across the organisation

People are another important part of moving AI beyond experimentation.

Foundational training can give employees a basic understanding of AI, but practical use is also important. There is a difference between completing a course and understanding how tools such as generative AI can actually be used as part of day-to-day work.

Giving people opportunities to experiment in a safe and secure environment helps build that understanding. It also needs to sit alongside existing training around areas such as data, compliance and security so that people understand the impact their use of AI can have.

AI literacy will vary depending on somebody’s role. A developer will have different requirements from somebody using AI to improve a business process, while leadership teams need a different understanding again. Role-based AI learning can help address those differences.

AI champions can also play a useful role within priority functions. These are people who can build their own knowledge through practical use and pass that learning on to the wider organisation.

AI literacy will continue to develop through that combination of foundational knowledge, role-based learning and practical experimentation.

Move from pilots to measurable value

Creating more pilots and proofs of concept will not, on its own, move AI to enterprise scale.

The important step is connecting those pilots to the wider organisation. That means understanding usage and costs, putting the right cyber and governance around them and measuring the business impact they are creating.

Once a high-value use case is in place, it should also continue to develop. User feedback can help organisations iterate and improve what has been created rather than treating deployment as the end of the process.

This is also why taking an enterprise view of AI matters. Individual teams can create useful applications, but organisations need visibility of those use cases and an understanding of how they fit into the wider business.

AI is already capable of creating powerful things incredibly quickly. The focus now needs to be on identifying the right business use cases, measuring their value and putting the governance, data, security and skills around them that allow organisations to use AI effectively at enterprise level.

That is the shift from experimenting with AI to making it a real, governed and measurable part of the organisation.

‍