What’s next for Enterprise AI?

Interest in the enterprise applications of AI, especially generative AI, really took off in 2023. Enterprises engaged with the potential for AI technologies to transform their operations, supply-chain, communications, customer proposition and decision making. They invested in developing a deeper understanding of potential sources of value and possible applications of AI technologies. Many got excited by the potential to deploy thinking machines capable of undertaking highly complex, business critical, functions and decision making. So where next for enterprise AI in 2024??

Moving AI from ideate and discover to design and implement will be a multi-year journey for most enterprises. The complexity and scale of the technology enablers and business change required are, for most, significant. Like cloud computing before it, implementing AI is likely to deliver ‘game-changing’ capabilities and value, but will require multiple years of complex delivery and change management. For many enterprises, 2024 is year one of this journey.?

We might think of 2024 as the year we lay the foundations for our AI enabled transformation journey. There’s no shortage of thought leadership offering advice around where to start with AI and how to build an AI strategy, business case, delivery model etc. Project One adopts a slightly different approach to the question of enterprise AI. We start from fundamental, pragmatic business change principles and work from there: Is there a compelling business case? Can we shape a credible implementation plan? How might AI impact our operating model? How will we manage the change and bring our people with us on the AI journey? Are we thinking about risk in the right way??

From this perspective, here are some of the questions enterprises might ask in year one of the AI transformation journey:

What problem(s) can AI help us solve??

As with all technologies reaching the pinnacle of the hype cycle, AI is often a solution in search of a problem. Understanding AI’s potential is a useful exercise, but investing in AI on this basis is unlikely to yield great results. The ‘build it and they will come’ approach is fraught with risk. Properly designed and implemented, AI will be able to solve problems that until now have been beyond the remit of a technology solution, ones that involve the execution of complex, creative tasks requiring human-like ingenuity, critical thinking and judgement. But these problems need to be carefully scoped and defined to enable the solution to be properly designed and implemented to solve them.? Adopting a systems-thinking approach could prove useful, reflecting on how AI sits within the broader organisational, cultural, operational, commercial and regulatory landscape, rather than approaching AI as an ‘IT’ problem statement.?

How should we plan to implement AI??

Enterprise AI is evolving rapidly. Looking ahead, the only thing we can be sure of is that we don’t know exactly where we will get to, what kind of AI products will be available and what use cases they will be able to meet. At the same time, it’s clear that if we don’t start laying the foundations for AI now we will fall behind quickly. Faced with so much ambiguity, where to start? Taking your medium-term enterprise strategy as a starting point is a good option. Aligning our thinking around when, where, and how to implement AI with strategic priorities will help shape the kind of AI use cases worth exploring, structure our AI roadmap and articulate our AI north star. At the same time, agility needs to be engineered into the implementation plan from day one. With the technology evolving so quickly, experimentation, flexibility, speed and innovation are critical. In this regard, the buy vs build question is particularly important. Planning to build AI technologies isn’t feasible for most enterprises so, by default, these services will be purchased. However, much AI will probably rely on access to large amounts of easy to access and process data. Getting data ready for AI is a critical AI implementation enabler than can begin today and is likely to involve a range of buy and build solutions.?

What impact might AI have on our operating model??

The automation of a swathe of complex, creative functions currently undertaken exclusively by people will be highly disruptive for most enterprises. We should assume that realising the benefits of enterprise AI will involve fundamental changes to operating models. These changes could be incrementally phased over an extended period to reduce disruption or executed rapidly to maximise benefit. Either way, thinking about this dynamic now is a worthwhile exercise as it raises profound questions related to workforce planning and management, corporate culture, organisational design and enterprise operations that will take time to work through. One of the key dynamics to reflect on is how to retain specialist enterprise specific skills and knowledge to avoid a risky over-reliance on decision making AI platforms that can’t explain why they have done what they have done and can be challenging to ‘re-teach’. To put it another way, how can we achieve a hybrid people-AI operating model so specialist skills and experience can inform, augment, check and challenge automated workflows and decision making??

How will we manage the level of business change driven by AI??

Business change management will be an important part of the AI transformation story given how intense and disruptive AI driven change is likely to be for most. Upskilling, workforce management, process re-engineering, organisation design, culture and leadership are likely to be important themes. Alongside these more traditional business change topics, thinking about how to address many people’s perception that AI is a threat or a negative, to be mistrusted or even feared, might be useful at this early stage in our AI journey. We can’t know yet, but it may well be that AI implementation generates more resistance than any technology-enabled change before necessitating a phased, gradualist approach to mitigate the fear factor.?

What risks might AI create??

In the medium term, AI is likely to create a range of new security, workforce, reputation, brand, legal and regulatory risks. Securing the huge amounts of personal and proprietary data AI is likely to require, addressing biases that emerge as AI learns, managing workforce displacement and skill obsolesce, ensuring compliance with complex regulations that aren’t in place yet, assuring the stability and performance of evermore complex, highly integrated IT systems. These are just a few of the emerging risk dynamics around AI that enterprises are starting to think through. Progressing this thinking and starting to build the mechanisms to manage and mitigate these new risks is likely to be an important theme in 2024.?

2024 probably isn’t the critical inflection point for enterprise AI, the market for enterprise AI products likely needs a few years to mature before ubiquitous adoption is practicable. However, it may well be year one of a multi-year AI driven transformation journey, where enterprises lay the foundations that will enable them to fully exploit AI technologies at pace, once the services and products are ready.