How to make a success of Gen AI in 2025

Plan on undertaking 10 AI projects, with at least one going into production by the end of 2025. This accounts for the many challenges in getting to a highly accurate, production-ready solution, as well as spreading investment risk across various areas of the business. 

More than 90% of CIOs said that managing cost limits their ability to get value from AI for their enterprise, according to a Gartner survey of over 300 CIOs in June and July 2024. In fact, Gartner believes that cost is as big an AI risk as security or hallucinations.

Whilst getting started with Gen AI is relatively easy, pushing projects to a level where they deliver real business value and reaching a production-level of accuracy and reliability, requires significantly more effort. Budgeting for example £500k or less per project allows sufficient investment to experiment, iterate, and develop a strong proof-of-concept before selecting viable solutions for deployment into production.  

The world we live in is full of constant information and disinformation, creating misconceptions about how to implement Gen AI. To navigate effectively, securing ‘air cover’ from key stakeholders is critical. Having backing from the board, and C-level leadership, provides the necessary alignment and support to push AI initiatives forward, ensuring the projects do not get stalled by organisational inertia or misconceptions. Gen AI must be incorporated throughout the organisation and not remain isolated to niche teams or departments.  

AI innovation can come from unexpected places, so creating an environment where experimentation is encouraged is key. Why pursue this approach? It really is simple: AI’s potential lies in delivering increased value to the customer. But achieving this goal requires overcoming the silos within the organisation and driving transformation at the operational level, not just automating current processes. Gen AI should be viewed as a tool to unlock new possibilities, not just a method to make existing tasks slightly more efficient. One way to accomplish this is to allow any one person or team to submit AI innovation ideas and business cases to the AI funding committee in the organisation, for them to select the 10 to be funded in 2025. 

The real power of Gen AI comes when it is integrated into a broader organisational transformation. AI is not just about automating what we already do; it’s about rethinking how we operate. To leverage Gen AI effectively, we must re-evaluate and potentially overhaul our operating models. A pragmatic and balanced approach is necessary, viewing AI as a catalyst for transformation rather than a quick-fix solution. Going back to the idea of submitting a business case for funding, these should include new ways to look at operational problems.  

Research firm International Data Corporation (IDC)  projects worldwide spending on technology to support AI strategies will reach $337 billion in 2025 — and more than double to $749 billion by 2028.  As businesses continue to grow with AI, the conversation is changing from hype to practicality. The initial excitement has been tempered by the reality that ROI can be very elusive. Headlines have moved from “AI is a game-changer” to “Where’s the ROI?”. Increased budgets for AI and data initiatives,  come the expectation of clear outcomes and measurable returns.  

We must set well-defined success criteria from the outset. Is 60% accuracy acceptable for certain applications? What are the security concerns? How much risk is the organisation willing to tolerate? These are questions that must be addressed early on to avoid wasted efforts and missed opportunities. As for our AI project business case, it should include clear targets to measure success. 

Underlying every successful AI initiative is a solid foundation of data. AI is only as good as the data on which it is built, and the ability to differentiate your AI capabilities often comes down to how well you manage, collect and utilise data. Without this foundation, even the best AI algorithms will struggle to perform. As we innovate our operational processes, look at the underlying data and how it can be used to train AI models that re-envision the process. 

As we continue to explore Gen AI, it is important to keep Amara’s Law in mind: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” This certainly applies to AI, where the short-term buzz often overshadows the slower, yet more profound, long-term impact. As Gen AI matures, the real value will emerge over time and not in immediate and flashy results. Look for incremental improvements in operational process that can be achieved quickly and moved to production. 

In summary, Gen AI presents an incredible opportunity, but requires a structured, realistic and long-term approach to realise success. We must focus on setting clear goals, building a strong data foundation, and ensuring the organisation is aligned in order to successfully integrate AI into our business. By transforming our operating models and embracing a balanced approach, we can unlock the full potential of AI whilst avoiding the pitfalls of hype-driven disillusionment.