Change management really started becoming a “thing” in the late 80s and early 90s, and within two decades, businesses were diligently factoring change management into their systems and processes and experts were charging eye-watering rates to assess the success of organizational transformation. Despite the hype around it, change management remains significantly underfunded in strategic budgets, and few businesses preempt the need for effective and structured change management within their plans and growth targets. It is therefore hardly surprising that inadequate change management has become one of the most significant causes of AI programme failure or poor AI return on investment in the organization. The models themselves aren’t underperforming necessarily, rather the challenge is that people are struggling to make full use of their potential due to outdated workflows, missing feedback loops, and lack of visibility.
AI programmes tend to stumble at the point where investment needs to translate into changed working behaviors such as skills development, redesigned roles, manager capability, employee trust, and executive follow-through. The Deloitte 2026 study into more than 3,200 companies found that the AI skills gap was the single biggest barrier to integrating AI into existing workflows. As discussed in a previous article, the gap between tech potential and tech deliverable is bridged by people.
Key takeaways:
· Gartner believes that by 2027 a significant percentage of companies will lose their top AI talent because they don’t have a clear AI people strategy.
· Only 20% of leaders believe their workforce is AI-ready.
· South Africa leads Africa on the AI readiness scale, but only scores 63/100 on workforce AI literacy and 53/100 on its AI literacy pipeline.
Why do most AI pilots never reach production?
A significant number of AI pilots stall when the tools hit real workflows. Solutions that work cleanly in demo environments often lose momentum when they come up against inconsistent data, undefined ownership, or teams who were not part of the decision-making process. The technology isn’t the constraint – the real limitation comes down to the fact that adoption wasn’t designed as part of delivery, it was bolted on at the end.
Incidentally, the same Deloitte survey showed that only 25% of companies had moved 40% or more of their AI pilots into production even though they were expanding their experimentation and investment.
What does AI readiness look like?
Genuine AI readiness isn’t a score card; it’s how people behave towards (and with) technology. Mint has adopted a customized AI Readiness checklist that places people and adoption right up at the top of the AI integration process alongside strategy, data management, and delivery. It is a structured evaluation that ensures teams understand how AI supports their work and provides training that focuses on interpretation and use cases.
To improve adoption, change management must be planned as part of the company’s AI readiness rollout strategy and the feedback needs to be continuously incorporated to improve how the solutions merge with workflows and real-world use cases.
How does role-based change management work?
Change management starts with the people doing the job and not the tool the company wants to deploy. Mint’s approach involves the business user early in the process so that they understand what the role of AI is, why it is relevant to the process in the first place, and how it can be used to support their role. It’s key to move the implementation of the technology from something that feels like an imposed system to something that will simplify or enhance the working environment and user experience. Training needs to be practical and relevant so that behavioral changes are lasting and people consistently and continuously engage with the technology.
How does Mint’s Listen phase close the change management gap?
This is precisely why Mint’s AI Enablement Framework starts with Listen rather than immediately building the AI systems. This ensures that we establish strategic priorities, identify where teams are struggling and with what, and what people define as a successful outcome. Is it working for them and with them or is it adding more steps to their workflows? Is it reducing time or making the process unnecessarily longer? This phase surfaces the questions so that problems are identified early and addressed before the point of no return during implementation which is a costly place to start fixing..
Companies who are serious about closing the AI change management gap aren’t spending more on models, but rather they are investing in the human experience that makes AI usable in a more deliberate way by using role-based capability building, early employee involvement, and accountable leadership from the outset.
Speak to Mint about starting your AI programme with the Listen phase, before you spend on tools.