How AI transformations succeed (quietly)

1. Culture is the key 

We’ve long believed that successful AI transformation is just as cultural as it is technical. If you don’t bring your people with you, if you don’t change how decisions are made and tasks are done, the tech won’t stick. That’s why our transformation programs centre around people and process from day one.

1. The real problem isn’t the technology.

AI technology is still in its “frontier period”, but it turns out that the key factors in both success and failure were much more human. 

MIT’s report, based on 300 public implementations and 200+ interviews, calls out a massive gap between adoption and impact. The *GenAI Divide* is real: significant investments, tiny returns.

MIT didn’t point fingers at bad models. It highlighted the failure of coordination, design, and implementation….. and leadership. 

That’s consistent with what we see in AI transformation, and its not-so-distant cousin, digital transformation. 

1. A consistent theme of transformation 

The 95% failure rate for AI pilots sounds awfully familiar: it’s an echo of the infamous 65% failure rate (or variations of that number, most notably by BCG) for digital transformations. Both are rooted in behaviour and governance, not software. Transformation is not a product drop. 

It’s an operational and cultural shift. Without embedding into workflows and involving the people who do the work, AI will remain stagecraft.

The ingredients that make up an ambitious AI project failure repeat the same mistakes: scope without alignment, front-office glamour, and no behavioural change. 

1. Where the value lies: the back office  

One of the findings that resonated was the finding that the back office is where the successful AI transformations often lie. Budgets frequently gravitate toward sales and marketing, but ROI is commonly realised in the more mundane aspects of the business. 

The report found:

“Most companies poured their AI budgets into sales and marketing… but the real money‑savers were the boring tasks nobody talks about: automating paperwork, processing invoices, and handling routine administrative work”.

Another lens says it even more starkly: while more than half of AI budgets go to visible, front‑office use cases, the highest returns come from back‑office automation, such as procurement, legal, HR, IT, and compliance workflows.

1. Low visibility, high ROI 

This is why the quiet achiever consistently outpaces the flashy bragger. AI transformation is effective when it’s embedded in back-office processes, which often involve structured, rule-based data, clear outcomes, and lower visibility. That’s where efficiency compounds.

At JOURNEY, we’ve seen this clearly across different industries like Lemmi in finance and MedAire in aviation. The focus has been on embedding intelligence into daily workflows and compliance, not just flashy front-end dashboards. They’re the systems that quietly keep operations running and elevate the customer experience. 

1. Solve & Evolve™: a proven enterprise framework. 

We’re proud to have partnered with customers and been part of successful transformations. But that consistency is not by accident. 

Like anything worthwhile, these successes have happened through unglamorous stuff: working together, strong structure, bold thinking, and practical small steps. 

This is precisely why we developed our Solve & Evolve™ methodology.

• Solve: untangles real-world blockers—cultural, process, design—to get pilots embedded and operational.

• Evolve™: anchors tech tactics to lasting strategic vision, ensuring each implementation integrates and adapts, rather than making noise.

It’s not glamorous, but it works. And when 95% of AI pilots fail to deliver ROI, it’s this sort of process that is necessary to be one of the successful few. 

1. Partnership and trust matters

The MIT report showed that implementations with external partners succeed approximately twice as often as internal builds (around 67% vs 33% success rate) because vendors tend to align tools with actual workflows. 

When asked what made a good partner, the results were consistent, and again, had much less to do with technology:

1. Headlines fade, results don’t.   

Paul Roetzer from the Marketing AI Institute offers some clear-eyed views on the report in his excellent podcast. He warned the 95% claim is directional, not definitive, due to methodology limitations. He points out that the report used the number “0” far too often to be credible. 

Like him, although the report seemed to validate many of the things we see in practice, I’m also saying we should treat headlines with a healthy dose of scepticism.

But even if the 95% stat is off, the pattern is real: transformation doesn’t fail for lack of ambition. It fails for lack of alignment.

JOURNEY helps enterprise businesses design and deliver AI and digital transformation that sticks. Get in touch to learn how Solve & Evolve™ can work for your team.