By Dr Maurizio Pilu, PhD (AI), eMBA, Fellow IET
August 2025
MIT recently reported that 95% of [Gen]AI projects fail. The stat got the web and investors in a frenzy. In fact, the stat is far more nuanced, and should be seen in the context of a boom in AI experimentation and well-established AI tools such as CoPilot and ChatGPT widely used in shadow or IT-endorsed forms.
Yet, is this surprising? Not really. It aligns with other statistics that report between 80-90% of failures in innovation projects related to new tech adoption.
Too often, organisations approach AI like an IT deployment with budgets, milestones, and a “go live no matter what” mindset. But AI isn’t (yet) IT. And neither is other emerging tech used, e.g. in safety-critical operations.
It’s innovation, with all the uncertainty, iteration, and discovery that innovation demands. Treating it like a predictable roll-out is why so many initiatives never deliver.
Yet with those poor stats flying around, innovation is, on one hand, obviously critical (especially now, with the AI revolution underway) and on the other hand getting a bad reputation.
What’s missing in many organisations is an operating model for innovation itself, something that elevates it from creative experimentation to a critical business function that needs to deliver impact and value for money.
If DevOps professionalised software delivery, what’s needed is to systematize and professionalise the approach to innovation, what I like to call #InnoOps.
Across my own career, having handled tens of millions in innovation budgets in and with corporates, startups, public sector bodies, and non-profits, I’ve seen the same patterns derail promising initiatives:
- Misfit use cases: initiatives that don’t address real business priorities.
- Ownership gaps: promising ideas often stall without a clear, accountable owner.
- Pilot paralysis: successful experiments fail to scale into real-world impact.
- Misaligned metrics: success measured by outputs, not outcomes.
- Cultural friction: experimentation remains a side-project, not embedded in daily operations.
Bue while everyone would know the theory, and there are standards emerging (such as the ISO 56000 series), addressing these in practice is the hard part in most organisations.
#InnoOps is all about operationalisation and day-to-day actual implementation of innovation in an organisation. It addresses the common practical challenges in the same way as other xOps do (like DevOps or ModelOps), giving organisations a structured way to run innovation with method and discipline to maximize impact and value for money.
While every organisation will implement #InnoOps it in its own way, its principles are clear and include:
- Accountability: clear ownership for turning ideas into results.
- Scalable pathways: moving pilots to meaningful adoption.
- Outcome focus: metrics aligned with value creation, not activity.
- Strategic alignment: initiatives tied to business goals.
- Cultural integration: experimentation becomes an integral part of the organisation’s DNA.
Would you like to discuss how could #InnoOps help your organisation shape its innovation journey?