AI is here whether we like it or not.

For many, the scale and intensity of vendor driven hype is worrying; the uncertain amongst us may prefer to wait and see if others pursuing the promise of AI manage to strike a profitable balance between risk and reward.

Those determined not to be late to the party, perhaps already teetering on the edge of enterprise-wide AI investment, won’t be able to rely on hard-won lessons from the past. Nearly 50 years of research and practice has helped refine our collective understanding of IT adoption and exploitation. However, whilst some aspects of AI adoption will be familiar to many (i.e. the pursuit of new business outcomes, developing new capabilities that enable business owners to then realise benefits), there is a fundamental difference when considering AI.

AI outcomes are not only likely to be less predictable than those secured from non-AI systems, but AI outcomes could become increasingly undesirable or even damaging to the organization.

What we’re talking about here is an unprecedented level of ongoing uncertainty in outcome, aka, an increase in risk.

So, can anything be done to address this?

For the less gung-ho, sponsors and investors may wish to consider a rethink in expectation management as they start to make the case for enterprise-wide AI adoption. They should actively:

• Diminish stakeholder expectations for tangible, financial returns from AI investment, by:
  • Limiting AI adoption i.e. invest in small, localised low risk projects that serve to ‘educate’ more than they are expected to produce acceptable (£) returns
  • ‘Over-engineering’ solutions – emphasising the importance of restricted access and usage over the need for quick returns returns

Whilst,

• Enhancing expectations for additional investments in:
  • New management information capabilities that proactively monitor how AI outputs impact expected business outcomes
  • Managing the ongoing costs of ‘failure’. Due to its non-deterministic nature, AI is likely to introduce unexpected (wanted and unwanted!) outcomes that are increasingly difficult to investigate, unpick and correct as companies struggle to develop ever more sophisticated, preventative measures. Even for organisations that believe track and fully understand the current costs of failures, AI will introduce a whole new level of opaqueness that will require additional resources to help unravel unexpected issues.

For those planning to sell the idea of AI adoption internally, this stance may appear overly defensive or even counter productive; however, an over-cautious approach may pay dividends in the long run as levels of trust improve and AI initiatives take hold.

What else should enterprise wide AI investors be thinking about?

Some of the (many) other questions I’m still wrestling with, especially before advising clients about AI adoption include:

1. Does an increase in digitally maturity really give one sustained competitive advantage? Has this link been established?
2. Does AI really represent a new critical resource?** One that directly contributes to sustained competitive advantage? Has that link been established?
3. If AI does represent a new critical resource, a strategic imperative, then experience suggests that AI uptake would form part of a wider (transformation) programme. Would such a programme then comprise of multiple discrete AI projects with few interdependencies, coupled with wider business change projects e.g. impacting corporate standards, organizational controls and structures?
4. Given the trend towards cloud, investment cases are likely to be focussed almost entirely on payback for OpEx (rather that CapEx)? Is that right?
5. Purveyors of AI have focused on the potential impact of AI on operational performance; however, what we’re really talking about here is an increase in levels of automation that eventually lead to a deferment in decision making (or an ‘outsourcing’ of decision making’). How will that impact the roles of ‘manager’ and ‘leader’? Has any work been done in this area?

Standing at the precipice of enterprise-wide AI adoption appears equally terrifying and exhilarating; the promise of AI seems terrific.

But are we ready for it?

* Thanks to Stephen E. Morris for the title for this article

** By critical resource, I’m thinking in terms of Barney who suggests that firms can achieve ‘sustained competitive advantage’ (or SCA) by developing certain critical internal ‘resources’ (Wernerfelt, 1984) to develop superior financial performance. Barney goes onto to describe such resources as being, ‘Valuable’, ‘Rare’, ‘Inimitable’ and ‘Non-Substitutable’ (or VRIN)’ (p.105) (Barney, 1991); it is in the gift of the organisation to acquire, maintain and develop such resources and that these resources should focus on creating ‘value’ as a means for achieving SCA (Porter, 1985).

Image attributed to:

Valdermaras, D. ‘Three-persons-standing-near-the-edge-of-a-cliff-during-day’. Available at https://unsplash.com/photos/three-persons-standing-near-the-edge-of-a-cliff-during-day-Nx2FTJfKX68 (Accessed 09/06/25)

Note:

As with all my articles, I never use AI to generate content (sans AI); however, in keeping in line with the title of this article, I couldn’t resist asking Co-Pilot ‘How do gold rushes end?’

This is its reply:

‘A gold rush typically concludes when the easily accessible gold is depleted, leading to a decline in prospectors and miners. This can happen for several reasons:

1. Resource Depletion: Once the surface gold is mined, it becomes harder and more expensive to extract remaining deposits.
2. Economic Factors: The cost of mining may outweigh the profits, especially if gold prices drop or operational costs rise.
3. Environmental Regulations: Increased scrutiny and regulations can limit mining activities.
4. Social Dynamics: As the initial excitement fades, communities may disperse, leading to a decline in population and support services.

Ultimately, a gold rush often transitions into a more stable mining industry or a ghost town, reflecting the cyclical nature of resource extraction and economic opportunity. Exploring these factors can provide a deeper understanding of the lifecycle of a gold rush.’ (09/06/25)