In thirty years and more than 400 client engagements, one thing never changed. Every core competency has a shelf life. AI is shortnening that shelf life and which competencies are even possible to build in the first place.
We call that competence velocity: the speed at which a competence is built, proven, or made obsolete. Until the mid-nineties it was slow. A company that moved faster than its industry's pace of change was safe, and industry boundaries defined the set of competitors that needed to be watched. Digital transformation had far-reaching implications by blurring those boundaries. AI is similar in its broad impact but also different in that it's bounded only by our imagination. It can destroy competence-based advantages in some places while creating them in others, at a pace far greater than before and in ways that at times give cause for wonder as well as concern.
Emerging technology has always been one of the lenses we apply to a client's situation. AI is the first one to make us re-examine the analysis itself.
Four moves follow from that, and we are running all four with our clients:
1. Test which of your competencies AI is eroding, and which new ones it lets you build
Prahalad and Hamel asked one question. What is this company uniquely good at, and what future does that entitle it to build?
We hold every claimed competency to four tests. Does it create real value for a customer? Is it genuinely unique? Can it be sustained? Can it be extended into a market you are currently not considering? Most leadership teams believe there are a dozen or more competencies their competitive advantage relies on. Across our engagements, an average of 80 per cent of those competencies fail. One or two survive as true core competencies.
That ratio holds true regardless of AI. But two things around it have moved. The first is how often the surviving competencies need re-testing. A competence that passed two years ago may now be something any competitor can easily acquire or quickly build themselves. The second is who the challenger is. The competency once threatened by a known rival in your own sector is now just as likely to be threatened from outside your industry, and outside the geography you believe is yours. The challenger builds faster than you can defend.
Two questions are worth answering before starting anything else. Do you know which of your competencies pass that filter, or are you still defending a dozen of which most won't pass the tests? And which one could a company outside your industry render irrelevant from your customers' perspective?
Case
manufacturing
One client sold a highly successful product. Apply the four tests to that business and the product created real customer value. It failed on uniqueness. Competitors sold something close enough, and margins were eroding.
The same client had been collecting data for years on how that product performed once it left the factory, and had built some capability in industrial IoT. Neither had been treated as a competence. Combined they passed all four tests. No competitor held that data, none could reconstruct it, and it pointed directly at a market the company was not currently operating in.
Machine learning turned it into the ability to predict equipment failure before it happened. That became a services business, sold to address the problem the customer actually had, which is the cost and disruption of equipment breaking down. Their customers saw emergency breakdowns fall by 75 per cent and maintenance costs fall by 70 per cent. The product barely changed but the competitive position moved significantly.
2. Capture the efficiency gains and the learnings from AI, but then move beyond them
Every company directs AI at increasing efficiency first, the same way it directed automation at efficiency in the nineties. The use cases are easy to find and the payback is quick. That work is worth doing because it improves your cost structure and frees the budget for the innovation your customers will actually value. It also takes your teams up the AI learning curve while the stakes are low.
Radical Simplification, the Strategos approach to step-change improvement in core operational processes, is the right way to deploy AI in this case. It does not optimise the current design. It challenges the assumptions that design was built on, then asks what the process would look like if you designed it from an AI-first perspective today.
Deploying off-the-shelf AI tools produces task-level speed-ups of 10 to 15 per cent. Your competitors are buying the same tools from the same vendors and getting the same 10 to 15 per cent. None of it reaches your customer unless it is converted into something they value.
The innovation work therefore runs in parallel, as soon as you have travelled up that initial learning curve.
3. Explore how AI drives new options for where you play and how you win
Differentiation still means offering customers value that competitors cannot match. Start with your existing innovation portfolio. Which positions in it can AI strengthen, and which could it invalidate? That is a short exercise but it quickly highlights changes needed in the portfolio.
Options exist now that were not viable before, and some that were unthinkable because the competencies could not be built. Opportunities that were once uneconomic to explore can be tested in weeks. Competitive positions that used to depend on assets you own may now depend on how you apply AI to deliver new propositions for your customers.
The strategies available for pursuing those options have changed too. This is where most companies fall short. They ask how AI can execute the strategy they already have, instead of challenging the underlying assumptions and recalibrating that strategy.
Case
food production
Keeping animals healthy has traditionally meant preventative medicine, or treatment once illness appears. Sick animals infect others, and the losses spread across the whole operation. Preventative treatment is expensive, and most animals never needed it. The global animal pharmaceuticals industry built on that solution is worth $46 billion a year.
In this case AI was used to read behavioural patterns and flag animals at the earliest stage of a developing health problem, while intervention is still cheap. Treatment becomes targeted. Cost falls, the risk of building resistance falls with it, and fewer animals are lost.
Consider where the winning core competence now resides. The traditional competence relies on drug discovery. This one runs on understanding behavioural data, which no pharmaceutical company holds and none of that $46 billion market was built on. That is a direct challenge coming from outside the industry.
4. Embed AI and build new capabilities to increase innovation performance
AI cannot create a winning strategy or run an innovation engine on its own. We apply it to something narrower: speed, volume, quality and range of imagination in work that people still drive and own. Someone must verify the output. Someone must be accountable for the decisions.
Governance and evidence standards in particular carry more weight than they used to. AI models aim to please and can invent sources. When they do, the answers are always plausible. A strategy built on non-existing evidence is worse than one built on nothing, because it carries a false sense of confidence in our assumptions. We have built verification into our system. It takes every source referenced in any research, confirms it exists, and confirms it contains the claim attributed.
That is the shape of the capability question. BCG's allocation rule for AI work is instructive. Ten per cent of the effort goes to algorithms. Twenty per cent goes to data and technology. The remaining 70 per cent goes to people, process redesign and the change in how the work is actually performed. Seventy per cent of the work lies where companies were already struggling before AI arrived.
The gains are real where the discipline is already there. On one recent engagement, discovery work that would have taken two weeks ran in two days. The principles have not changed but the methods and tools for strategy and innovation have.
This is also the strongest argument we know for working Future Back. When competencies decayed slowly, planning forward from your current position was defensible, because the position existed long enough to plan from. At today's competence velocity you are optimising a position that may already be dissolving. Start instead from the position you intend to hold in the future, then trace back the competencies, bets and experiments that get you there.
Companies that run this exercise once end up with a better strategy. Organisations that embed this as an enduring capability have something that helps improve overall resilience, and they come prepared when the next big shift arrives.
The same question, at a different speed
Core competencies have always eroded. The only question that has ever mattered is whether an organisation can build new ones faster than the old ones decay. AI raises competence velocity on both sides of that equation.
Most companies use AI to execute their existing strategy better. The ambitious ones use it to find new options in markets they have not explored yet, in their core business, adjacent to it, and beyond it. And to find them first.
