According to Kilshaw, today’s AI-driven transformation differs significantly from previous technological shifts in several crucial ways.
The speed and pace of evolution are happening at unprecedented levels, with Kilshaw explaining, “In just two and a half years, we’ve gone from the initial release of ChatGPT to models that surpass PhD-level intelligence across several domains. Even for those of us focused full-time on this space, the pace of change makes it challenging to keep up.” Many prior technical advances provided knowledge and support (remember Microsoft’s Paperclip?) but were not comparable to human intelligence. Today’s AI capabilities have swiftly matched or surpassed human expertise, for example, outperforming humans in tasks like medical imaging analysis and creative work. The pace of this huge step change is generating existential anxiety for many professionals.
The gap between early and late adopters produces significant tension as individuals and organizations embrace AI at varying speeds. Kilshaw cited research by Slack/Salesforce, which categorized people into distinct personas: for example, the ‘maximalist’ who embraces AI fully, the cautious ‘observer,’ and the resistant ‘rebel’. These different adoption rates create organizational friction as teams attempt to collaborate across varying levels of AI integration and acceptance. Even within departments, the uneven distribution of AI skills and comfort levels presents management challenges that previous technological transitions did not produce to the same degree.
The concepts of trust and explainability represent another key difference in AI adoption compared to previous technologies. “If you’re not transparent about the ‘black box’ and how AI makes decisions, someone who’s been making those decisions as a human for 15 years will doubt the veracity of the recommendations,” Kilshaw notes. This challenge is even more substantial with more recent reasoning models. Recent research by Anthropic (Claude) found that approximately three-quarters of responses from generative models lacked transparent explanations about their reasoning process. The ethical stakes are also much higher than with previous technologies, with Kilshaw cautioning, “Concerns about rogue AI, data misuse, and a competitive ‘race to the bottom’ are significant, especially in the absence of robust global collaboration or governance.”
Beyond technical considerations, AI raises questions about which applications are appropriate, even when technically feasible. “Just because you can do something cheaper or faster with AI doesn’t mean you should be doing it in the first place,” Kilshaw emphasizes. “Consider examples like exclusively using AI for performance reviews or recruiting – you could do it very efficiently – but is it the right thing to do ethically?” It’s important to consider where it’s right to use decision automation versus decision intelligence, where there is a “human in the loop”.
These questions push organizations to develop more sophisticated frameworks for evaluating not just whether AI can improve a process, but whether such improvements align with organizational values and maintain human dignity in the workplace. This ethical dimension represents perhaps the most significant departure from previous technological revolutions, which primarily focused on efficiency gains rather than fundamental questions about the human-machine relationship in the new world of work.
“A key transformation pattern is using AI to break down organizational silos through data integration.”
Several common themes emerge about how AI is reshaping organizations:
A key transformation pattern is using AI to break down organizational silos through data integration. At Sanofi, the pharmaceutical giant formed through more than 400 acquisitions, AI played a crucial role in unifying disparate parts of the organization.
“Sanofi created an app called ‘plai’ that was available on employees’ phones. It takes company data from across the ecosystem and enables AI-powered decision intelligence across portfolio planning, commercial pipelines, forecasting, and manufacturing quality,” Kilshaw explains.
This approach helped realize CEO Paul Hudson’s vision of creating “One Sanofi” by enabling radical transparency across the company’s diverse operations. In regulated industries like pharmaceuticals, this integration must navigate complex regulatory boundaries while still unlocking value.
In research-intensive sectors, AI is revolutionizing development cycles and discovery processes. “What Sanofi has been doing with OpenAI and specialist providers is focused on shortening the life cycle between molecule discovery and getting through clinical trials and FDA approval,” Kilshaw explains. “If you can shorten that cycle, it’s huge – you get more efficiency from your R&D pipeline and help more patients more quickly.”
AI also helps identify potential failures earlier: “If you can spot data that suggests a drug is unlikely to be successful in phase one rather than phase two, you save significantly. Clinical trials become increasingly expensive as they progress, given the larger patient groups.”
Organizations need common data architecture and methodologies while maintaining flexibility for customized applications.
Successful AI implementation requires careful balancing of standardization and differentiation. Organizations need common data architecture and methodologies while maintaining flexibility for customized applications.
This balance extends to acquisitions and integrations as well. For example, when bringing AI capabilities into established organizations, maintaining some degree of separation can preserve the innovative culture of acquired teams while still leveraging organizational scale and resources.
As AI capabilities grow, what remains distinctly human becomes increasingly valuable. Kilshaw identifies several key capabilities that will define successful professionals in an AI-enhanced future:
Learning agility: “Learning is the only competence that will never go out of fashion, and that’s more relevant now than ever,” Kilshaw asserts. Organizations must create psychological safety where people can experiment with AI tools without fear of making mistakes. If AI can increase our productivity, could those gains be repurposed towards dedicated investments in your personal growth, to keep pace with disruption?
Trust but verify: “Don’t take AI at its word. You need ways to spot inconsistencies, fallacies, and hallucinations,” cautions Kilshaw. “The same way you wouldn’t follow GPS off a cliff just because it tells you to; you must apply common sense to AI recommendations.”
An opportunity to leapfrog: “AI isn’t just a faster horse,” Kilshaw explains, referencing Henry Ford’s famous – and possibly apocryphal – quote in which he said if he asked his customers what they wanted, they would say ‘faster horses’ rather than an automobile. “Don’t just digitize or add AI to existing processes. Step back and ask what AI allows you to do that was previously impossible through new levels of breakthrough capability or capacity.”
If AI is reshuffling organizational operations and skills, it means the learning and development function needs to adapt too.
If AI is reshuffling organizational operations and skills, it means the learning and development function needs to adapt too. Here is how:
AI impacts the learning and development function in two ways: the function itself is changing (inside-out effect), and the function needs to help the organization learn and embrace the new technology (outside-in effect).
Traditional L&D has long followed the ADDIE model – Analysis, Design, Development, Implementation, and Evaluation – a structured approach involving needs assessment, instructional design, content creation, delivery, and measurement of outcomes.
Similarly, the Kirkpatrick evaluation model has been the standard for measuring learning effectiveness across four levels: reaction (participant satisfaction), learning (knowledge acquisition), behavior (application of skills), and results (business impact).
According to Kilshaw, AI is now completely reimagining these established processes.
“The traditional ADDIE model involves structured steps: interviews, focus groups, design, testing, RFPs, outsourcing, and ‘smiley sheets’ at the end for evaluation,” Kilshaw explains. “I’m building a system that instead takes curriculum documentation, throws it into a vector store, and uses an AI interface that tracks emotional context, frustration levels, and learning challenges in real-time.”
This approach transforms both content creation and measurement. “Instead of traditional learning models, you can hyper-personalize education, start new programs in an hour, and track learner progress individually,” Kilshaw notes. “This system has 50 measures of learner progress, including detecting frustration through mouse movements or voice tone.” The implications for learning evaluation are profound: “The whole idea of Kirkpatrick and four levels of learning evaluation has just disappeared,” Kilshaw asserts. “You can now continuously measure emotional context, learning efficiency, and change adoption in real-time, providing significantly more insight than traditional evaluations.”
Everyone needs to get into the mindset that AI is not managed by the digital department, the same way there wasn't an Excel department.
How can organizations help their employees adapt to this rapidly evolving technological landscape? Kilshaw recommends several practical approaches: “Everyone needs to get into the mindset that AI is not managed by the digital department, the same way there wasn’t an Excel department where you would send your spreadsheets 20 years ago,” Kilshaw states. “This is something we all need to embrace directly.”
He emphasizes domain-specific applications: “If I’m in finance, how do I use AI for financial benefits? If I’m in sales or marketing, how does it help me? Contextualizing AI within your work makes it tangible and creates real use cases.”
Organizations must identify and offer structured opportunities for experimentation: “Create time and capacity and psychological safety where people can play with these tools. Give employees the resources, teach them about ethics, and show them how to apply AI to their specific work challenges.”
This is where learning and development play a crucial role in helping organizations adapt. At Shell, for instance, Kilshaw helped apply category management insights from his Nike consumer goods experience to their approach to decarbonization. “Shell wanted to move from a general decarbonization approach across the value chain to a more sector-targeted one, recognizing that you decarbonize aviation differently than road transport, which is different to construction,” he explains.
This cross-pollination of ideas demonstrates how L&D can facilitate unconventional thinking: “AI probably wouldn’t have thought of that approach — bringing a consumer goods category model into oil and gas decarbonization. Sometimes you need more than the human-in-the-loop; you need divergent versus convergent thinking.”
What's left for humans includes empathy, emotional intelligence, leadership, critical thinking, strategy, and connecting the dots.
For professionals navigating this rapidly changing landscape, Kilshaw offers pragmatic advice, having worked across several differing industries. “Think about what transferable skills you bring that would be valuable (or sorely missing) in another industry,” he suggests. “Be open to exploring and challenging your assumptions about different industries – not all pharma is slow; not all consumer companies are fast. There are subjectively good and bad companies in every sector, with good and bad being highly personal and subjective.”
Kilshaw emphasizes the value of cross-industry experience: “I’ve found that high-performing teams include people with massively different skill sets and experiences. When diverse teams tackle problems, you get five different approaches instead of groupthink.”
While acknowledging that the human-AI transfer of effort and work is a “one-way baton pass to a machine as it’s not going to come back to you at some point,” he warns against over-specialization in knowledge domains. “AI is targeting knowledge-heavy domains – like medical diagnostics and constitutional law – first. What’s left for humans includes empathy, emotional intelligence, leadership, critical thinking, strategy, and connecting the dots. These skills are more portable and transferable than domain-specific expertise, so I believe it will promote more cross-industry moves in the future than less.”
The AI revolution is fundamentally different from previous technological shifts – faster, more capable, and potentially more disruptive to established ways of working.
Organizations that succeed will balance standardization with differentiation, embrace continuous learning, and focus human talent on what remains distinctly human.
As Kilshaw reflects, the organizations best positioned to thrive are those that view AI not merely as a tool for efficiency but as a catalyst for reimagining what’s possible. Whether in pharmaceuticals, energy, or consumer goods, the competitive advantage will go to those who can blend the best of human creativity with AI’s analytical power while maintaining a clear ethical compass to guide this transformation.
All views expressed herein are those of the author and have been specifically developed and published in accordance with the principles of academic freedom. As such, such views are not necessarily held or endorsed by TONOMUS or its affiliates.
Leading consumer goods companies are strategically embracing AI across their entire value chains, creating unprecedented business capabilities and market differentiation. IMD’s AI Maturity Index reveals the strategies that separate AI innovators from...
AI and sustainability are transforming the business landscape. In their new book, Michael Wade and Konstantinos Trantopoulos introduce the concept of twin transformation – a leadership imperative for companies seeking to remain...
Brian Evergreen shares radical insights from his book Autonomous Transformation: Creating a More Human Future in the Era of AI with IMD’s Stefan Michel....
AI's true revolution isn't in chatbots but in transforming industrial operations through dark factories and predictive maintenance. This will require leaders to balance technological advancement with responsible societal impact....
Explore first person business intelligence from top minds curated for a global executive audience