Enterprises are pouring resources into two AI frontiers: generative assistants (think ChatGPT) and digital twins (virtual replicas of physical assets). Both promise transformation. But both, in their current architectures, quietly sideline the very thing that makes an organisation unique, its collective human expertise.
Bloor’s argument is simple: in the rush to automate, we risk hollowing out the core value that keeps businesses resilient and differentiated. Let’s examine why.
What Is a Digital Twin And What Is It Not?
A digital twin is a real-time, virtual representation of a physical object or system. It’s a discipline honed in heavy industry: manufacturing, aerospace, infrastructure. The twin simulates, predicts, and optimises the physical asset’s performance; machines, not minds.
Key features:
- Models things, not people.
- Focuses on efficiency, maintenance, and operational risk.
- Mature field, but no bridge to capturing a person’s know-how or judgement.
Generative AI: Powerful, But Whose Value Does It Amplify?
Generative AI, in contrast, promises to turbocharge knowledge work: drafting documents, summarising research, answering questions. The narrative is seductive—AI as the universal assistant, levelling up every desk job.
But here’s the catch: generative AI doesn’t preserve what your people uniquely know. It produces outputs based on vast, generic training data. It’s synthetic, not sovereign. The expertise it draws on is collective, but not contextual to your business. When your staff leave, does their judgement remain? Not unless you’ve captured it elsewhere.
Generative AI is a productivity amplifier. But unless it is paired with a mechanism to encode, own, and continuously update your organisation’s tacit knowledge, it risks commoditising what should be a source of competitive advantage.
The Shared Blind Spot: Human Expertise at Risk
Both digital twins and generative AI miss the mark on one crucial front: neither is designed to preserve, transfer, or enhance the human expertise embedded in your enterprise. The industrial logic behind both is about optimisation—of assets, of workflows, of generic outputs. But the digital logic for the next wave of value is about capability: who owns it, who grows it, and how it’s protected when people move on.
Strategic Depth
- Bloor’s model insists on “pull-through operating model thinking” – how do these technologies fit into, or disrupt, your core value creation system?
- True transformation connects AI investment to macroeconomic outcomes, not just tactical wins.
- The real differentiator is not feature comparison, but operating model resilience: what happens to your business when the AI stops, the twin fails, or the workforce turns over?
Ownership, Value, and Liability: The Unspoken Questions
Ownership:
With both generative AI and digital twins, the knowledge and models are typically owned by vendors or reside in silos. There’s no institutional memory, no mechanism for individuals to “take their expertise with them” or for enterprises to meaningfully transfer judgement across teams.
Value:
What’s being automated is not always what’s most valuable. Productivity gains are real, but what about the context-specific insights, the “why” behind the “what”? If you automate without preserving the underlying logic, you’re at risk of building brittle systems – efficient, but fragile.
Liability:
If your business relies on either tool as a proxy for expertise, you inherit hidden liabilities. When judgement errors occur, who is accountable: the machine, the model, or the person who is no longer there?
A Real-World Scenario: Expertise Lost in Translation
Imagine a manufacturer with a fleet of digital twins monitoring every machine. The twins predict wear and optimise maintenance schedules. But one veteran engineer retires, taking with her decades of tacit knowledge about how those machines behave under edge-case conditions, knowledge the twins never encoded.
Now layer in generative AI: new staff can ask questions, but the answers are based on generic manuals, not the engineer’s lived experience. The result? When the unexpected happens, the business is less resilient, not more.
The Operating Model Fault Line: Building for Resilience, Not Just Efficiency
Bloor’s analysis is that the next wave of AI investment must be operating model first, not tool first. The core question is: how do you design systems where human expertise is continually captured, updated, and owned, by the enterprise and its people, not just by vendors or platforms?
Key Principles:
- Move from output to outcomes: Don’t just automate tasks, measure the impact on capability and resilience.
- From automation to augmentation: Use AI to amplify human judgement, not replace it.
- From workers to enterprises of capability: Treat your workforce as a dynamic system of value creation, not a cost to be automated away.
What Enterprises Must Do Differently
Audit Knowledge Flows: Map where expertise lives, how it’s transferred, and where it’s at risk of loss. Don’t let digital twins or generative AI become black holes for context.
Design for Ownership: Build systems where both the individual and the enterprise can own, update, and transfer expertise. This means personal AI agents, not just vendor-owned assistants.
Mitigate Liability: Ensure that decision-making logic is transparent and traceable. Don’t allow “the model did it” to become an excuse for brittle outcomes.
Link to Macroeconomics: Connect your AI investments to broader operating model risks: talent churn, regulatory exposure, competitive resilience.
Pull-Through: The Fusion Workforce of the Future
The real opportunity is to create a fusion workforce, where human and digital capabilities are not in competition, but in constant dialogue. This means moving beyond the industrial logic of tools and towards a digital logic of capability, ownership, and long-term value.
Reflective Questions for Enterprise Leaders
- Where does your organisation’s unique expertise actually live?
- If your top talent walked out tomorrow, what would your digital systems remember – and what would they forget?
- Are you investing in tools, or in the architecture of capability and resilience?
Conclusion: Don’t Automate Away Your Advantage
Enterprises that treat AI as a tool for short-term automation will win some battles, but risk losing the war for long-term resilience. The organisations that thrive will be those that build operating models where expertise is a first-class asset—preserved, owned, and continually enhanced.
Final Thought:
Before you invest in the next AI wave, ask: are you building a system that protects what only your people know?
Frequently Asked Questions (FAQs)
No. A digital twin models a physical object or system, a machine, a facility, a network. A generative AI assistant is a general-purpose tool trained on public data. They solve different problems and aren’t typically confused with each other.
No. Neither was built to. A digital twin has no person to model in the first place, and a generative AI assistant has no persistent memory of any one person’s judgement between sessions.
No, closing this gap is about preserving and multiplying what an expert already knows, not substituting for them. The highest-value judgement calls still sit with the person.
A category built specifically for it, a personal digital twin, trained on one person’s own material and owned by them or their organisation, rather than a platform. Bloor Research’s version of this is Digital Me.
Print Article for this blog Download here