Organisations and societies face a crossroads when it comes to Artificial Intelligence. Will the AI of the future be defined by raw computational power, or by its ability to address real human needs, unlock equity, and act responsibly? In the latest episode of the Cambridge Executive Business Insights: Rethinking AI podcast, Jaideep Prabhu, Professor of Marketing at Cambridge Judge Business School, sits down with Mark Bloomfield, founder of Turbulence, to cut through hype and examine how AI can be a lever for thoughtful, resource-efficient innovation.
Framing the AI discussion: from bigger to smarter
An emphasis on purposeful and frugal AI, a theme resonant throughout the episode, asks us to move away from the allure of headline-grabbing, resource-intensive models towards technology that is inclusive, efficient, and deeply connected to human realities. In an era where “move fast and break things” is giving way to “move fast and make things,” the invitation is clear: how can we make technology not just powerful, but meaningful?
Democratisation and the rise of AI as organisational capability
According to Mark Bloomfield, the last three years have seen a genuine democratisation of AI. Tools like ChatGPT and Claude have brought powerful generative models to the fingertips of millions, moving AI out of the realms of technical specialists and into the hands of everyday problem solvers. Yet with this shift comes both opportunity and challenge.
AI is now becoming critical infrastructure for most organisations. No longer the preserve of “nerds in the corner”, its promise stretches beyond automating routine tasks to the full-scale reimagination of processes, growth strategies, and even organisational culture. But this very broadness of potential creates new complexities for leaders: how to sift real value from hype, and navigate the “AI FOMO” that has overtaken boardrooms.
Efficiency versus growth: finding the right balance
A recurring tension in the modern enterprise is whether AI should be prioritised as a vehicle for efficiency, for example for reducing costs and automating workflows, or as a catalyst for innovation and growth. Mark Bloomfield acknowledges that, in a world of quarterly results and investor scrutiny, boards often default to efficiency, encouraged by vendor pitches promising rapid, measurable ROI.
Yet the deeper, longer play of AI is in enabling growth. Generative AI’s capacity for simulation, that is, allowing teams to test, refine, and de-risk ideas virtually, can accelerate innovation cycles and open new frontiers of value. The organisations that will thrive, Bloomfield argues, are those able to pursue both streams: extracting efficiency gains in the short term, while investing resource and attention in the complex, sometimes uncomfortable, work of reimagining their business for the future.
The frugal AI imperative: doing more with less
So what does “frugal AI” look like in practice? Here, the discussion draws on Bloomfield’s background in aerospace and his fascination with nature-inspired optimisation, particularly the famously efficient foraging behaviours of ant colonies. Ants, he points out, learn from each other, coordinate without centralised control, and continually optimise paths with stunning computational thrift.
Translating these insights into the business world, AI becomes a tool not just for brute-force analysis but for de-risking, rapid prototyping, and collaborative problem-solving at scale. For example, creating AI “personas” can allow teams to challenge assumptions, scrutinise ideas, and simulate outcomes swiftly and at low cost, unlocking frugal innovation that might once have required significant investment.
Critically, this approach isn’t about compromise. Rather, it asks executives to find friction points – those high-pain processes or organisational bottlenecks – and target AI interventions where they most readily free up human capacity for creative, higher-value work.
Navigating AI hype, FOMO, and organisational culture
With “AI transformation” now a staple of business magazines and consultancy pitches, leaders often feel enormous pressure to “do something” with AI simply to keep up with the Joneses. The risk, says Bloomfield, is that companies find themselves rolling out tools like Microsoft Copilot or Google Gemini without strategic intent, mistaking box-ticking for value creation.
To counteract this, Bloomfield urges leaders to distinguish between AI curiosity, initial experimentation with new tools, and genuine capability, where AI is embedded in the reimagining of workflows and processes. Language matters here: being “AI-first” is less important than being “problem-first”, putting organisational pain points front and centre and letting AI follow as an enabler.
Equally vital is modelling. Cultural change, Bloomfield argues, happens not just through top-down pronouncements but by leaders at every level demonstrating how they want AI to be used, both intentionally and responsibly, and in the service of organisational learning. When paired with a culture where experiments and lessons are widely shared (avoiding duplication and re-inventing the wheel), this builds trust and helps teams focus on meaningful impact rather than technology for its own sake.
A case in point: fixing real problems with AI
Behind every successful AI journey lies a concrete use case, often in the unglamorous, behind-the-scenes processes that, while “low risk,” soak up time and create churn. Mark Bloomfield recounts his work with a US financial services company, where the temptation was to expect a transformation purely from rolling out an off-the-shelf tool. However, after a costly false start, the company instead focused on a key source of organisational friction: high staff turnover in a particular department.
The root cause? Overburdened HR teams with little capacity for development or meaningful employee support. By deploying AI agents to handle transactional queries, the company dramatically increased HR’s ability to focus on high-impact work, and crucially, made it clear up front that technology was there to augment, not replace, human talent. The result: a dramatic drop in attrition, a more engaged workforce, and a flywheel of receptiveness to further AI-enabled problem solving.
Friction audits and the art of human-centred AI
Central to Bloomfield’s philosophy is the concept of a “friction audit”, systematically identifying points in the organisation which elicit frustration, eye-rolls, or sighs from teams, and targeting these as candidates for AI intervention. In doing so, organisations can rapidly build confidence and showcase ROI, whilst dispelling fears that AI is solely an engine for cost-cutting or job displacement.
The aim, he asserts, is not to subordinate humans to machines, but to ensure that AI removes drudgery, liberating capacity for the sort of uniquely human work such as judgement, creativity, relationship-building, that sustains long-term competitive advantage.
From “AI-enabled” to “problem-native”: a new lens on transformation
Successful AI-adoption, Bloomfield contends, isn’t about attaching technology to every business unit, but about cultivating an organisation that is “problem-native”: one that focusses relentlessly on the issues faced by customers and colleagues, and is open to creatively deploying AI capabilities (and, indeed, any resource) to solve them.
This philosophy requires both humility and bravery. Humility, to revisit processes that “worked” in the analogue era and ask: do we even need to do this anymore? Bravery, to empower teams to take ownership of AI-driven experiments, recognising that the journey may be non-linear and that AI capability itself is evolving at unprecedented speed.
No “point B”: continuous learning and reimagination
If traditional change management describes moving from “point A” to a pre-defined “point B”, AI upends this model entirely. Today’s technology is, as Bloomfield puts it, “the worst it will ever be again.” Organisations must get comfortable not with single, linear transitions, but with continuous cycles of learning, capacity-building, and reimagination.
Pragmatically, this means both top-down and bottom-up mechanisms: leaders modelling (not merely mandating) intentional AI use, and front-line teams empowered, with time and resources, to test and adapt workflows as capabilities evolve.
Personalising AI transformation: Turbulence as a living example
Much of Bloomfield’s advice is grounded in his own experience building Turbulence, where he made the intentional choice to remain a solo founder and leverage AI not to displace people, but to give himself more human capacity for creativity and judgement.
From building AI “non-executive directors” tasked to challenge his own decision-making, to employing AI researchers who summarise troves of insights and data, his approach is not about automating away autonomy, but about reclaiming time for reflection, strategy, and curiosity.
Yet this approach demands discipline: Bloomfield is clear that humans must resist the temptation to fully “outsource” their thinking to AI, building in regular reflection – “AI mental audits”- to ensure that judgement, authenticity, and intuition are preserved.
Scaling beyond pilots: risk, governance, and trust
The journey from successful AI pilot to scaled enterprise capability is fraught with technical, regulatory and often cultural obstacles. A key challenge is risk: as AI systems move from back-office process automation to customer-facing roles, anxieties about bias, hallucination, and unpredictable outcomes multiply.
Bloomfield advocates for a rigorous simulation approach, using generative AI itself to create synthetic customers, simulate board debates or regulator responses, and stress-test new ideas before full-scale rollout. Governance then becomes an enabler, not a roadblock, granting space for experimentation while ensuring responsible oversight as initiatives mature.
Human-centricity to the fore: AI as a productivity multiplier, not a threat
Perhaps the most pernicious myth in contemporary AI discourse is that of mass human obsolescence. Yet in organisation after organisation, including those Bloomfield advises, the evidence runs counter to the dystopian narrative. Human skills remain indispensable: overseeing, refining, and integrating AI solutions in context requires judgement, collaboration, and, above all, curiosity.
This is especially evident in the next generation, as illustrated by Bloomfield’s own children, whose playful interrogations of AI, sometimes to further their own interests, highlight the centrality of curiosity, adaptability, and ethical reasoning for the AI-literate workforce of the future.
Responsible AI, ethics, and the sustainability paradox
As the AI landscape broadens, so too does scrutiny over its impact on privacy, fairness, accountability, and, increasingly, sustainability. Energy consumption, water use, and the extractive toll of massive data centres raise searching questions for a technology sector that aspires to do more good than harm.
Yet Bloomfield is optimistic, pointing to a likely shift away from ever-larger language models towards more efficient, localised approaches: “The phraseology itself contains three assumptions: large, language, model… But if the future is small (smaller models, less resource-intensive) then frugal AI becomes not just an ethical imperative, but ultimately a business one.”
Ultimately, trust sits at the heart of responsible AI. Whether it’s transparency in model reasoning, explainability in automated decisions, or accountability mechanisms for error and bias, organisations must be intentional and proactive if they want to build cultures, both internally and with customers, where AI is welcomed as a partner.
Transitions, empathy, and building for the human future
As the episode closes, Bloomfield offers a reflection apt for a period of discontinuous change: “Change is situational. Transitions are psychological.” Underlying this technological revolution is a human transition, one that demands empathy, warmth, and a shared effort to understand what roles humans will play, and what values must shape the future.
Practical steps for leaders: small, focused, and human-first
For those seeking an actionable starting point, Bloomfield is unequivocal: “Start small. Pick a process or workflow behind the scenes that’s low-risk, high-friction. Be open to how AI can help you solve and reimagine it. Prove the value, and then build the flywheel from there. This builds confidence, trust, and literacy.”
Frugal, effective AI adoption is less about revolution than about a steady accumulation of human-driven wins.
Towards a future of purposeful, frugal, and human-centric AI
The vision emerging from this conversation is neither utopian nor apocalyptic. Instead, it is a call to treat AI as a continuous partner in organisational evolution: a capability to be embedded, scrutinised, and reimagined in perpetuity. Whether through the design of more efficient HR systems, the empowerment of employees to solve friction points, or the refusal to cede judgement and creativity to algorithms, purposeful AI is within our collective grasp.
As businesses, public sector organisations and individuals seek a path through hype, risk, and uncertainty, the lessons from Cambridge Judge Business School’s conversation with Mark Bloomfield are clear: start with real problems, foreground human value, and let technology follow as a capable ally.
To delve deeper into these themes and explore more real-world examples, listen to the full episode of the Cambridge Executive Business Insights: Rethinking AI podcast, wherever you get your podcasts.
