What a board should ask when AI enters the work, from someone who has designed more than 30 operating models for successful businesses globally.
Tess Hilson-Greener explains her work by imagining a sceptical board. “AI will increase productivity and reduce employee costs. Why do we need to redesign the organisation?” Her first question back, she says, would be: “Where is AI entering your organisation?”
Map that properly, she says, and the scale of the question becomes clear. AI is entering workflows, jobs, enterprise systems, customer interactions, knowledge, decisions and, increasingly, autonomous actions. The organisation introduces it formally, and employees introduce it informally. The informal route isn’t small: Microsoft and LinkedIn’s 2024 Work Trend Index found that 78% of AI users were bringing their own AI tools to work.
Then comes what changes as a consequence. “If AI reduces a two-hour task to twenty minutes, you have created capacity; you have not automatically created productivity.”
Tess has spent more than three decades introducing technology into organisations and has designed more than 30 operating models for global businesses. Today she is the founder of the AI-Mediated Foundation, the creator of the PXB Ecosystem™, an enterprise operating architecture for organisations where AI is part of the work, and the author of The AI-Centric HR Operating Model, published by Kogan Page on 3 October 2026.
Dashboards and Policies
A board may already have AI dashboards, Tess says, but they measure what you have decided to count: licences, usage, time saved, costs. “They cannot, on their own, tell you whether AI has created a more capable and productive organisation.” The same goes for an AI policy and an ethics framework. Her question is what it costs to get AI wrong, and badly designed governance has a cost too: duplicated checking, extra approval layers, slower decisions. “They must be designed into the way work operates,” she says of ethics and governance, rather than wrapped around AI afterwards.
Those questions come out of a long career.
Thirty Years Before AI
Tess’s career has spanned more than three decades across the public and private sectors, academia and global organisations including Xerox, Canon, British Airways, FedEx, Siemens, BMW/Rolls-Royce, SITA, Astellas Pharma, Gilead Sciences, E.ON, Deutsche Bank, Santander, Deloitte, Inmarsat, Adecco and the Office for National Statistics.
Her experience encompasses commercial leadership, technology, business development, operations, organisational design, HR and enterprise transformation. She has built businesses and functions, including establishing enterprises for investors, and has designed and led major transformation, learning and capability programmes. She has developed thousands of managers and leaders, while the learning strategies she has created have reached tens of thousands of employees.
Her career also extends into academia and further education, where she has built commercial capability and connected learning with the realities of business and the changing world of work.
That breadth is central to her thinking about AI. She has experienced change from multiple perspectives as a business builder, commercial leader, transformation practitioner, organisational designer, educator and developer of talent while witnessing its consequences for investors, leaders, employees and business performance.
The Question That Became PXB
“The starting point was much more personal than that question suggests,” Tess says. In 2023 her father died. She had stepped away from her role as HR Director, Organisational Development and Diversity at Inmarsat to care for him. Afterwards, for the first time in a very busy career, she had the space to stop, reflect and think seriously about what she wanted to do next.
She began researching what AI would mean for the future of work. “The more I researched, the more uncomfortable I became with what I was seeing.” Much of the conversation focused on tools, use cases, efficiency and what AI could replace or accelerate. But she found herself asking a different set of questions: what happens to the organisation itself when AI becomes part of the work? What happens to the people working alongside it? And how do organisations redesign for a world in which human and artificial intelligence increasingly coexist?
The economics also concerned her. Claims about “AI impact” were frequently expressed through adoption, time saved, reduced headcount or individual productivity gains. “But none of those measures, on their own, demonstrates that you have created a better-performing organisation.” Saving time creates capacity. What matters is what happens to that capacity, what else changes around it, what new costs are introduced, and whether the organisation ultimately creates greater performance and value.
Those questions led her beyond the boundaries of HR. “AI does not respect functional boundaries,” she says. Changes to work affect roles and capability. Changes to decisions affect authority and accountability. Changes to customer interactions affect experience. All of them can affect organisational performance. She began building the architecture she could not see elsewhere, which became the PXB Ecosystem™.
What had started as a question about AI and the future of work had become a much larger one:
“How do you design an organisation when intelligence within it is no longer exclusively human?”
Her answer begins with three stages.
Three Stages
AI-enabled, AI-centric and AI-mediated aren’t three levels of technology maturity, Tess says. “They describe a changing relationship between AI, work, and the organisation.” An AI-enabled organisation puts AI into the organisation it already has, and jobs, workflows, decision rights and measures of performance stay largely as they were. An AI-centric one asks, instead of “Where can we use AI?”, “Now that this capability exists, how should we redesign the work and the organisation?”
“AI-enabled changes the tools. AI-centric changes the design of work and the organisation. AI-mediated changes how human and artificial intelligence operate together as part of the organisational system.”
Tess Hilson-Greener, Founder, AI-Mediated Foundation
In an AI-mediated organisation, AI “increasingly becomes part of the connective intelligence of the organisation,” she says. Employees may use it to reach organisational knowledge or allocate work, leaders may experience their organisations through AI-generated interpretations of what’s happening, and customers may deal with AI before, or instead of, another human.
How does she tell whether an organisation has crossed from one stage to the next? “I do not count AI licences,” she says. “I look for evidence in the organisation.” Has work been redesigned, and have roles and decision rights changed? Can the organisation say where human judgement is required? And, she says, crucially: can it demonstrate better business performance as a consequence?
The AIM Transformation Market Tracker, which she says analyses live UK recruitment activity rather than opinion surveys, points the same way. Its strongest signal from the first half of 2026 was a shift from managing organisational change towards designing organisations capable of operating alongside AI. The research also identifies a gap: organisations are investing in operating models, capability and governance but still typically developing them independently. Buying the components of transformation, she says, “is not the same as transforming the organisation.”
The obvious objection is that organisations have absorbed every new technology before.
The Case Against Her
Put it to Tess that ordinary change management has absorbed every new technology for a century, and that naming a new organisational form adds vocabulary rather than capability, and she calls it “a good challenge.” Then she says what she thinks is different. “We have never introduced a new form of intelligence at this scale that can perform aspects of cognitive work, potentially outperform employees in particular tasks, operate continuously and increasingly act in pursuit of a goal.”
This is where her idea of AI as the ultimate Completer Finisher comes in, the subject of a recent article on LinkedIn. In Belbin’s original team-role concept, she says, the Completer Finisher drives towards completion and persists until the job is finished. AI potentially takes that characteristic further: it can pursue an objective relentlessly, and at a speed that can outpace the humans responsible for overseeing it. A human may remain accountable while no longer being the most capable intelligence involved in the work.
“A name in an accountability box does not establish meaningful control.”
Tess, Founder, AI-Mediated Foundation
What she wants to know is who or what influenced the decision, who had authority to act, where human judgement entered, and what would cause the AI to stop. That’s why she distinguishes decision influence, decision authority and decision accountability, which can separate as AI capability increases. The question she now asks is different from change management’s: “How do we design an organisation when the technology is itself becoming an increasingly capable participant in the work?” In her words: “That is not simply new vocabulary. It is a different organisational condition.”
That leaves what a leader should do while none of it is settled.
What Carries Over and What Doesn’t
For a senior leader deciding now, the judgement that transfers most directly is that “technology does not transform an organisation,” Tess says. Her advice is to “not start with the technology. Start with the organisation you are trying to create.” Be clear about the business outcome, understand the work that creates it, then find where human capability adds value, where AI adds value and where the combination produces something better.
Some assumptions from the last thirty years she wouldn’t carry over. The biggest is the assumption that an organisation should be designed around humans as the principal source of intelligence. “Increasingly, that assumption is no longer safe to make.” She’d also question designing a future state and stabilising around it, because AI capability is moving too quickly for a static target operating model to remain “target” for long. And she wouldn’t carry over the assumption that accountability automatically means control.
“We are going to need leaders who are comfortable saying, ‘We don’t yet know.’”
Tess, Founder, AI-Mediated Foundation
The challenge, she says, is not to predict perfectly what the organisation will look like in five years. It’s to build one “capable of making good decisions while the environment is changing.”
That is the thinking she’s carrying into her book, her research and her Foundation.
Where Tess Is Now
The AI-Centric HR Operating Model: An Ethical Approach to Improve Workforce and Business Performance was published by Kogan Page on 3 October 2026. Tess’s work is now extending beyond the AI-centric operating model into a much bigger question: what happens to organisations when intelligence within them is no longer exclusively human?
That question sits at the heart of her research into the AI-Mediated Organisation and the work of the AI-Mediated Foundation. Her focus is not on AI simply as a technology, but on how increasingly capable non-human intelligence changes the way organisations function: how work is designed, how decisions are made, where authority sits, how capability develops, how people experience work, and how organisations maintain governance and accountability as AI becomes more deeply embedded in their operations.
The work is deliberately multidisciplinary. Tess combines more than three decades of experience across public and private sectors, global organisations and academia with extensive research and continuing engagement with leading thinkers and practitioners across AI, business and academia. Her book is grounded in research and evidence, while the AIM Foundation is developing a wider research agenda around the organisational, human and economic consequences of AI. The premise is that questions of this scale cannot be answered by technology, HR, organisational design, economics or any other discipline working independently.
In 2027, Tess will take this work to a wider international audience through speaking engagements, working sessions, and consultation with governments and organisations. The purpose is not simply to explain what AI can do, but to help leaders and policymakers understand what its growing capability means for organisations, people and economies. Her work encompasses leadership, operating models, workforce and organisational capability, judgement, decision authority, governance, AI economics, accountability and the evolving relationship between human and artificial intelligence.
The PXB Ecosystem™ provides the organisational architecture through which much of this thinking can be translated into practice. People, Experience and Business form its core. Around them sit interconnected components including the Strategic AI Pillars, WorkScope, the Judgement Spine, BALANCE, BREATHES and the Resources Hub. Their purpose is not to create another collection of AI frameworks, Tess says, but to connect consequences that organisations often address separately. “The point is that the consequences are connected.”
PXB enables organisations to examine where AI is entering their operating model and trace what changes as a consequence across work, people, capability, experience, decisions, governance and business performance. It provides an architecture for redesigning how human and artificial intelligence operate together rather than simply adding AI to an organisational model designed when human intelligence was the principal source of intelligence within the organisation.
For boards, that redesign also has an economic dimension. Tess’s starting point is that leaders need to understand not only where AI is entering the organisation, but what is being invested, where capacity is being created, whether that capacity is translating into measurable productivity and business performance, how workforce costs and roles are changing, and what additional costs are being created through verification, governance and oversight. Reducing the cost or time required to perform an individual task does not automatically improve the economics of the organisation as a whole. As she says earlier in the article, if AI reduces a two-hour task to twenty minutes, the organisation has created capacity. It has not automatically created productivity.
These questions increasingly extend beyond individual organisations. Governments also need to understand the cumulative consequences for employment, skills, productivity, investment, competitiveness and economic growth, including the potential economic cost when the pace at which AI changes or removes work is not matched by the creation of new productive capacity and value.
Tess does not see the answers as belonging to one person, one profession or one discipline. The work continues to evolve through research, organisational evidence and collaboration with people working at the leading edge of AI, organisations and academia. The AI-Mediated Foundation provides a platform for bringing those perspectives together as the relationship between human and artificial intelligence continues to change.
“No one discipline owns the answer to what happens when intelligence within an organisation is no longer exclusively human.”
Six Reflections from Tess
Where is AI entering your organisation? Map it across workflows, jobs, systems, customers, knowledge, decisions and actions.
What are the economics? Understand the total investment in AI, the capacity being created, the costs of verification, governance and oversight, the impact on workforce costs, and whether any of this is translating into measurable productivity, performance and value.
Where did the saved capacity go? Time saved is capacity, not automatically productivity. Ask what happened to that capacity and whether the work around it was redesigned to convert the saving into organisational value.
Where does human judgement enter? Ask who or what influenced the decision, where human judgement was exercised, who had authority to act and who remains accountable.
What would make the AI stop? Can the organisation identify the limits of AI authority and reconstruct the chain from objective and influence through decision and action to consequence and accountability?
What remains fundamentally true? Keep what holds, recognise what AI has made newly possible, and redesign the space between the two.
From the Diary of Tess Hilson-Greener
Tess Hilson-Greener is a UK-based author and researcher and the Founder of the AI-Mediated Foundation.
| Role | Founder, AI-Mediated Foundation |
| Based in | United Kingdom |
| Previously worked with | Inmarsat, Deloitte, Office for National Statistics, Deutsche Bank, Santander, Astellas Pharma, Gilead Sciences, BMW/Rolls-Royce and other global organisations |
“A name in an accountability box does not establish meaningful control.”
Editor’s Note
Executives Diary chose Tess Hilson-Greener because her work starts beneath the usual conversation about AI tools, with how an organisation is designed when intelligence within it is no longer exclusively human. She brings more than 30 operating models and three decades of introducing technology into organisations to that question, and a book published by Kogan Page on 3 October 2026. The editors were persuaded by how clearly she separates what AI changes from what a board remains responsible for.
Executives Diary, Editorial
Sources
Microsoft, “Microsoft and LinkedIn release the 2024 Work Trend Index on the state of AI at work,” 2024. https://blogs.microsoft.com/blog/2024/05/08/microsoft-and-linkedin-release-the-2024-work-trend-index-on-the-state-of-ai-at-work/
