Part 5 of 5 — Series Finale  |  Previous: From Copilot to Agent: why the shift to Agentic AI demands a different kind of governance

AI won’t take your job. Someone using AI will. But the deeper truth — the one that runs through every blog in this series — is that the organisations who understand this, and build their teams accordingly, will leave everyone else behind.

This is the fifth and final blog in Brooklyn’s AI readiness series. We started with Processes. We moved through Transformation Approach, Data Readiness, and Agentic AI governance. Every one of those parts had a people dimension running through it. It was always pointing here. Because people are the most important and most consistently underfunded part of any AI transformation.

The Capability Gap is Already Here

Deloitte’s 2025 Global CPO Survey found that 57% of Chief Procurement Officers cite siloed working as their top barrier to value delivery [1]. The people who own the process are separated from the people who own the data, who are separated from the people who run the technology. In that model, AI doesn’t fix the problem. It makes it worse — faster.

The organisations navigating this well have broken down those silos through small, aligned, multi-skilled teams that bring process understanding, data ownership, technology capability, and change thinking into the same room. Amazon’s two-pizza teams. Agile squads. Change-and-run merged. Whatever you call it, the principle is the same: people with overlapping skills working toward a common goal, with the authority and accountability to make decisions rather than escalate them.

AI-First Doesn’t Mean Fewer People

One of the most persistent misconceptions in enterprise AI strategy is that AI adoption is primarily a headcount story. It isn’t. And procurement is a clear case in point.

Think about due diligence. Today, most procurement organisations do a thorough job on a handful of strategic suppliers and a checkbox exercise on everything below that. Not because people don’t want to do it better — but because they don’t have the capacity. Ask any procurement lead whether they’re doing everything they’d want to do, at the standard they’d want to do it. The answer is always no. There are two or three times more things they want to do than they have people to do them.

AI changes what’s possible at scale. Automated due diligence on every supplier, not just the top tier. Continuous obligation monitoring rather than periodic reviews. Real-time risk alerting rather than quarterly reports. That is not a fewer-people story. That is a do-more-with-the-same story covering more ground, at higher standards, freeing human expertise for the work that requires human judgement.

The right framing: An AI-first procurement team is not a smaller team. It is a team doing different things, covering ground that was previously impossible to cover, operating at a standard that was previously impossible to maintain, and focusing human attention on the decisions, relationships, and exceptions that AI cannot handle.

What the AI-Ready Procurement Team Actually Looks Like

The roles that matter in an AI-first procurement team are not the same as the roles that matter today. All of them share a common thread: they require people who can work across disciplines, govern technology as well as use it, and think in terms of processes and outcomes rather than tasks.

Process Owners: People who map, govern, and continuously improve the end-to-end processes that AI executes on. They define what good looks like at each step, own the escalation paths, and are accountable for the outputs of both the human and the agent in their process area.

Data Stewards: Named owners of specific data domains — supplier master data, contract data, spend classifications — responsible for quality, completeness, and consistency. As argued in Part 3, data quality is a discipline requiring someone accountable for maintaining it every day.

Agent Owners / AI Governance Leads: The people who are accountable for what the agents do — who review outputs, tune prompts, monitor escalation patterns, and ensure agent behaviour remains within defined boundaries. In the RACI model, the agent is Responsible. A human is always Accountable. This role is that human, at scale.

Supplier Relationship Specialists: The work AI cannot do: human-to-human interaction, relationship depth, negotiation, trust. As AI handles more transactional and monitoring work, human value in supplier management concentrates in relationship quality. These roles become more important, not less.

The Expert Generalist: The role that cuts across all of the above. Someone who spans process, data, technology, and commercial understanding. What Brooklyn has done with the Product Quality Engineer — consolidating product management, QA, and design into a single role supported by AI tools — is a live example. In procurement, the equivalent consolidation is already visible: contract manager, supplier relationship manager, and supplier risk manager are roles that have proliferated to the point of fragmentation, and that AI now makes collapsible into a broader, single remit.

The Expert Generalist Imperative

The T-shaped professional — broad across multiple disciplines, deep in one or two — has been the aspirational hiring model in technology for years. What AI has done is make it the necessary model everywhere. The technical/non-technical divide is dissolving. In the same way everyone became a people manager and everyone became a marketer in the modern organisation, everyone is becoming a technology practitioner. And in the age of Agentic AI, everyone is becoming an agent manager.

Everyone is Now a Manager — of Knowledge and Capability

Across three roles, AI can now take on 10–30% of the tasks that were previously performed manually by each person. Those tasks don’t disappear — they are broken off and assigned to a governed AI Agent that you define, train, tune, and review. That agent becomes a member of the team — a digital colleague with a defined remit, a governance framework, and a line of accountability back to a named human.

Managing that agent looks remarkably similar to managing a junior team member. You onboard it with clear instructions. You review its work periodically. You identify where it’s going wrong and correct it — not by telling it off, but by refining the prompt and the data it works with. You define what it must do, and equally importantly what it must not do. You give it guardrails. You build an opt-out path for when it hits the edge of its competence and a human needs to take over.

Carbon and silicon: The AI-ready organisation has both carbon-based and silicon-based knowledge working together. The human team owns the judgement, the relationships, the exceptions, and the governance. The agent team handles the volume, the monitoring, the data consolidation, and the pattern recognition. Neither replaces the other. The old framework — people, process, technology, governance — has a new layer: the AI entity, governed like a team member, accountable to a human, auditable at every step.

Hire for AI Readiness — Starting Now

Every role you hire for from this point forward should include an explicit AI dimension. Not an AI specialist role — every role. What AI tools does this person use? How have they applied AI in their previous work? How do they think about AI augmenting their own capability?

This is not about screening for AI expertise. It is about screening for AI curiosity and adaptability — the mindset that treats AI as a capability to develop rather than a threat to resist. The organisations building this into their hiring criteria now will have a compounding advantage over the next three to five years that will be very difficult to close later. The person who loses their job to AI is not replaced by a machine. They are outcompeted by a colleague who understood how to use machines better than they did.

The Run Model is Dead. Long Live Continuous Improvement.

Run organisations are still being staffed in many enterprises as if they are pure operators — lean to the point where there is no capacity to invest, iterate, or improve. The software has moved on entirely. The platforms update continuously. The AI capabilities evolve weekly. But the team structure is still built for a world where you deployed a version, ran it for two years, and did a big programme to move to the next one.

That model is self-harming. Organisations that hold back updates to ‘stabilise on a version’ end up, often unintentionally, on a branch they can’t get back from. The people capability you need for an AI-first operating model is not a change team and a run team. It is a team that does both, continuously — tuning agents, refining prompts, improving data quality, spotting escalation patterns, feeding learnings back into process design. The change/run distinction, in the old sense, is over.

Closing the Loop: the AI Readiness Framework in Full

The Brooklyn AI Readiness Series — five parts, one framework:

1. Understand your processes. Map end-to-end. Define inputs, outputs, and what good looks like at every step. Without this, nothing else works.
2. Fix your transformation approach. Start with an MVP. Establish design authority. Invest in capability, not just technology. Don’t try to boil the ocean.
3. Govern your data. Start with your cleanest data. Define minimum viable standards. Establish systems of record. Data quality is a discipline, not a project.
4. Deploy agents with the right governance. Human-in-the-loop by design. Full observability. Named accountability. EU AI Act compliance built in, not bolted on.
5. Build the team that makes it sustainable. Expert generalists. AI literacy as a hiring baseline. Agent management as a core competency. Continuous improvement as the operating model.

None of these five parts works without the others. That is the point of writing them in sequence. You cannot govern an agent on a process you haven’t mapped. You cannot trust data you haven’t stewarded. You cannot build a transformation on a team that has been stripped of the capacity to change. And you cannot sustain any of it without the people who have both the capability and the culture to make continuous improvement their default mode.

The organisations that will win in the AI era are not the ones that bought the best tools. They are the ones that did the foundational work — on process, on data, on governance, and above all on people — before they expected the tools to deliver.

About the Author

Nick Francis, Chief Technology and Marketing Officer at Brooklyn Solutions

Nick Francis is a well-established and experienced CxO delivering Digital & Security-focused Transformation through the design, build, and deployment of cost-effective, highly automated industry-leading solutions. Nick has experience working across the private and public sectors in industries such as Financial Services, Insurance, Legal, Utilities, Retail, Public Sector and Government. Specialised in transformation activity to optimise processes, operational expenditure, and increase productivity. Significant experience in compliance, risk & control activities in highly regulated industries, standardisation of technologies, streamlining of internal processes and continuous improvement driving consistency and efficiency across an organisation whilst holding Customer, Colleague and Partner experience at a premium.

References

[1] Deloitte Global CPO Survey 2025: 57% of Chief Procurement Officers cite siloed working as their top barrier to value delivery.

[2] Kienbaum 2025 study on organisational transformative capacity: under 2% of transformation budgets typically allocated to capability building and reskilling.

[3] Gartner (2025–2026): over 40% of agentic AI projects expected to be cancelled by 2027, primarily due to governance gaps and people capability shortfalls.

[4] Professor Scott Galloway, NYU Stern — “AI is not going to take your job. Someone who understands AI is going to take your job.” Masters of Scale podcast (October 2024), Diary of a CEO with Steven Bartlett, and OMR 2024 conference.

[5] Martin Fowler, Unmesh Joshi & Gitanjali Venkatraman — “Expert Generalists”, martinfowler.com (July 2025). Defines the expert generalist as someone who can “dissect unfamiliar challenges, spot first-principles patterns, and make confident design decisions with the assurance of a specialist” while spanning multiple domains.