Welcome
Welcome back to The Project Guy, the newsletter that won't ask you to "circle back" on anything. Ever.
This month we’re tackling how AI is reshaping the way we structure project delivery teams. It’s a topic I’ve been chewing on for a while, and I wanted to do it justice rather than skim the surface.
Grab a coffee and settle in..
How AI Is Reshaping the Shape of Project Delivery Teams
AI is bringing about a fundamental shift in how project teams structure themselves to deliver work. The classic project delivery pyramid — wide base of analysts and doers, a middle band of specialists and technical practitioners, a thin slice for project managers, and a small apex of senior leaders — was built around one core reality: detailed analysis, documentation, and coordination required extensive human effort.
That reality no longer holds.
AI tools like Claude and ChatGPT can digest 300-page requirements documents and produce comprehensive analysis before your coffee cools. Discovery phases that once required three analysts working for six weeks now take a single AI-literate facilitator six hours. The broad base of the project delivery pyramid — analysts, coordinators, junior project support — is being hollowed out not by cost-cutting, but by capability displacement.
The future of project delivery isn’t a flatter pyramid. It’s a diamond. A small layer of AI-literate facilitators at the base, a strong and experienced expert middle, and a focused apex of senior leaders who own outcomes.
The shift is already underway — lets dive into what it means for how you structure your teams in the future.
The Pyramid: A Structure Built for Its Time
The pyramid model of project delivery emerged in the late twentieth century as a rational response to the complexity of managing Projects.
It succeeded because it created clear reporting lines and scaled proportionally: few decision-makers at the top for speed and coherence, many hands at the base for execution. Decisions flowed down, work and insight flowed up, and everyone knew precisely where they fitted in.
For consultancies, the pyramid also created significant commercial leverage. An army of junior analysts carrying out documentation and discovery work justified high project fees, and clients had little alternative: if the work needed doing, it required the people to do it. The pyramid model wasn’t just operationally logical — it was commercially powerful.
Two conditions made this structure work:
• Detailed analysis and documentation required significant human effort.
• Clients and organisations were willing to pay for that human effort.
Both conditions are now being fundamentally disrupted.
Why the Pyramid Is Transforming
Generative AI can now record and process stakeholder interviews, generate project documentation and risk registers, produce analysis and status summaries, and flag issues in real time — in a fraction of the time and at a fraction of the cost of human effort. The work that once filled the base of the pyramid has not disappeared; it has simply become automatable.
This is where it’s worth addressing the Gartner prediction that has circulated widely in project management discussions: that "by 2030, 80% of project management tasks will be run by AI".
Here’s the nuance that often gets lost in the headline though.
Not all roles are being reshaped equally. The experienced project manager’s day-to-day reality — building stakeholder trust, navigating organisational politics, making judgement calls under uncertainty, delivering difficult news with credibility — sits firmly in the people space.
These aren’t tasks you can hand to an algorithm. They require emotional intelligence, context, and the kind of relationship capital that only comes from showing up consistently and earning it over time.
Where Gartner is absolutely being proved right, however, is at the base of the pyramid. The analytical, coordinative, and administrative work historically carried out by junior project staff — the layer that has justified large team headcounts for decades — is being eroded by automation at pace.
Data collection, status reporting, schedule tracking, risk register maintenance. These are precisely the tasks Gartner was pointing at. And clients are no longer prepared to pay for slower, costlier human effort when AI can produce comparable outputs in hours rather than weeks.
So the prediction isn’t wrong. It’s just landing exactly where you’d expect it to — at the task level, not the leadership level.
The Diamond: A New Structure for Project Delivery
As the base of the pyramid erodes, the emerging structure is not a flatter version of what came before. It is a fundamentally different shape: a diamond. Every role is reshaped in some way by the evolution of AI.
The diamond model retains the apex of senior leaders and sponsors who own outcomes and provide strategic direction. But even here, AI provides unprecedented visibility and analytics for decision-making.
Project managers in the diamond model shift their role from tactical coordination to strategic orchestration. With AI monitoring plan progress, flagging risks, and tracking resource allocation in real time, the project manager’s attention moves upstream: to stakeholder alignment, decision-making frameworks, and ensuring the team is focused on the right problems.
The middle of the diamond consists of experienced project architects, subject matter experts, and technical specialists that form a robust expert layer that is now the engine of delivery. AI handles routine analysis and documentation; these practitioners focus on what AI cannot do — synthesis, judgement, client relationships, and complex problem-solving. They are brought in as needed rather than maintained on permanent headcount.
At the base of the diamond, instead of a large layer of junior support staff, a small number of AI-literate facilitators manage the technology stack — orchestrating the tools, prompting the analysis, reviewing outputs, and ensuring quality. They replace the broad base not by doing less, but by doing it differently.
The Benefits of the Diamond Model
The diamond model offers substantive advantages over its predecessor — but only when it is implemented deliberately.
• Lower cost of delivery. AI consistently handles a significant proportion of routine project work, reducing the headcount required without reducing output quality.
• Faster execution. Discovery, analysis, and documentation phases that previously took weeks can be completed in days or hours, compressing timelines without sacrificing depth.
• Scalable delivery capability. Teams can flex their expert middle up or down based on project complexity, without the overhead of a large permanent base layer.
• Aligned with stakeholder expectations. Clients increasingly expect small expert teams with measurable impact. The diamond delivers this — and with transparent, auditable AI-generated outputs, quality is easier to demonstrate.
• Human effort focused where it matters. By automating the routine, the diamond frees experienced practitioners for the relational, judgement-intensive work that determines whether a project succeeds or fails.
The Risks to Manage
The diamond model is not without its challenges. Three risks warrant particular attention.
AI Output Quality
AI-generated outputs still require human review. Hallucinations — plausible-sounding but factually incorrect outputs — remain a genuine risk, particularly in documentation-heavy phases. The AI-literate facilitator layer exists precisely to manage this: reviewing, editing, and validating outputs before they are used. This is not a reason to avoid AI; it is a reason to deploy it thoughtfully.
The Career Pipeline Problem
The traditional pyramid created a clear career pathway: junior analysts learned their craft by doing, moving progressively into more senior roles. The diamond disrupts this. If the entry-level roles that once formed the learning ground are replaced by AI, organisations need to think carefully about how the next generation of experienced practitioners develops. New models for early-career development — structured exposure to AI-augmented delivery, mentoring within the expert middle, deliberate knowledge transfer — will be necessary to prevent a capability gap a decade from now.
Re-Bloating the Base
Perhaps the most insidious risk is organisational inertia. Leaders accustomed to the pyramid model may unconsciously revert to its logic — hiring too many mid-level roles, allowing the base to creep back, and effectively rebuilding the structure they were trying to move away from. The diamond requires active, deliberate governance to maintain its shape. Without that governance, it drifts back toward the pyramid.
This Is Already Happening
The diamond model is not a prediction. It is a description of what the most forward-thinking project organisations are already doing. Teams are delivering faster with fewer people. Consultancies are restructuring their delivery models. Clients are scrutinising headcount on project proposals with a sharper eye than they were two years ago.
Project delivery teams that cling to the pyramid are not preserving quality or protecting their people. They are paying for headcount that no longer justifies its cost and missing a shift that clients are already demanding.
The strongest project teams of the next five years will not be the biggest.
They will be the most deliberately designed: a tight expert middle, AI-literate facilitators managing the technology stack, and senior leaders with genuine strategic oversight. That is the team that hits milestones faster, earns stakeholder trust, and scales without chaos.
Until next time,
Guy Thorpe - aka The Project Guy