Design The System: AI, Agile, and the Future of Work

AI is moving from a simple productivity tool to becoming part of the operating system of work itself. The episode explores how Agile made work visible, remote work made it digital, and now AI is beginning to interact with, automate, and execute those workflows.

The discussion highlights a broader shift in consulting, organisational design, and knowledge work: value is moving away from manpower and toward orchestration, judgment, systems thinking, governance, and operational redesign. It also explains how developments like Claude Skills and Anthropic’s MCP could turn organisational knowledge into reusable, executable workflows.

Find us here: www.agileleanireland.org

Transcript

“Design The System: AI, Agile, and the Future of Work”

Welcome back.

Over the last few months at ALI, we’ve been exploring a huge question:

What actually happens when AI stops being a tool sitting beside work… and starts becoming part of the operating system of work itself?

Because that feels like where we are heading.

And honestly, a lot of the recent AI news stories all seem disconnected on the surface:
 consulting firms buying AI companies,
 AI agents entering workflows,
 local models emerging,
 AI standards forming,
 traditional jobs changing,
 new organisational structures appearing.

But underneath all of it, there is a much bigger pattern emerging.

The structure of work itself is changing.

For years, Agile made work visible.

That was one of the biggest shifts in modern organisations.

User stories.
 Sprint boards.
 Stand-ups.
 Retrospectives.
 Backlogs.

Suddenly work became structured, traceable, measurable, and collaborative.

Then remote work accelerated something else.

It made work digital.

Conversations moved into Slack.
 Documentation moved into shared systems.
 Meetings became recorded.
 Processes became observable.

In many ways, remote work made organisational behaviour machine-readable.

And now AI arrives on top of that.

Which means the same workflows Agile and digital transformation made visible… AI can increasingly interact with, analyse, automate, augment, and eventually execute.

That is why we keep saying this is not just a technology shift.

It is a systems shift.

And honestly, a lot of organisations still seem to be thinking about AI as if it is just another productivity tool.

But the more we experiment, the clearer it becomes:
 AI is starting to become operational infrastructure.

Take consulting.

Over the last few weeks we’ve seen IBM, Accenture, McKinsey, OpenAI, Anthropic, and others all making major moves around AI transformation.

At first glance, it looks like firms simply expanding AI services.

But we think something deeper is happening.

Traditional consulting scaled through manpower.
 Large teams.
 Hierarchy.
 Delivery structures.
 Repeatable methodologies.
 Specialist silos.

But AI changes the economics of knowledge work itself.

Research.
 Analysis.
 Documentation.
 Reporting.
 Workflow coordination.
 Code generation.
 Operational support.

All increasingly becoming AI-assisted.

So the value shifts.

Less in manpower.

More in orchestration.
 Judgment.
 Systems thinking.
 Governance.
 Integration.
 Operational redesign.

Which raises a fascinating question:

What happens when organisations no longer scale primarily through people… but through capability?

Because that feels like where this is going.

And that shift becomes even clearer when you look at things like Claude Skills.

We recently spent a lot of time experimenting with Claude Skills at ALI.

And honestly, they may represent one of the most important practical AI developments so far.

Because they move AI away from being a generic chatbot and toward becoming a reusable operational specialist.

Instead of constantly re-prompting:
 “write in this tone”
 “avoid jargon”
 “follow this structure”
 “generate tests first”

…those behaviours become embedded.

Reusable.
 Consistent.
 Shareable.

And what became really interesting to us was this:

Skills start to resemble an evolution of SOPs.

Traditional SOPs documented how work should happen.

AI Skills begin operationalising that knowledge directly into executable workflows and AI behaviour.

That feels significant.

Because now organisational knowledge itself becomes scalable in a completely different way.

And then we extended those experiments into MCP — Anthropic’s Model Context Protocol.

This may sound technical, but the implications are massive.

Because MCP is essentially becoming the plumbing layer for AI-native organisations.

It allows AI systems to securely connect into real enterprise environments:
 tools,
 systems,
 data,
 documentation,
 CRMs,
 knowledge bases,
 workflows,
 operational processes.

That changes everything.

Because once AI can reliably interact with operational systems, it stops behaving like a chatbot sitting outside the organisation.

It starts becoming part of the organisation itself.

Part of the workflow.
 Part of the operating model.
 Part of decision-making.

And that is why we think the real AI race may not ultimately be about who has the smartest model.

It may be about who designs the best systems around those models.

Who orchestrates best.
 Who governs best.
 Who operationalises best.

And we think this also explains why AIQ matters so much.

Because as AI becomes more conversational, fluent, persuasive, and embedded into workflows, humans may increasingly mistake fluency for judgment.

That is dangerous.

The future challenge may not be whether AI becomes conscious.

It may be whether humans start behaving as if it is.

Whether we defer to it too quickly.
 Trust it too easily.
 Outsource judgment too casually.

And this is why leadership becomes more important, not less.

Governance becomes more important, not less.

Human judgment becomes more important, not less.

Because in AI-native organisations, the challenge is no longer simply:
 “How do we use AI?”

It becomes:
 “How do we design systems where humans and AI collaborate intelligently, safely, adaptively, and responsibly?”

And honestly, we think this is where Agile becomes surprisingly relevant again.

Not because Agile is about ceremonies.

But because Agile was always fundamentally about adaptation, visibility, learning, feedback loops, collaboration, and continuous improvement.

Those capabilities become even more important in AI-native systems.

Because AI accelerates change.

And when change accelerates, organisational adaptability becomes survival infrastructure.

That is why we keep returning to the same core idea at ALI:

Design the system.
 Don’t just automate it.

Because automation without redesign simply scales old problems faster.

The organisations that succeed in this next phase may not be the ones deploying the most AI.

They may be the ones best able to redesign workflows, leadership, governance, operating models, and human + AI collaboration systems intentionally.

And honestly, it still feels like most organisations are only at the beginning of understanding how structural this shift may become.

We are not just watching new tools emerge.

We may be watching the early formation of AI-native organisations themselves.

And that changes everything.