---

Why Councils Struggle to Digitise Physical Assets? How AI Actually Helps.

The Gap Between Ambition and Reality

Almost every council I’ve worked with wants to “go digital.” They want digital twins, predictive maintenance, and clearer insight into what’s happening across their networks. But when you actually step into the operational environment—whether it’s pavements, pipes, tree roots, moisture levels or traffic sensors—you quickly find the same underlying issue: the council often doesn’t have a clean, trusted, or complete record of what assets exist in the first place.

This isn’t due to a lack of effort. It’s because councils inherited decades of fragmented systems, inconsistent datasets, and organisational structures that were never designed for modern digital asset management.

---

Councils Are Data-Poor Where It Matters Most

There’s a common assumption that councils have well-structured geospatial and asset databases. In reality, most have a mix of outdated GIS layers, inconsistent condition ratings, and partially digitised legacy records. Different teams collect information differently, and historical data is often unreliable enough that staff hesitate to use it for decision-making.

This creates a major barrier for AI. Models don’t fail because they’re “not powerful enough.” They fail because:

• datasets are incomplete,
• asset definitions vary between teams, and
• the input data isn’t standardised or trustworthy.

Digitisation ends up being a data quality and governance challenge long before it becomes a technology challenge.

---

The Silo Problem: Too Many Teams, Too Many Systems

Councils typically divide responsibilities across transport, water, parks, stormwater, planning, GIS, contracting teams, and external regulators. Each unit has its own workflows, systems, and priorities.

Digitising physical assets requires these groups to operate as a single integrated system. But councils are not structured that way. The result is:

• duplicated work,
• inconsistent asset definitions,
• incompatible datasets, and
• no clear “owner” of digital infrastructure.

Even when one team is enthusiastic about AI, the surrounding workflow can block it from being effective.

---

The Hidden Challenge: Ground Truthing

In the 18-month pavement AI trial I worked on, the most resource-intensive part wasn’t the modelling. It was the validation: ensuring field data, labelled samples, and condition assessments were consistent across staff and contractors.

Councils often underestimate how much labour, coordination, and discipline is required to build high-quality ground truth. Without this foundation, even the best digital twin becomes a polished visualisation rather than a tool for real decision-making.

---

Vendors Often Oversell What AI Can Do

Councils are regularly pitched AI systems claiming to “automatically detect everything.” Usually the models were trained overseas, don’t reflect local conditions, and rely heavily on post-processing by contractors. Integration with council systems is afterthought at best.

The result is that councils buy tools that look impressive in demos but create more manual checking and data-cleaning work than expected. The problem is not incompetence; it’s the mismatch between vendor promises and operational reality.

---

Where AI Actually Adds Value

Despite the challenges, I’m more optimistic today than I’ve ever been. When used properly, AI doesn’t replace inspectors, engineers, or asset managers. Its real value lies in consistency, scalability, and network-wide insight.

1. AI speeds up asset capture

A single drive or drone flight can capture data equivalent to weeks or months of manual fieldwork.

2. It provides a consistent baseline

AI applies the same rules across the entire network—something impossible to achieve manually.

3. It reveals patterns hidden in the data

Deterioration, environmental influences, risk clusters—AI surfaces these quickly and objectively.

4. It enables proactive, cost-neutral maintenance

Instead of reacting to failures, councils can plan interventions earlier and more strategically.

5. It supports integration of above- and below-ground assets

This is where the next decade of council infrastructure management is heading:
roads, pipes, trees, stormwater flow, moisture profiles, traffic volumes—all feeding into a single AI-supported digital ecosystem.

---

Why This Is Finally Possible

Five years ago, building a system like this required bespoke tools, specialised staff, and very large budgets. Today:

• sensors are affordable,
• models are pre-trained and adaptable,
• cloud infrastructure is cheap, and
• councils can run targeted pilots without major procurement cycles.

The technology has matured. The need has become urgent. And the distance between those two points is finally closing.

---

The Real Issue Isn’t Vision — It’s Infrastructure

Councils aren’t struggling because they lack ambition or capability. They’re struggling because they are expected to manage complex, growing infrastructure with limited resources, siloed workflows, and data systems that were never built for modern AI-driven asset management.

AI provides something councils have never had before:
a consistent, objective, scalable view of the entire asset network.

Once that foundation is in place, everything else—planning, budgeting, resilience, compliance—becomes dramatically easier.

Digitisation is no longer optional. And for the first time, AI makes it realistically achievable.

---