AI maturity in the installation industry

A sector that works at two speeds

The installation industry consists of two worlds that rarely ask the same question. On one side is the physical work: technicians laying pipes, connecting equipment, resolving faults on site. On the other side is the organisation around it: planning, work preparation, administration, customer contact, inventory management. The former cannot be automated; the latter can, to varying degrees. That ratio — a great deal of manual work, with an organisational shell that processes information — determines where AI maturity can grow in this sector and where the boundary lies that software cannot resolve.

The shell is larger than is often assumed. Work orders, time registration, material orders, technician scheduling, communication with clients and subcontractors: these are processes that largely consist of data and decision rules. Whether an organisation can do something with that does not depend on the will to innovate, but on what is already in place. A company that plans using a whiteboard and phone calls has a different starting point than a company with a planning system that already collects data on turnaround times and availability.

Why pilots often get stuck here

Installation companies that try AI often start with the visible: a chatbot for customer questions, a tool that helps draft quotations. That works temporarily, until it turns out the underlying data is not consistent. Material codes that differ per branch, time registration that is partly done on paper, customer data that is named differently in three systems — that is not an AI problem, it is a data management problem that precedes AI.

The organisational dimension also plays a stronger role here than in office work. Installation companies often work with a fixed core and a fluctuating group of subcontractors or temporary staff. Who decides on a new system, who maintains it, and whether technicians enter data themselves or leave that to the back office — those questions determine whether an AI application is embraced or ends up standing alongside the work rather than within it. A system chosen by the back office but meant to be used by technicians in the field often gets stuck at that transition.

IT infrastructure as bottleneck and opportunity

The physical nature of the work brings with it a specific IT requirement: mobile access, offline working at locations without a connection, linking fieldwork to office systems. Many installation companies have partly built up this infrastructure — out of necessity, not from an AI strategy — but not always in a way that allows systems to talk to each other. A planning tool and an invoicing system that exist side by side without shared data do not provide a foundation that AI can add anything to.

This is where this sector distinguishes itself from a sector such as healthcare, where AI maturity in healthcare is mainly held back by regulation and accountability, or from wholesale, where AI maturity in wholesale revolves more around inventory and order data than mobile infrastructure. In the installation industry, the question is rather: does one coherent picture of a job exist, from first contact to completion, or does that information fall apart into separate systems that nobody connects.

What the fundamental dimensions mean here

The maturity measurement from hybridresourcing looks at seven dimensions, of which organisation, IT infrastructure and data management precede the rest. In the installation industry, that is a logical order, not a choice. Without unambiguous data on materials, hours and customers, no AI application has anything reliable to build on. Without infrastructure that connects fieldwork and office, every application remains limited to one of the two worlds. And without an organisation that makes clear who is responsible for systems used by both permanent staff and flexible workers, every initiative gets bogged down in unclarity about ownership.

This industry partly shares that pattern with manufacturing and the transport sector. In AI maturity in manufacturing, the combination of physical process and digital shell also plays a role, and in AI maturity in the transport sector, mobility and dispersed work locations are a comparable issue. Anyone working in the installation industry will therefore recognise part of their situation in other sectors with a great deal of work outside the office, but the specific combination of permits, safety regulations and material knowledge makes the data management dimension weigh more heavily here than elsewhere.

Because a plot round has multiple people score separately — a director, a planner, a technician, an IT manager — it becomes visible whether these pictures align. In a sector with such a difference between office and field, that spread is often larger than expected, and that spread itself is informative: it shows where an organisation is unanimous about its readiness and where it is not.

When capacity has been established

This page describes whether an organisation can bear AI — whether the fundamentals of organisation, infrastructure and data are in order. That is a different question from which part of the work AI could actually take over. For that question, the work scan from FTE TO AI calculates per task which part of the work qualifies for takeover, based on what the work precisely involves and not on the readiness of the organisation behind it. Anyone who wants to know where the installation industry stands would do well to keep both questions separate: first whether the foundation is in place, then which work can actually shift.