Connected Britain returned to London this year, bringing around 8,000 people from operators, altnets, government and the supply chain into one building. The agenda covered faster full fibre and 5G rollouts, the digital skills gap and the role of AI in running networks. Each topic came back to the same question:

How quickly can the work on site keep pace with the plans?

The plans are moving fast. Ofcom's Spring 2026 update found full fibre available to 24.9 million UK homes, 82% of the total, as of January 2026. That is 1.2 million more homes in six months, or roughly 46,000 a week. If every planned build goes ahead, Ofcom expects full fibre to reach 92% of homes by the end of 2028. On mobile, VodafoneThree is investing £11 billion to take 5G Standalone to 99% of the population by 2030 and 99.96% by 2034.

Every one of those homes and masts depends on a person on site completing a survey, an install, an inspection or a handover. 

We hosted a session with Francesca Serravalle, Head of Network Monetisation Strategy at VodafoneThree, and Kapil Singhal, CEO and Co-founder of Vyntelligence, to look at what happens when AI is built around that work. Kapil set out where physical work AI is heading, and Francesca brought the view from inside one of the UK's largest network upgrades. We also launched natural-language agent building for telecoms.

Why the workforce can't keep pace with the build

Telecoms and utilities have more work than people to do it. The Digital Connectivity Forum's September 2026 report, Expanding the Network, found that around 60% of UK telecoms engineers are over 50 and estimated a 30,000-person engineering shortfall over the next ten years. Third-party estimates reviewed in the report put the number of engineers leaving the wider industry by 2030 at up to 200,000, against a graduate pipeline of around 14,000 a year.

Some of those exits reflect the full fibre build cycle maturing. The knowledge leaving with experienced engineers is the harder loss. The engineers nearing retirement are the people who know the network best: where faults occur, what an issue looks like and what needs to happen next.

For an industry facing a skills shortage, AI that works alongside field teams can keep that knowledge in the business and give engineers back time for the work that needs them.

The rise of physical work AI

Gartner named Physical AI a top strategic technology trend for 2026: intelligence that senses, decides and acts in the real world.

Physical work AI brings that intelligence to the people who build, inspect and maintain infrastructure. It captures evidence on site, understands it in the context of the asset and the job, and triggers the action that follows.

Vyntelligence delivers it through Agentic Video Intelligence, starting with the most widely deployed sensor in the field: the engineer's phone.

Field teams send a short guided video in place of a form, and the AI returns asset health, site intelligence, safety and quality assurance from a single clip. Every video also captures what experienced engineers know, so that knowledge stays with the business.

For Kapil, AI offers "a once-in-a-lifetime opportunity to rethink how we do our work." Francesca sees operators moving the same way, towards a data-driven network, with field data feeding an AI platform that drives outcomes. Asked whether organisations are ready for that shift, she said:

“Organisations are ready. We can already see the change taking place."

Francesca Serravalle, Head of Network Monetisation Strategy, VodafoneThree

Already live across Britain's networks

Physical work AI runs through every layer of Britain's connectivity:

Across telecom deployments, Vyn customers save 65% of pre-install visits, complete handovers three times faster and see 97.5% fewer snags. Each of those gains gives time back to field teams.

What's next: AI agents built by the people who know the work

Expectations of AI keep rising. Early AI was asked to identify what was on screen. Today, customers expect it to decide what happens next and act on it.

At Connected Britain, Vyntelligence launched natural-language agent building for telecoms, following its utilities launch earlier this year. It puts the power to build AI agents in the hands of operations teams.

An operations lead can describe a fault, the rules that decide whether it's "monitor", "48-hour repair" or "emergency", and what should happen next, from approval to dispatch. The system builds the agent. Checks and actions that teams handled by hand six months ago now run on their own.

Experienced engineers will retire. Their expertise can now carry through every job.

Four questions to test if your field work can keep pace

Four questions can help operators test where they stand:

1. When your most experienced engineers retire, how much of what they know stays with the business?

2. How old is your field data by the time it reaches a system?

3. Would your field evidence hold up for a customer, a regulator or an invoice?

4. Does your AI trigger the next step, or wait for someone to do it?

Building a truly connected Britain depends on the work done on site. The operators moving fastest are simplifying that work, capturing what their engineers know, and giving AI the context to act on it.

Find out how physical work AI can help your teams keep pace