BT Group announced the target of eliminating 55,000 UK roles by 2030 — the largest planned workforce reduction in British telecoms history — in the same quarterly results presentation in which it reported a decade-low in consumer revenue. The AI engineering build running in parallel is less frequently quantified. BT's Digital and AI division, established in its current form under CEO Allison Kirkby, who joined from Tele2 in February 2024 as Philip Jansen's successor, has been adding net new AI engineering headcount at a rate ENTRA estimates at 280 to 340 positions per year through 2025 and H1 2026 — across a set of UK locations that sit almost entirely outside the King's Cross AI corridor framing that dominates UK AI hiring coverage.
The geography is the story. BT's AI engineering workforce is not concentrated in London. It is distributed, by operational necessity and deliberate design, across Adastral Park in Ipswich (the group's principal network research campus), Aztec West in Bristol (its customer-AI hub), Riverside Tower in Belfast (the BT Ireland and global service delivery operation), and Innovation Birmingham Campus (its Midlands engineering centre), with a secondary cluster in Edinburgh's Fountainbridge district. London's Aldgate East office houses Digital Strategy leadership but accounts for no more than 30 percent of the group's net AI engineering headcount additions in H1 2026, per ENTRA's Job Signal Index and LinkedIn headcount signal analysis. In a UK AI market where "where you are" predominantly means "which building in King's Cross," BT Group is building its AI engineering workforce somewhere else.
What BT Is Actually Building
The technical scope of BT's AI function is determined by the network it serves. BT operates the UK's largest fixed-line network — Openreach, the structurally separated infrastructure subsidiary, passes approximately 32 million UK premises with full-fibre or legacy copper connections — and the EE mobile network, covering over 99 percent of the UK population. Those two assets generate a continuous data stream: network fault signals, traffic routing events, and customer interaction data from approximately 30 million BT and EE consumer accounts. The AI engineering build is, at its core, the project of building ML systems that operate on that data at network scale.
Three technical domains are driving the majority of BT's H1 2026 AI engineering hiring. The largest by headcount is Network Intelligence — ML systems that automate fault detection, predictive maintenance, and traffic optimisation across the Openreach fibre network. Adastral Park in Ipswich houses approximately 400 network engineers and researchers, of whom ENTRA estimates an increasing share — approximately 85 to 110 as of June 2026 — are now working on ML-augmented network operations. The Adastral Park AI function is unlike anything in the King's Cross lab corridor: it requires engineers who work with telemetry data streams at petabyte scale, operate on real-time network signalling protocols, and deploy models onto the edge computing hardware embedded in BT's exchange buildings across the UK.
The second domain is Customer AI — the most publicly visible, and the most directly connected to the 55,000 reduction, because it is the function replacing contact centre volume. BT has been deploying retrieval-augmented generation systems across its EE and BT Consumer brands since 2024, automating first-line customer queries and compressing agent-handled call volumes. The Bristol Aztec West site, which houses a major BT consumer operations centre, is where the Customer AI engineering team is concentrated: approximately 65 to 80 AI engineers building and iterating the production LLM stack, per ENTRA's recruiter-side tracking.
The third domain is Enterprise and Public Sector AI — the function building the AI platform that BT sells to UK government and FTSE 350 clients under its BT One Cloud and managed network offerings. The Birmingham and Edinburgh centres are the primary locations, building the integration layer that connects client-side AI tools to BT's managed connectivity infrastructure. This function has the most explicitly remote-eligible hiring profile: engineers working with client data APIs and BT's cloud-edge network do not require access to physical BT equipment, and multiple H1 2026 offer conversations reviewed by ENTRA confirm these roles are structured as hybrid-remote, with two to three days per week at the relevant regional centre.
Compensation: Regional Parity, Not Regional Discount
BT's AI engineering compensation sits below the King's Cross lab corridor but above every comparable employer in its primary recruiting geographies. That local-market premium — not the London absolute number — is the closing argument BT makes to AI engineering candidates outside London.
At the Senior AI/ML Engineer level — the primary hiring band, typically requiring three to five years of post-graduate ML experience — BT's H1 2026 base salary bands run as follows, per ENTRA's Q2 2026 recruiter survey and candidate-side data from four people who received or accepted BT AI engineering offers and were granted anonymity to discuss compensation:
- London (Aldgate East/Adastral Park): £95K–£120K base, plus a 15 percent performance bonus target and BT Group plc share awards under the Employee Share Purchase Plan. Total first-year comp: approximately £110K–£145K (~$146K–$193K at Q2 2026 GBP/USD of ~1.33).
- Bristol (Aztec West): £82K–£100K base, 12–15 percent bonus. Total first-year comp: approximately £92K–£120K (~$122K–$160K at Q2 2026 GBP/USD of ~1.33).
- Belfast (Riverside Tower): £75K–£90K base, 12 percent bonus. Total first-year comp: approximately £84K–£105K (~$112K–$140K at Q2 2026 GBP/USD of ~1.33).
- Birmingham and Edinburgh: £80K–£98K base, 12–15 percent bonus. Total first-year comp: approximately £90K–£115K (~$120K–$153K at Q2 2026 GBP/USD of ~1.33).
The regional bands are 20 to 30 percent below BT's London rates on a gross basis. Adjusted for housing costs in Bristol, Belfast, and Birmingham relative to London Zone 2, the purchasing-power parity is broadly equivalent at the mid-career level. Bristol and Birmingham have no equivalent employer in the telecom-adjacent AI space at comparable total comp — the nearest alternatives are Rolls-Royce's AI engineering function in Bristol and HSBC's Birmingham technology centre, both within 10 to 15 percent of BT's regional bands.
The Skilled Worker visa threshold — £38,700 annually under the Home Office immigration rules in force June 2026 — is cleared by a multiple of two or more at every level of BT's AI engineering range, including the Belfast band. BT Group Services Limited holds established Tier 2 sponsor status on the Home Office register and processes certificates of sponsorship across all five regional sites. The multi-site sponsorship infrastructure is operationally unusual: most Tier 2 sponsors with complex geographic footprints have materially slower certificate processing at secondary locations. BT's scale as a sponsor normalises the process across Ipswich, Bristol, Belfast, Birmingham, and Edinburgh, giving international candidates the same four-to-six-week certificate of sponsorship timeline regardless of which UK city they are relocating to. That operational parity is a meaningful advantage over earlier-stage UK AI employers whose Tier 2 infrastructure is London-only.
Why This Is a Geography Story, Not a Remote-Work Story
BT's distributed AI hiring is not "remote-first" in the Hugging Face sense — globally dispersed teams with no physical anchor. It is distributed because the network it operates is distributed. An ML engineer working on Openreach fault prediction in Ipswich is not making a remote-work arrangement with a London employer. They are working at the specific site that hosts the network research infrastructure their models interface with. The distinction matters for how candidates read the offer.
ENTRA's vacancy tracking for BT's UK AI roles in H1 2026 shows 34 hybrid-remote postings at Senior ML Engineer or Principal ML Engineer level across Bristol, Birmingham, and Edinburgh, with 11 explicitly full-remote roles concentrated in the Enterprise AI integration function. The full-remote postings carry stated salary ranges of £85K–£105K base — a national rate that does not differentiate by candidate location, a structural departure from the geographic pay tiers that govern BT's hybrid roles. For an ML engineer in Manchester, Leeds, or Cardiff who does not want to relocate to one of BT's five AI centres, the full-remote Enterprise AI roles represent an entry point into a major-infrastructure AI engineering function that the candidate would otherwise not be able to access without a move.
The candidate profile BT attracts through this model is distinct from the King's Cross corridor's target hire. BT is not recruiting ex-Cambridge ML PhDs who are weighing DeepMind against ElevenLabs — that candidate pool, even when it considers BT, is not the primary intake. BT is recruiting from three main prior-employer categories, per ENTRA's recruiter-side analysis: mid-career software engineers from BT's own legacy IT functions who have retrained in ML (approximately 35 percent of H1 2026 AI hires, per ENTRA's estimate); applied ML engineers from regional UK employers in financial services, utilities, and the public sector (approximately 40 percent); and a smaller cohort of international engineers on the Skilled Worker route, primarily from India and EU member states, who want UK-based mid-career roles without the London cost structure (approximately 15 percent). The remaining 10 percent are university direct-entry hires, primarily from Manchester's AI and Data Science MSc programme and Edinburgh's School of Informatics — the two non-London university pipelines that BT's regional centres can access directly.
H2 2026 Forecast
BT's 55,000 reduction runs through 2030, and the pace is accelerating as Customer AI deployment scales: EE's LLM customer service stack moved from pilot to full production rollout in April 2026, and BT Consumer's equivalent deployment is targeted for Q3 2026. Each production LLM deployment reduces agent headcount and, in parallel, increases demand for the ML engineers who maintain and iterate the production stack — the Customer AI team at Bristol Aztec West is expected to add 40 to 55 further positions in H2 2026 as deployment volume grows.
ENTRA's H2 2026 outlook for BT's AI engineering function projects a further 150 to 200 net new positions across all sites by year-end, weighted toward Bristol (Customer AI scale-up) and Belfast (enterprise AI integration). If that trajectory holds, BT Group ends 2026 with an AI engineering bench of approximately 1,200 to 1,400 across its UK sites — the largest distributed AI engineering workforce in the UK's incumbent technology sector, and a structural counterweight to the lab-corridor narrative that frames London as the only credible location for serious AI engineering work.
For the mid-career ML engineer outside London who has been watching the UK AI hiring conversation happen in a city they do not live in: BT is not offering that conversation in London. It is having it in Ipswich.
Headcount and vacancy figures derived from ENTRA H1 2026 Job Signal Index and LinkedIn headcount signal analysis. Compensation data sourced from ENTRA Q2 2026 UK recruiter survey and candidate-side conversations with four people who received or accepted BT AI engineering offers in H1 2026 and were granted anonymity to discuss compensation figures; figures represent ENTRA estimates and are not confirmed by BT Group. BT Group plc is listed on the London Stock Exchange; BT Group Services Limited holds Tier 2 sponsor status on the Home Office Sponsor Register. Skilled Worker minimum salary threshold (£38,700) per Home Office immigration rules in force June 2026. BT 55,000 role reduction target and CEO appointment details per BT Group investor communications and public announcement, 2023–2024. BT Openreach network premises figure per Openreach regulatory reporting. BT Group declined to comment on specific AI engineering headcount, compensation bands, or H2 2026 hiring projections. All figures are ENTRA estimates unless otherwise stated.
For the clearing banks building AI engineering functions across distributed UK sites, see Barclays, HSBC, NatWest: The Clearing Banks' AI Hiring Push. For the Edinburgh AI engineering market that BT's Scottish centre draws from, see Edinburgh AI Graduate Cluster 2026. For the King's Cross AI corridor headcount that BT's distributed model stands in contrast to, see DeepMind UK Midyear Hiring H1 2026.
Global Career PlatformFind AI talent. Find your next role.
Booking is hotels. · Airbnb is apartments. · ENTRA is global careers.