RE: Canaccord Genuity reiterates Buy with £9 target21 Apr 2026 13:06
The Central Argument
The UK Government has declared its intention to become an AI-led state and is spending billions to get there. But its execution problem is operational, not strategic. Government lacks the internal capability to embed AI into day-to-day public administration. That gap is what a business process outsourcer fills — and a BPO led by a CEO who has built AI products, run hyperscale cloud operations, and executed this transformation twice before is categorically different from one led by a career outsourcing executive.
This is not a prediction of success. It is an argument that the probability of successful execution is materially higher under this CEO than under any plausible alternative. The thesis requires serious engagement with the failure conditions, which are substantial.
Part One: The Government's AI Problem
The UK public sector awarded 1,809 AI contracts worth £3.8 billion between 2018 and March 2026. 2025 was the biggest year on record at £1.17 billion — a 102% jump on 2024. The government's AI Commercial Strategy is explicit: the priority is buying AI-enabled services from the market, not building in-house.
But ambition and execution are different things. Government sits on an estate predating the internet, spending approximately £525 million per year on shared services delivered through 286 fragmented legacy platforms. 54% of civil servants have received no AI training whatsoever. Only 17% use AI for advanced tasks.
Why AI Does Not Eliminate the BPO Layer
Consultancies design transformation. Hyperscalers supply infrastructure. SaaS vendors enable workflows. None of them run the processes. As Hernandez told analysts: "We do not build the engines, we fly the plane, and for some customers, we actually run the airline." Government does not need software that can automate payroll — it needs a counterparty who will take contractual accountability for 250,000 civil servants' pay being correct every month.
The risk to this framing is that AI may compress the BPO layer rather than augment it. If government departments deploy AI tools directly through SaaS platforms with minimal human oversight, outsourced volumes could shrink. Hernandez's counter — that government processes are too complex and exception-heavy for generic AI tooling — is plausible but unproven at scale
Part Two: The Career DNA of Adolfo Hernandez
The critical distinction is that Hernandez is a computer scientist who became an operator — not an operator who learned to talk about technology. BSc Computer Science (Universidad de Granada), MBA (IDE-CESEM), career beginning at IBM in London and Munich across enterprise software and open systems.
Phase One — Enterprise Technology at Scale. Sun Microsystems SVP Global Services, running a $5bn+ services business with global P&L responsibility. Alcatel-Lucent EVP, incubating cloud platforms including Cloudband and API management products — building cloud products inside a large enterprise b
Sign In
Follow the stocks
