House of Lords Launches Inquiry into FCA Regulatory Sandboxes: Value, Innovation, and AI Testing

The House of Lords Financial Services Regulation Committee has opened a formal inquiry into the UK's regulatory sandbox regime, with a call for evidence that launched on 17 September 2026. Committee chair Baroness Noakes set out the problem directly: there has been, in her words, surprisingly little evaluation of how sandboxes work in practice and what effect they actually have, even as new types of sandboxes keep being introduced.

That's a fair description of where things stand. Four separate sandbox vehicles now sit inside the UK financial perimeter, each authorised, structured and overseen differently, and none has been through a systematic, comparative parliamentary evaluation before now.

The committee's stated aim is to examine how sandboxes are used by both regulators and the firms that participate, and whether they're succeeding in driving genuine innovation. In practice, that splits into three separate strands worth tracking separately: whether each sandbox delivers value for money against its running cost, whether its outputs reach consumers rather than staying confined to participating firms, and whether admitting a small, selected cohort into a sandbox gives those firms a competitive advantage that unselected rivals don't get. That third strand isn't language the committee has used publicly, but it's the practical question sitting underneath 'value for money' that any authorised firm competing against a sandbox alumnus will recognise.

The oldest of the four vehicles is the FCA Regulatory Sandbox, which allows live market testing with real consumers under tailored, time-limited regulatory waivers. The FCA Digital Sandbox sits earlier in a firm's development pipeline, offering synthetic data and API access so a product can be de-risked before a firm even reaches the authorisation gateway.

The Digital Securities Sandbox is structurally different from both. It's run jointly with the Bank of England under powers granted by the Financial Services and Markets Act 2023, and it exists specifically to test tokenised securities and distributed ledger settlement within the UK's financial market infrastructure regime — which means systemic-risk questions apply to it in a way they simply don't for the other three.

The fourth, and newest, is the FCA's Supercharged Sandbox, launched in 2025 to test autonomous, 'agentic' AI models in a controlled environment before those models touch live customer outcomes. Agentic AI doesn't map cleanly onto supervisory categories built around a human decision-maker and a defined line of accountability, and the Lords' interest in AI testing concerns sits squarely on top of that programme.

For firms and legal practices preparing written evidence, the committee's core interests translate into a fairly practical set of questions to answer with real data rather than general impressions: what measurable outcome came out of sandbox participation, whether it changed an authorisation timeline under the Financial Services and Markets Act in a way that can actually be evidenced, whether cohort selection criteria were published and applied consistently, and — for AI-specific testing — what supervisory gap the sandbox closed that the ordinary authorisation process wouldn't have.

Written evidence is due by 23 October 2026. The committee's own published call for evidence is the place to confirm the exact submission cut-off time, since that detail wasn't consistently available across the sources reviewed for this piece.

The practical stakes sit downstream of the inquiry itself. If the committee concludes that sandbox participation confers a measurable authorisation advantage without a matching public benefit, that finding feeds directly into how the FCA structures cohort selection under FSMA going forward — meaning firms currently inside a sandbox, or applying to one, have more than a compliance-diary reason to pay attention to how this inquiry develops.