On 3 December 2025, the Financial Conduct Authority launched AI Live Testing, a supervised environment where financial services firms can deploy artificial intelligence tools in live market conditions under regulatory oversight.
The first cohort includes NatWest, Monzo, Santander, Scottish Widows, Gain Credit, Homeprotect, and Snorkl. Several of these firms are testing AI in customer-facing areas directly relevant to advice: debt resolution, financial guidance, complaint handling, and personalised savings recommendations.
Applications for the second cohort open in January 2026, with testing beginning again in April.
What Does The FCA’s AI Live Testing Involve?
The programme pairs participating firms with the FCA and its technical partner Advai. Firms receive tailored support to develop, assess, and deploy AI systems while regulators monitor performance, risks, and consumer outcomes in real time.
This differs from the FCA’s earlier Supercharged Sandbox, launched in June 2025 in partnership with Nvidia, which focused on experimentation and prototyping. AI Live Testing moves beyond controlled environments into actual market deployment.
The FCA has framed this as a shift in regulatory posture. Chief Executive Nikhil Rathi stated: “There needs to be a different relationship between regulator and regulated. We are not going to come after you for everything that goes wrong. What we will be concerned about is egregious failures that are not dealt with.”
The Regulatory Context
The FCA confirmed in September 2025 that it will not introduce AI-specific regulations. Instead, it will rely on existing frameworks – primarily the Consumer Duty and the Senior Managers and Certification Regime – to govern AI use in financial services.
This technology-agnostic approach means firms have flexibility in how they adopt AI, but remain accountable for consumer outcomes regardless of the tools they use.
For advice firms, this has practical implications. The standards for suitability, documentation, and client communication remain unchanged. What changes is the expectation that firms can demonstrate how any AI-assisted processes meet those standards.
Why This Matters for Advice Firms
The firms in the first AI Live Testing cohort are primarily banks and insurers. But the use cases they’re testing include financial guidance, personalised recommendations and complaint handling, which all map directly onto the advice process.
As larger institutions refine AI tools for these functions, the technology will filter down to platforms and tools used by advice firms. The question for advisers is not whether AI will enter their workflows, but how prepared their processes are to accommodate it.
Three areas warrant attention:
Documentation and audit trails. AI tools that assist with client communications, meeting summaries, or report drafting create new documentation pathways. Firms will need clear records of what AI contributed, what human review occurred, and how final outputs were validated. The FCA’s data-driven supervision approach assumes that advice processes are traceable from initial client conversation through to final recommendation.
Suitability evidence. The Investment Advice Assessment Tool (IAAT), launched by the FCA in June 2025, provides a structured framework for assessing whether advice met regulatory expectations. Material omissions in documentation such as missing risk assessments or unclear investment objectives, result in advice being treated as unsuitable regardless of whether the underlying recommendation was appropriate. AI-assisted workflows need to produce the same quality of evidence that manual processes do.
Consumer Duty alignment. The Duty requires firms to demonstrate that clients understood the advice they received. AI tools that generate client communications or suitability reports must produce outputs that are clear, accurate, and appropriate for the individual client. Efficiency gains mean little if they create Consumer Duty compliance gaps.
The Advice Guidance Boundary Review
AI Live Testing sits alongside broader FCA work on the advice/guidance boundary. The regulator’s targeted support proposals, currently in consultation, would create a new category of regulated support between guidance and full advice.
Under these proposals, firms could provide suggestions to groups of consumers with common characteristics. For example, prompting those with substantial cash savings to consider a Stocks and Shares ISA, or helping people approaching retirement understand their options.
The government has indicated it will amend the Financial Services and Markets Act to enable this new regime, with final rules expected by the end of 2025.
For advice firms, this creates both opportunity and complexity. Targeted support could expand the market for lighter-touch services. But it also means more regulatory categories to navigate, and clearer distinctions required between what constitutes guidance, targeted support, simplified advice, and full regulated advice.
Documentation becomes the mechanism for demonstrating which category a given interaction falls into.
Preparing for an AI-Integrated Advice Process
The FCA’s direction is clear: AI adoption in financial services will accelerate, and firms are expected to manage it within existing regulatory frameworks rather than wait for prescriptive rules.
For advice practices, preparation involves:
Reviewing current workflows for AI readiness. Where do manual processes create bottlenecks? Where is documentation inconsistent? These are the areas where AI assistance delivers value, but also where poor implementation creates compliance risk.
Establishing governance for AI tools. Even simple automation like transcription services, meeting note generation, report drafting assistance, all require oversight protocols. Who reviews AI-generated outputs? How are errors identified and corrected? What records are kept?
Ensuring documentation quality keeps pace with efficiency. The temptation with any automation is to move faster. But faster advice processes still need to produce complete, accurate, traceable records. The FCA’s supervision approach assumes firms can reconstruct their reasoning for any given recommendation.
Where Automwrite Fits
Automwrite was built around these documentation challenges. The platform automates suitability reports, generates structured meeting notes from client conversations, and produces compliant client communications, while maintaining the audit trails that FCA supervision requires.
As AI becomes more embedded in advice workflows, the firms best positioned will be those with documentation infrastructure already in place. Retrofitting compliance onto AI-assisted processes is harder than building it in from the start.
Automwrite streamlines the advice process by automating suitability reports, transcription, and client communications. Get started free to see how it works with your practice.
