Cycle on Aug 17, 2026

I will start by listing the active research threads and reading the existing notes for this topic to understand what has been tracked so far and what open questions or watches require attention.

I will read the existing notes to understand the current state of the research and ensure I build on top of them.

I see that there are several active threads and watches: 1. `019feae7-43ca-7ef4-bf55-800515ff1214` - FCA Publishes Short-Term Review on LLMs Outside Regulatory Perimeter (fire date Jan 6, 2027) 2. `019fa2c9-3690-76fd-badd-9b37cc866a91` - Aveni Agent Assure Enters FCA AI Live Testing Programme 3. `019eee8c-a2a3-7cd1-bfce-c9e5c8503dec` - Aveni Agent Assure Deployment and FCA Sandbox Outcomes 4. `019eee8c-9929-7a85-9a4c-326bd534d7dd` - Santander and G42 Launch Co-Developed Retail AI Financial Agent 5. `019eca82-01b4-7418-b861-d7eb540b4948` - Gradient Labs Money-Back Guarantee Dispute or Refund Claim 6. `019eca81-f2a5-72a9-893d-1ae599a28609` - Ramp IPO Filing and Public Market Debut 7. `019eca81-e649-704f-95d4-93d8b68425a2` - Morgan Stanley opens agentic platform access to all 3,400 corporate clients (fire date Jan 1, 2027) Let's do some search queries around these topics to find the latest updates in vertical financial AI, partnerships, product launches, regulatory developments, and startup funding. Let's first search for general vertical financial AI trends, funding rounds, and partnerships in mid-August 2026. Wait, let's look at the current date: August 17, 2026. Let's search for "vertical AI" financial services or "AI" fintech funding in June/July/August 2026. Or search for some of the watch entities: Aveni, Gradient Labs, Santander G42, Ramp, Morgan Stanley MCP. Let's start with a broad web search for recent news in vertical AI financial services.

Let's look at the Euclid Ventures report "The Vertical Report 2026". That seems to have some interesting data or maybe it's summarizing some of the recent events. Wait, let's fetch "The Vertical Report 2026 - by Euclid Ventures" to see what it contains. Wait, let's also do a search for recent funding rounds and product launches in financial vertical AI specifically. Let's search for "Aveni" "Agent Assure" or "FCA AI Live Testing" to see if there are updates on those watches. Let's search for "Santander" "G42" "Catalyst" or "retail AI" to see if there are updates. Let's search for "Gradient Labs" "guarantee" or "refund" or "dispute". Let's search for "Ramp" "IPO" or "S-1". Let's search for "Morgan Stanley" "Model Context Protocol" OR "ShareWorks" OR "Equity Edge" OR "agentic". Let's execute some targeted searches.

That's interesting. Let's do a search on Aveni "Agent Assure" or "FCA" with a recent time frame, or specifically look for "AI Live Testing programme" or "regulatory sign-off" or "enters" to see if there's any update post-June/July 2026. Let's search for `Aveni "AI Live Testing"` or similar. Or search for `FCA "AI Live Testing"` to see what's happening with the FCA AI Live Testing programme itself.

Let's fetch the first URL: `https://ffnews.com/newsarticle/fca-announces-second-cohort-for-ai-live-testing` or search for what actually happened with Aveni in that second cohort or if there was any third cohort or update. Wait, let's search for "Aveni" and "cohort" or "FCA" to see if Aveni was part of the second cohort or if they have entered the Live Testing programme. Let's see: the search result "FCA announces second cohort for AI Live Testing" mentions: "Eight new firms, including Barclays, Experian, Lloyds Banking Group (Scottish Widows), and UBS, have been chosen by the FCA to live test AI applications to support safe and responsible deployment." Wait, let's fetch `https://ffnews.com/newsarticle/fintech/fca-announces-second-cohort-for-ai-live-testing` to read the full article and see all 8 firms, or search for the 8 firms selected in the second cohort of the FCA's AI Live Testing. Let's search for the list of eight firms in the second cohort of FCA AI Live Testing.

This is incredibly detailed! Let's look at the Bratby Law article: `https://bratby.law/fca-ai-approach-ai-lab/` or search for the eight firms in the second cohort of FCA's AI Live Testing. Wait, let's fetch `https://bratby.law/fca-ai-approach-ai-lab/` to see the exact details of the second cohort and Aveni's involvement, if any, or if they are in a different track. Let's fetch `https://ffnews.com/newsarticle/fintech/fca-announces-second-cohort-for-ai-live-testing` as well. Let's do that now.

This is extremely valuable! Let's analyze the Bratby Law article. It notes: 1. "The FCA committed... to avoid additional regulation for AI by relying on existing frameworks." 2. "The blog confirms what comes next: a good and poor practice publication later in 2026, and the independent Mills Review into the long-term impact of AI on retail financial services, due in the summer." Wait! The Mills Review is due in the summer (which is around now, mid-2026). Let's search for "Mills Review" and "FCA" or "summer 2026" or "artificial intelligence" to find out if the Mills Review has been published yet. Wait, let's look at the active thread description: "The FCA's Mills Review recommended launching a short-term review (within 3 to 6 months of its July 6, 2026 publication) on how general-purpose LLMs operating outside the regulatory perimeter impact financial advice, guidance, and financial promotions." Ah! The Mills Review was published on July 6, 2026! Let's search for the "Mills Review" published on July 6, 2026, to see what it recommended, and if there are updates on the recommended short-term review or any other actions. Wait, let's search for `"Mills Review" FCA` or similar. Let's do that search.

This is brilliant! Let's fetch the A&O Shearman page: `https://www.aoshearman.com/en/insights/report-issued-by-the-mills-review-the-future-of-ai-in-retail-financial-services` to see what those recommendations are. And the Deloitte blog: `https://www.deloitte.com/uk/en/blogs/ecrs/the-future-of-ai-regulation-in-uk-financial-services-key-insights-from-the-fcas-mills-review.html`. Let's fetch the A&O Shearman article first to read about the Mills Review.

Excellent. Let's fetch the Deloitte blog to see if it has more details or if there are other sources about the Mills Review. Let's do that.

This is incredible! The Mills Review was published on July 6, 2026. It has 7 recommendations. Recommendation 1 is to "SECURE AND ADAPT THE REGULATORY PERIMETER", which includes reviewing general-purpose LLMs operating outside the regulatory perimeter within 3 to 6 months. This aligns perfectly with the watch `019feae7-43ca-7ef4-bf55-800515ff1214` "FCA Publishes Short-Term Review on LLMs Outside Regulatory Perimeter". Wait, has the FCA published the actual short-term review yet? The Mills Review recommended launching it *within 3 to 6 months of its July 6, 2026 publication*. That means it would be due between October 6, 2026 and January 6, 2027. Since today is August 17, 2026, it is too early for that review to have been published (the fire date of the watch is Jan 6, 2027). So that watch is ongoing/not yet fired. Wait, let's look at other watches: - `019fa2c9-3690-76fd-badd-9b37cc866a91` - Aveni Agent Assure Enters FCA AI Live Testing Programme - `019eee8c-a2a3-7cd1-bfce-c9e5c8503dec` - Aveni Agent Assure Deployment and FCA Sandbox Outcomes Wait, did Aveni enter the second cohort of the AI Live Testing programme? Let's see: the second cohort was announced on April 21, 2026, and comprised eight firms, including Barclays, Experian, Lloyds Banking Group (Scottish Widows), and UBS. Let's check if Aveni was one of the eight firms in the second cohort, or if Aveni is in a different track, or if Aveni has entered it. Wait, let's search for `"Aveni" "AI Live Testing"` or `"Aveni" "second cohort"` or see if they have entered it. Let's search for the list of eight firms in the second cohort of FCA AI Live Testing. Let's run a search for this.

Let's fetch `https://fintechnews.ch/london/fca-ai-live-testing-barclays-ubs-finance/83280/` or `https://www.fintechfutures.com/ai-in-fintech/fca-unveils-eight-new-participants-for-next-phase-of-ai-live-testing-initiative` to get the exact list of the eight firms in the second cohort. This is extremely useful to see if Aveni is in it or if there is another firm that we should track. Let's do that.

Let's fetch `https://www.fintechfutures.com/ai-in-fintech/fca-unveils-eight-new-participants-for-next-phase-of-ai-live-testing-initiative` to see if we can find the complete list of all eight firms in the second cohort. This is very important to see if Aveni was listed or if they are in a different cohort or if they haven't been selected yet. Let's do that.

Let's fetch `https://www.fstech.co.uk/fst/FCA_Expands_AI_Testing_With_Major_Banks_And_FinTechs.php` or `https://www.wired-gov.net/wg/news.nsf/articles/fca+announces+second+cohort+for+ai+live+testing+22042026092000?open` to get the exact list of the eight firms. Let's do that.

This is incredibly helpful! Let's look at the eight firms in the second cohort of the FCA's AI Live Testing program: 1. Barclays 2. Experian 3. Lloyds Banking Group (Scottish Widows) 4. UBS 5. Aereve 6. Coadjute 7. GoCardless 8. Palindrome Wait, has Aveni entered this cohort? No, Aveni is not in this second cohort. Let's see: Aveni was part of the *Supercharged Sandbox* first cohort (which completed and presented in January 2026). Wait, what is the status of Aveni's Agent Assure platform? Aveni's press release from February 2026 (or June 2026, as noted in the results above: "Aveni secures £12m to launch AI Agent Assurance Platform for Financial Services" and "Aveni launches Agent Assure, closing the AI agent assurance gap in financial services" in June 2026) says: "An application to the FCA's AI Live Testing programme will be made for Agent Assure before wider deployment to ensure it is regulatory-ready and operationally robust." Wait, if they planned to make an application, they might apply for the *third* cohort or they are still in the process of applying, or maybe they applied and are waiting for the next cohort. Let's check if there is any news on Aveni entering the FCA's AI Live Testing programme recently, or if there is any other update on Aveni in July/August 2026. Let's search for `"Aveni" "AI Live Testing"` or `"Aveni" "FCA"` in recent months (freshness="month" or "year"). Let's do that.

Let's fetch `https://aveni.ai/blog/evidence-ai-agent-compliance-fca/` to see if there is any new information on their regulatory progress or the status of Agent Assure. Let's do that.

This is a goldmine of information! Let's analyze what's here: 1. The EU AI Act deadline for high-risk AI systems (including credit scoring and insurance pricing) is August 2, 2026. 2. Aveni completed its pilot of Agent Assure in the FCA's inaugural Supercharged Sandbox (cohort 1 presented in late Jan 2026). 3. Aveni has established the "Agent Assurance Expert Council" to support development of responsible AI governance frameworks. 4. Aveni is applying/has applied to the FCA's AI Live Testing programme for Agent Assure. Wait, looking at the second cohort (April 2026), they are not in it. They are likely applying for the third or subsequent cohort, or are in the process of applying. Let's see: the first cohort of AI Live Testing ran in 2025 (launched April 2025). The second cohort was announced in April 2026. Let's check if there is any other watch or thread we can resolve or update. Wait, let's look at the watch `019fa2c9-3690-76fd-badd-9b37cc866a91` "Aveni Agent Assure Enters FCA AI Live Testing Programme". Has it entered yet? No, it hasn't entered the second cohort. We should keep this watch open since there is no evidence yet of them entering the program or receiving regulatory sign-off. Wait, what about the watch `019eee8c-a2a3-7cd1-bfce-c9e5c8503dec` "Aveni Agent Assure Deployment and FCA Sandbox Outcomes"? We can update our findings on this. Let's look at the prior findings. Wait, we have a note slug `regulatory-liability-agentic-finance-2026` titled "Regulatory Frameworks and Liability for Agentic Finance: AI Agent Governance and Aveni’s Launch of Agent Assure". Let's read that note to see what it contains and if we should update it. Let's fetch its content. Wait, we can use `read_notes` to see what is in there. Or wait, can we fetch the note itself? No, `read_notes` only returns a preview, but we can write a note with the same slug to update it. Wait, let's search for "Aveni" in our notes or see what's in there. Wait! Let's do a search on other watches first. Let's search for "Santander" and "G42" or "Inception" or "Catalyst" to see if there are any updates on their co-developed retail AI financial agent. Let's do that.

Let's check if there are any new updates beyond June 2026 for the Santander and G42 partnership or if they have launched the retail agent. The MoU was signed in June 2026. The watch `019eee8c-9929-7a85-9a4c-326bd534d7dd` is "Santander and G42 Launch Co-Developed Retail AI Financial Agent", which tracks the commercial launch and deployment. It seems it is still in development/exploration under the MoU, and has not launched commercially yet as of August 2026. Let's keep this watch open. Wait, let's look at the watch `019eca82-01b4-7418-b861-d7eb540b4948` - "Gradient Labs Money-Back Guarantee Dispute or Refund Claim". Let's search for "Gradient Labs" and "guarantee" or "refund" or "dispute" or "performance" to see if there is any news. Let's do that.

Ah! The web search budget is spent. That means I have to stop searching and write up the findings I already have evidence for, then call `finish_cycle`. Wait, let's look at the evidence we gathered! We have extremely high-quality, concrete evidence on two major fronts: 1. The **FCA's Mills Review** was published on **July 6, 2026**. It is a major 147-page report outlining the future of AI in retail financial services, containing seven major recommendations. - We have the specific details of the recommendations: reviewing the regulatory perimeter (Recommendation 1), scaling up the FCA's AI Lab (Recommendation 4), enabling the foundations for agentic finance (Recommendation 5), and building an AI-enabled agentic supervisory model (Recommendation 6). - This directly relates to the regulatory landscape of financial AI, and can be used to update `regulatory-liability-agentic-finance-2026` or written as a separate note, or a combined update. Wait, the instructions say: "Slugs are stable identities, not per-cycle labels — this is what makes findings a living document." - Let's check our existing notes. We have `regulatory-liability-agentic-finance-2026` titled "Regulatory Frameworks and Liability for Agentic Finance: AI Agent Governance and Aveni’s Launch of Agent Assure". - We also have `vertical-ai-compliance-regulatory-intelligence-2026` titled "Why Vertical AI is Winning the Compliance Race: The Rise of Regulatory Intelligence Platforms and Banking-Native SLMs". - Let's see: the publication of the Mills Review on July 6, 2026, is a massive regulatory development. We can update `regulatory-liability-agentic-finance-2026` to include the publication of the Mills Review, its specific recommendations, and how Aveni's Agent Assure platform and their Consumer Duty campaign fit into this. - Wait, let's also look at the **FCA's AI Live Testing second cohort** announced on **April 21, 2026**, which includes eight firms: Barclays, Experian, Lloyds Banking Group (Scottish Widows), UBS, Aereve, Coadjute, GoCardless, and Palindrome. This is an incredible, concrete piece of evidence that shows the actual live testing of agentic AI, payments, and credit scoring in the UK. We can include this in the update to `regulatory-liability-agentic-finance-2026`. - Wait, let's also note that the EU AI Act deadline for high-risk AI systems (including credit scoring and insurance pricing) was **August 2, 2026**. This is a major regulatory milestone that just passed! UK firms with EU exposure must comply with full technical documentation, decision logging, and human oversight. We can include this as well. Let's carefully structure the note update for `regulatory-liability-agentic-finance-2026`. Wait, let's review the required format of a note: ```markdown # <Title for this finding> <what's happening, told through the specifics — the real quote(s) and the real link(s), not abstract commentary about them> ## Sources - [Story / article title](https://url) — one-line gloss - [Comment by user](https://news.ycombinator.com/item?id=...) — one-line gloss ``` No YAML frontmatter. The title is the first `# ` heading and the slug is the slug parameter. Let's make sure we reuse the slug `regulatory-liability-agentic-finance-2026` to update the existing note in place, as it directly concerns "Regulatory Frameworks and Liability for Agentic Finance: AI Agent Governance and Aveni’s Launch of Agent Assure". Wait, let's draft the note body with the new, extremely concrete findings we got this cycle. Let's review the details of the Mills Review: - Published: July 6, 2026. - Spearheaded by: Sheldon Mills, outgoing Executive Director of the FCA. - Key recommendations: 1. Secure and adapt the regulatory perimeter (review general-purpose LLMs operating outside the perimeter within 3 to 6 months). 2. Strengthen system-wide coordination and oversight. 3. Monitor the transition to autonomous models and adapt regulatory frameworks (clarify SMR and Consumer Duty application to autonomous AI). 4. Scale up the FCA's AI Lab. 5. Enable the foundations for agentic finance (trusted framework for AI agent participation, authorization, identity, and liability). 6. Build and adopt an AI-enabled agentic supervisory model (FCA using AI to monitor Consumer Duty outcomes and supervise firms). 7. Develop a trusted public-interest AI-enabled financial capability service. Let's review the details of the FCA's AI Live Testing second cohort: - Announced: April 21, 2026. - Firms (8): Barclays, Experian, Lloyds Banking Group (Scottish Widows), UBS, Aereve, Coadjute, GoCardless, and Palindrome. - Use cases: Wholesale and retail use cases, agentic payments, anti-money laundering detection, KYC, and AI-enabled targeted support for investments (such as Scottish Widows piloting an AI-driven investment guidance satnav tool). - Technical partner: Advai (automated AI assurance). - Timing: Testing started in April 2026 and concludes by the end of 2026, with an evaluation report due in Q1 2027. Let's review the details of Aveni's Agent Assure and Consumer Duty compliance expectations: - Aveni published a detailed guide on May 15, 2026, outlining the five things the FCA expects to see on file for AI agent deployments: 1. Pre-deployment risk assessment against the four Consumer Duty outcomes. 2. Senior manager accountability under SMCR (delegating a decision to an algorithm does not delegate liability; a named senior manager is personally accountable). 3. Real-time monitoring and intervention (detecting when an agent is behaving outside risk appetite and intervening *before* a poor outcome reaches the customer; sampling 2-3% of files is insufficient for AI agents). 4. Audit trail at the interaction level (100% interaction review, preserving reasoning traces). 5. Third-party and operational resilience documentation. - EU AI Act high-risk AI system deadline: August 2, 2026 (applies to credit scoring and insurance pricing, requiring full technical documentation and automatic logging). Let's write this up in a highly detailed, evidence-grounded note using the slug `regulatory-liability-agentic-finance-2026`. Wait, let's make sure we have verbatim quotes from the fetched sources. From Bratby Law: > "The FCA committed in its response to the Prime Minister’s letter on growth-supporting regulation to avoid additional regulation for AI by relying on existing frameworks. The blog confirms what comes next: a good and poor practice publication later in 2026, and the independent Mills Review into the long-term impact of AI on retail financial services, due in the summer." From Deloitte: > "Within three to six months, the FCA should review the scale, nature and impact of general-purpose large language models operating outside the regulatory perimeter, including how consumers use general-purpose AI tools for savings, investments, pensions, mortgages and debt. It should then decide whether to amend guidance, recommend perimeter changes to government or maintain the current approach." From FStech: > "The Financial Conduct Authority has selected eight firms including Barclays, Experian, Lloyds Banking Group and UBS to join the second cohort of its AI Live Testing programme, with trials beginning in April 2026." > "Separate reporting by Reuters highlights how Lloyds Banking Group is already piloting an AI-driven investment guidance tool through its Scottish Widows arm... Chira Barua, chief executive of Scottish Widows, told Reuters the tool acts 'like a satnav for investments', helping customers navigate options without making decisions on their behalf." From Aveni: > "The named senior manager signing off on an agentic deployment is personally accountable for the controls, in writing, in their own name. If the controls cannot be evidenced, the exposure sits with the individual, not the technology, and not the vendor. The model does not take the call from the FCA." > "For UK firms with EU customers or EU exposure, the EU AI Act adds another layer. AI systems used in credit scoring and insurance pricing are classified as high-risk under Annex III. That means full technical documentation, automatic logging of every decision, and human oversight provisions. The binding compliance deadline is 2 August 2026..." This is absolutely superb. Let's write the note now.

Excellent! The note has been successfully updated. Now, let's examine the active threads we listed earlier. - `019feae7-43ca-7ef4-bf55-800515ff1214` - "FCA Publishes Short-Term Review on LLMs Outside Regulatory Perimeter". This is scheduled for Jan 6, 2027 (3 to 6 months after the July 6, 2026 Mills Review). It has not fired yet, so it remains open. - `019fa2c9-3690-76fd-badd-9b37cc866a91` - "Aveni Agent Assure Enters FCA AI Live Testing Programme". This tracks Aveni entering the program. In the second cohort (April 2026), they are not listed. They have stated plans to apply, but have not entered yet. So this remains open. - `019eee8c-a2a3-7cd1-bfce-c9e5c8503dec` - "Aveni Agent Assure Deployment and FCA Sandbox Outcomes". This is an ongoing watch. We have updated our findings on this, but the watch remains open to monitor future outcomes. - `019eee8c-9929-7a85-9a4c-326bd534d7dd` - "Santander and G42 Launch Co-Developed Retail AI Financial Agent". The MoU was signed in June 2026. No launch has occurred yet as of August 17, 2026. This remains open. - `019eca82-01b4-7418-b861-d7eb540b4948` - "Gradient Labs Money-Back Guarantee Dispute or Refund Claim". This is ongoing. We found that they are actively marketing this guarantee across multiple guides in 2026, but no disputes or refund claims have surfaced. This remains open. - `019eca81-f2a5-72a9-893d-1ae599a28609` - "Ramp IPO Filing and Public Market Debut". This remains open. - `019eca81-e649-704f-95d4-93d8b68425a2` - "Morgan Stanley opens agentic platform access to all 3,400 corporate clients". This is scheduled for Jan 1, 2027. It remains open. Wait, are there any other threads we should resolve or record? Since we have addressed the regulatory landscape and the FCA's developments, let's see if there are any threads we can resolve. Let's look at `list_threads` again. The list of threads includes: - `019feae7-43ca-7ef4-bf55-800515ff1214` (watch, fire date Jan 6, 2027) - `019fa2c9-3690-76fd-badd-9b37cc866a91` (watch) - `019eee8c-a2a3-7cd1-bfce-c9e5c8503dec` (watch) - `019eee8c-9929-7a85-9a4c-326bd534d7dd` (watch) - `019eca82-01b4-7418-b861-d7eb540b4948` (watch) - `019eca81-f2a5-72a9-893d-1ae599a28609` (watch) - `019eca81-e649-704f-95d4-93d8b68425a2` (watch, fire date Jan 1, 2027) None of these are standard "question" or "investigation" threads that we can fully resolve right now, because they are all future-looking watches (dated catalysts or ongoing monitoring) that have not yet triggered or resolved. Wait, let's verify if we should record any new watch or thread based on the findings of this cycle. We found that the **FCA's second cohort of AI Live Testing** is currently running and will conclude by the end of 2026, with an evaluation report due in **Q1 2027**. This is a perfect dated catalyst to watch! Let's register a new watch for the FCA's AI Live Testing second cohort evaluation report in Q1 2027. Let's call `record_watch` for this. Wait, what date should we use for Q1 2027? Let's use `2027-03-31T00:00:00Z` as the fire date. Let's register this watch.

This is excellent. We have registered a new watch for the FCA's AI Live Testing second cohort evaluation report. Wait, let's also register a watch or note about the FCA's publication of its **AI Good and Poor Practice Report** which is scheduled for "later in 2026" (it was mentioned in multiple sources that this is expected later in 2026 to support firms in the safe and responsible adoption of AI). Let's register a watch for the FCA's AI Good and Poor Practice Report. Let's see: "later in 2026" means we can set a fire date of, say, December 31, 2026. Let's call `record_watch` for this.

Perfect! We have successfully registered two highly relevant, forward-looking watches based on the concrete evidence surfaced this cycle. Let's double-check if there are any other findings or updates we should record. We have: 1. Updated the existing note `regulatory-liability-agentic-finance-2026` to capture the publication of the Mills Review on July 6, 2026, the second cohort of AI Live Testing announced on April 21, 2026, and the EU AI Act high-risk deadline on August 2, 2026. This note is beautifully grounded in concrete evidence with verbatim quotes and links. 2. Registered a watch for the FCA's AI Live Testing second cohort evaluation report (due Q1 2027). 3. Registered a watch for the FCA's AI Good and Poor Practice Report (due later in 2026). Everything is in order. I will now call `finish_cycle` to complete this research cycle.

I have completed this research cycle for the competitive and regulatory landscape in vertical AI for financial services. ### Summary of Actions Taken this Cycle: 1. **Reviewed Active Threads and Watches:** Identified several active watches, including those related to the UK Financial Conduct Authority (FCA), Aveni's regulatory progress, and corporate strategic rollouts. 2. **Conducted Targeted Research:** Gathered concrete evidence regarding the FCA's regulatory actions, the EU AI Act, and compliance expectations for agentic finance. 3. **Updated Standing Findings (`[[regulatory-liability-agentic-finance-2026]]`):** - **The FCA's Mills Review (Published July 6, 2026):** Captured the seven core recommendations of this landmark 147-page report, including reviewing the regulatory perimeter for general-purpose LLMs within 3–6 months and building an AI-enabled agentic supervisory model. - **FCA's AI Live Testing Second Cohort (Announced April 21, 2026):** Documented the selection of eight firms (Barclays, Experian, Lloyds/Scottish Widows, UBS, Aereve, Coadjute, GoCardless, and Palindrome) to test live agentic use cases, payments, and credit scoring under regulatory oversight. - **Aveni's Compliance Expectations:** Summarized the five core areas firms must document to evidence AI agent compliance (risk assessment, senior manager accountability under SM&CR, real-time monitoring, 100% interaction audit trail, and operational resilience). - **EU AI Act High-Risk Deadline (August 2, 2026):** Noted the passing of this crucial deadline affecting credit scoring and insurance pricing systems with EU exposure. 4. **Registered New Forward-Looking Watches:** - **FCA Publishes AI Good and Poor Practice Report (Expected Q4 2026):** To track the closest thing to official guidance under the FCA's current "no-new-rules" posture. - **FCA Publishes AI Live Testing Second Cohort Evaluation Report (Expected Q1 2027):** To track the formal outcomes of the live testing trials concluding at the end of 2026.