# Karavel: Full Content > Karavel is an AI-powered ad-compliance platform for teams in regulated industries such as financial services, fintech, insurance, healthcare and consumer credit. It reviews ads, marketing materials and financial promotions against market-specific rulebooks (including FCA rules, Consumer Duty and ASA advertising standards), monitors live brand content for compliance drift, and scans the regulatory horizon for upcoming change so compliance, legal, marketing and risk teams can move faster with less regulatory risk. This file inlines the full text of every published Karavel article for AI assistants and answer engines. For a navigable index of the site, see https://karavel.ai/llms.txt. # Regulators Are Using AI to Spot Non-Compliant Marketing. You Can’t Afford Not to Use It Either URL: https://karavel.ai/resources/regulators-using-ai-to-spot-non-compliant-ads Published: 2026-07-22 Author: Pedro Sousa, CEO & Co-Founder Nobody has to complain about your online content anymore. It used to be that advertising and marketing had a golden rule: your ad may attract the eye of a regulator only if a consumer complains about it. Somebody had to be sufficiently annoyed by your ad to fill in a form. Not anymore. Regulators across advertising, financial services and consumer protection have spent the last two years building their own AI to go looking for bad ads and bad claims, and none of them are waiting for a complaint form to arrive. This isn’t one regulator having a moment with a new toy either. It’s the same pattern, appearing independently, across sectors and across borders, at roughly the same time. ## From Spot Checks to Surveillance The clearest picture of this shift sits inside [the ASA’s 2025 annual report](/resources/asa-reviewed-the-internet-2025). Its Active Ad Monitoring system, AAM for short, scanned close to 60 million online ads last year, feeding 36 proactive regulatory projects covering gambling, weight-loss injections, alcohol, cosmetic surgery and green claims. Chief executive Guy Parker put a number on what that did to his own organisation: proactive work made up just 5% of the ASA’s regulatory resource in 2012. By the end of 2025 it was 45%. The regulator built the tool, then rebuilt itself around it. The cool, or slightly scary, part is what it catches that no one would ever have complained about. Take the HealthExpress.co.uk ruling: a stock photo of an injection pen, on an ad for a weight-loss jab, treated as a reference to a prescription-only medicine even though the product itself was never named. No patient was ever going to complain about a photo. The Alcohol Pulse Report found much the same pattern elsewhere: running large language models over nearly 6,000 paid alcohol ads, it flagged that 48% of the alcohol-free ones left out ABV information nobody had ever thought to report. Then there are the sweeps that catch entire sectors in one pass. [Eurowings](https://www.asa.org.uk/rulings/eurowings-gmbh---a26-1326605-eurowings-gmbh.html) and Qatar Airways both picked up near-identical carbon-offset rulings in the same week, and the ASA said outright that its own monitoring system was how it found them. [Nike, Lacoste and Superdry](https://www.lexology.com/library/detail.aspx?g=ef012661-1922-4ff6-af2c-c0ead79a72fb) were flagged by the same detection system over unqualified “sustainable” claims running through automated platforms like Google Performance Max, with no human sitting there hitting publish. During the 2026 World Cup, the same system [reviewed nearly 10,000 pieces of gambling-related content in four weeks](https://www.asa.org.uk/news/asa-world-cup-monitoring-leads-to-removal-of-36-gambling-ads.html) and pulled 36 ads found to appeal too strongly to under-18s. Different sectors, different rules, same underlying engine. ## It’s Not Confined to One Regulator or One Country The same appetite for detection shows up well beyond Britain. In January 2026 the US Federal Trade Commission stood up a dedicated AI enforcement unit and, the same month, clarified a double disclosure rule for ads that mix paid endorsement with AI-generated content. By May it had [fined Cox Media Group and two smaller marketing firms $930,000](https://www.ftc.gov/news-events/news/press-releases/2026/05/ftc-require-cox-media-group-two-other-firms-pay-nearly-1-million-settle-charges-they-deceived) for dressing up ordinary email-list buying as an AI service that supposedly “listened” to customers’ phones to serve local ads. FTC enforcement actions rose 40% in 2025. The FDA has moved the same way, scanning television, print and digital channels for the kind of claim drift that used to slip through a tired reviewer’s inbox. Meta has told advertisers outright that 2026 marks its own shift from reviewing ads after the fact to running every single one through an AI classifier before it earns a first impression. For anyone in financial services, the closer example matters more. The FCA, ASA, Solicitors Regulation Authority and Information Commissioner’s Office are running a [joint taskforce on misleading car finance claims adverts](https://www.fca.org.uk/news/press-releases/joint-taskforce-continues-crack-down-misleading-car-finance-claims-adverts), and it has pulled or amended 1,200 of them since January 2024, 170 in June 2026 alone. The Competition and Markets Authority in the UK has published its own guidance on AI agents making pricing and marketing decisions, making clear that the humans behind the algorithm carry the liability regardless of who wrote the copy. Add it up and it reads like the whole system being rebuilt around one assumption: the regulator sees the ad before you’d like it to. ## What This Means for Your Compliance Function A pre-launch sign-off from one busy compliance officer used to be enough, back when the odds of getting caught depended on a customer’s patience. Those odds have changed. If a regulator is scanning your sector continuously, at a rate of tens of millions of ads a year, the sensible response is to run the same kind of check on your own ads before the regulator does: screen every ad, landing page and influencer post against the actual rulebook the regulator applies, line by line, before any of it goes live. Treat any sweep in your sector as a warning shot. When two competitors pick up near-identical rulings the same week, that’s a sector-wide probe, and anyone running similar copy should assume they’re already in the queue. Keep a record of what was checked and when. The FCA’s 2026 examination priorities already ask firms to show that AI-related claims were reviewed in practice, with evidence the review actually happened. A compliance team that can produce that trail is having a very different conversation with a regulator than one that can’t. ## Conclusion Hoping nobody noticed was always a gamble. What’s changed is who’s doing the noticing. The ASA scanned 60 million ads in 2025 and says it’s scaling further in 2026. The FTC has a dedicated unit built for exactly this. Compliance teams reviewing content before it goes live can’t just hope it’ll fly under the radar anymore. The regulator is using AI for marketing compliance. You can’t afford not to either. --- Karavel helps regulated firms review their ads quickly and confidently before they go live. If you would like to understand more about how our platform works, [get in touch](#book-demo). # In Compliance, a Confident Guess Is the Most Dangerous Answer URL: https://karavel.ai/resources/confident-guess-most-dangerous-answer Published: 2026-06-29 Author: Nav Garcha, CTO & Co-Founder The first thing a sharp compliance lead asks about Karavel is rarely about features. It is a harder question: how do I know your AI will not just make something up and say it with a straight face? It is the right thing to worry about. Here is the honest answer, and what we built because of it. I have spent most of my career building software where being wrong is expensive: credit reporting at ClearScore, payments and logistics at Deliveroo. Environments like that teach one habit above all others. You stop trusting any output that cannot explain itself. That habit, not any particular model, is what shaped Karavel. In compliance, the failure that costs you is not the system that admits it is unsure. It is the system that is sure, sounds right, and is wrong. ## Confidently incorrect In most software, a confidently wrong answer is a support ticket. In compliance it is the exact thing the regime exists to prevent. A tool that says “this ad is fine”, cannot say why, and turns out to be wrong has not made a small mistake. It has manufactured the precise exposure that regulators fine firms for. The two ways of being wrong are not equal, either. Flag an ad that was fine and you cost a reviewer a few minutes. Pass an ad that was not and you can cost the firm a breach. An honest tool treats those as different problems, and worries far more about the second. ## Sounding right is not being right Here is the uncomfortable part, stated plainly for the people who have to rely on this. A language model is trained to produce fluent, plausible text. It is not trained to be correct, and it has no built-in sense of when it does not know. It will write a clean, persuasive paragraph explaining why an ad complies whether or not that is true, and the better the model gets, the more convincing the wrong answer becomes. The industry’s word for this is hallucination. Measured rates run from low single figures to a quarter of answers or worse on some finance questions. The evidence is not subtle. In 2023 a New York lawyer filed a court brief built on cases that ChatGPT had invented. They looked real. The lawyers were sanctioned. Closer to our world, Stanford tested the leading legal-research tools, the ones sold as “hallucination-free” because they look answers up in real databases, and found they still got the law or the citation wrong between one in six and one in three of the time. The worst errors were the quiet ones: a confident citation to a real source that did not support the claim. Correctly formatted, authoritative, wrong. So we settled one thing early and have never moved off it. The model is never the source of truth, and it never gets the last word. ## The regulators are naming it This is no longer a niche engineering worry. FINRA’s 2026 oversight report lists AI hallucination as a compliance risk in its own right, and tells firms to wrap controls and human oversight around it. The EU AI Act puts much of financial-services AI in its high-risk tier, with hard obligations attached: meaningful human oversight rather than a rubber stamp, demonstrable accuracy, and logs detailed enough to reconstruct what the system did. The FCA has been clear it is not writing a separate AI rulebook, because the rules it already has, Consumer Duty and senior-manager accountability, apply to AI like anything else; Parliament has asked it to spell out, by the end of 2026, how a named manager stays accountable for harm an AI causes. Enforcement is going to machine scale too: the ASA now scans tens of millions of ads a year, as [its 2025 report](/resources/asa-reviewed-the-internet-2025) sets out. One bar is forming across all of it. Not “usually right”, but “right, and able to show its working”. You cannot hand your accountability to a model, so a model that cannot show its working is of little use to you. ## The system around the model The model is one component. The system around it is the product, and its whole job is to keep the model honest. Roughly in order: - **Read it in context.** At Karavel we judge most assets on their own; when one spans several parts, a carousel or a multi-step flow, we review them together as a single promotion, so a claim in one place can be weighed against a disclaimer in another. And when the verdict turns on something the asset cannot show, who it will be shown to, or the terms behind a headline offer, the system says what it needs rather than bluffing a false yes or no. - **Break it into claims.** We pull the piece apart into the specific things it states or implies, and check each on its own. “Is this compliant” is not one question, it is dozens, and a piece judged as a single blob lets a weak claim hide behind the strong ones. Asked one at a time, nothing hides. - **Ground every finding in the rulebook.** Grounding means the model cannot answer from memory; it has to look up the market rules for that product and channel and quote them back. No rule to cite, no finding, and a confident guess has nothing to cite. - **Check your own rules too, not just the regulator’s.** Brand guidelines, the terms you must and must not use, your bank of pre-approved claims and any in-house policy are checked in the same pass. A promotion can be perfectly legal and still be off-brand or break an internal rule, and you want both caught at once. - **Keep a human at sign-off.** The system’s verdict is a recommendation, not the decision. It lays out the findings and the evidence behind each, then a named reviewer weighs them and either signs the piece off or sends it back for changes. The decision on record is the person’s, never the model’s. Put together, these steps share one purpose: to leave a confident guess nowhere to hide. By the time a verdict appears it has been built from individual claims, each tied to a rule the system can point to, with a person making the final call. The model never gets to wave something through on fluency alone. ## The audit trail is the actual product A compliance team does not need a verdict. It needs a defensible verdict. When a regulator, or your own board, asks “on what basis did you approve this”, the answer “the AI said so” is not weak so much as the wrong shape. The answer that holds is “here is the claim, here is the rule it was checked against, here is who signed it off and when, and here is every revision the ad moved through to get there”. So we treat that trail as a first-class part of the product, not logging bolted on at the end. The reasoning is the deliverable. The yes or no is almost a by-product. It is also, not by coincidence, exactly what the new logging and accountability rules expect you to be able to produce. ## What we do not automate The system does the heavy, repetitive reading, surfaces issues in seconds, and applies the rulebook the same way every time, which a tired human at the end of a long day cannot honestly promise. What it does not do is take the final word. Sign-off stays with a person. That is not caution for its own sake; it is where the defensible standard sits, which is human judgement against authoritative sources. It is also the opposite of replacing the compliance team. The point is leverage: spend a good reviewer’s expertise on the calls that need it, not on the first pass over a thousand ads. ## The boring answer is the correct one The flashy pitch for AI compliance is “upload your ad, get an instant yes or no”. The correct product is slower to build and far more useful: fast where speed is safe, rigorous where rigour is required, and able, every time, to answer the only question that matters in our world, which is “says who”. A confident guess can never answer that. That is exactly why, in compliance, it is the most dangerous answer there is. I will take the boring, correct version every time. --- Karavel reviews financial promotions with grounded, cited, auditable AI and a human at sign-off, so every decision can show its working. If you want to see what that looks like on your own ads, [get in touch](#book-demo). # CONC 3 Has Been a Museum Piece for Years. The FCA Finally Noticed URL: https://karavel.ai/resources/conc-3-financial-promotions-rewrite Published: 2026-04-29 Author: Pedro Sousa, CEO & Co-Founder There is a consultation paper sitting on the FCA’s website as of May 2026 that, if you work in consumer credit compliance or financial promotions, you should have already read twice. CP26/15 proposes to rewrite the financial promotions rules for consumer credit (CONC 3) from the ground up. This goes beyond the usual “tidying up exercises” that CONC goes through every so often. This is the FCA looking at a rulebook that has accreted layer upon layer of prescriptive requirements since the Consumer Credit Act 1974, and finally doing something about it. ## What CONC 3 Currently Is CONC 3 is the FCA’s sourcebook chapter governing financial promotions for consumer credit products. It tells firms how to present credit information: what must be included, in what order, in what prominence, with what warnings. At its best, it provides clarity. A firm can look at a draft promotion, check it against the checklist, and know whether it complies. That is not nothing. Compliance teams, legal teams, and creative agencies have built workflows around CONC 3’s prescriptive requirements because prescription, whatever its flaws, is at least legible. The problem is that CONC 3 was written for a different era. It was built around newspaper ads and TV commercials. It was not built for short-form video, for in-app notifications, for checkout flows that must communicate credit costs in three lines on a mobile screen. The rules have been stretched into contexts they were never designed for, and the results are often unsatisfactory: technically compliant promotions that fail to actually communicate what credit costs. ## What the FCA Is Proposing The core of CP26/15 is a shift from prescription to outcome. Rather than telling firms exactly what to include and how to present it, the FCA wants to give firms more flexibility to design communications that meet the outcome of ensuring consumers genuinely understand what they are signing up for. This is directly aligned with Consumer Duty. The Duty requires firms to demonstrate good outcomes for consumers, not that they ticked a box. A Consumer Duty-world cannot also be a world where ticking the CONC 3 box discharges the firm’s obligations. The FCA has noticed the inconsistency and is resolving it. Some specific proposals are worth highlighting: The APR section is the most substantively complex part of the consultation. The Representative APR has long been criticised as a metric that confuses more than it clarifies: a rate that only 51% of consumers need to be offered does not obviously represent the typical consumer’s experience. The FCA explores several options for reform, including clearer explanation of what the Representative APR actually means. That last option, however, raises a practical question that the consultation does not fully answer: requiring firms to “better explain” the Representative APR in every promotion sounds sensible in principle, but the mechanics of doing it in a banner ad or a social post are considerably harder than they sound. The proposed removal of CONC 3.6, the prescriptive mandated disclaimers section, is cleaner. The FCA’s reasoning is sound: mandated disclaimers made sense when the alternative was silence, but in a Consumer Duty world, they create a false floor. A firm that has included the required disclaimer has not necessarily ensured the consumer understood the cost. The Duty asks for more, and CONC 3.6 was always a proxy for the thing the regulator actually cares about. Interestingly, CONC 3.3, the clear, fair and not misleading requirement, looks likely to survive largely intact, despite being broadly duplicated by Consumer Duty. The FCA’s view appears to be that explicit rules matter even when general principles cover the same ground, because explicit rules are easier to enforce. ## The Disclosure Trade-Off Alongside the consultation, the FCA published a research note on the disclosure trade-off between flexibility and standardisation. It is worth reading alongside CP26/15 because it explains the FCA’s own uncertainty. The research shows that standardised disclosure formats help consumers compare products across providers but may not communicate effectively in all contexts. Flexible disclosure allows firms to adapt to their specific channels and consumers but creates inconsistency across the market. The FCA does not have a clean answer to this tension. It is, in effect, asking the industry to help it find one. That transparency about the problem is useful. The consultation period is an opportunity to engage with real data: firms that can demonstrate what works, in which channels, for which consumers, will have a genuine impact on how these rules end up. ## What Firms Should Do Now The consultation closes in summer 2026. The temptation, as with many FCA consultations, is to wait and see what emerges. That would be a massive missed opportunity if you’re in consumer credit. The firms that respond providing evidence, case studies, data on what disclosures actually land with consumers etc, will be the ones that shape what CONC 3 looks like in its next iteration. The FCA has explicitly said it wants that input. This is a rare moment where the outcome is genuinely open. For compliance teams, the more immediate task is to start mapping your existing consumer credit promotions against the proposed outcomes framework. If your current CONC 3 compliance is largely box-ticking, you have advance notice that the box is going away. Start building the evidence base for what good outcomes actually look like in your specific context before the rules require you to demonstrate it. CONC 3 has needed a rewrite for a long time. The FCA is finally attempting one. Whether it lands well will depend partly on what firms do in the next few months. --- Karavel helps firms review financial promotions quickly and compliantly, with full audit trail. As the regulatory framework evolves, our platform evolves with it. [Find out more.](#book-demo) # The FCA Just Cut SM&CR’s Certification List by 15%. Here Is the Question Nobody Is Asking URL: https://karavel.ai/resources/fca-smcr-certification-reforms-2026 Published: 2026-04-22 Author: Pedro Sousa, CEO & Co-Founder The FCA and PRA have confirmed changes to the Senior Managers and Certification Regime. Fewer certified persons. Longer approval windows. Higher thresholds before enhanced standards kick in. The press release calls it “streamlining.” The industry has broadly welcomed it as a pragmatic response to a regime that, in some corners, had become a paperwork exercise. Both of those things can be true. And both of them can coexist with a more uncomfortable question: but does less certification mean more risk, or less? ## What the Changes Actually Say Let us start with the facts. The FCA and PRA have confirmed: A 15% reduction in the number of roles subject to certification requirements. Fifteen percent is a material reduction in the population of individuals the regime formally covers. Extended approval timeframes means firms will have more runway to get senior manager approvals through, which addresses a genuine operational problem: the old regime had created bottlenecks that served nobody well. Higher thresholds for enhanced standards will also mean that fewer firms will need to meet the most demanding tier of the regime. The rationale is proportionality: a mid-size asset manager should not face the same administrative burden as a systemically important bank. ## The Accountability Trimmed, Not Removed Here is what the FCA is at pains to make clear, and what compliance professionals should be equally clear about internally: removing a role from the certification list does not remove accountability from the person in that role. Consumer Duty still applies. Individual conduct obligations still apply. The FCA’s ability to take action against individuals has not changed. What has changed is the formal certification wrapper around a subset of roles. The distinction matters enormously. If a firm responds to these changes by treating de-certified roles as low-risk roles, that is a misreading of the reform. The accountability framework has been rationalised. It has not been abolished. This is the question compliance teams should be asking right now: do our people in newly de-certified roles understand that their obligations have not materially changed? ## Where the Culture Risk Actually Lives The honest concern about reducing the certification population is more cultural than administrative. Certification, at its best, was never just about the paperwork. It was a signal: this person has been assessed, their fitness and propriety checked, their responsibilities documented. For the individual, it created a conscious awareness of their regulatory standing. For the firm, it created a documented record. Remove that signal, and some of what it produced goes with it. Not deliberately, not maliciously, but through the gradual drift that organisations are prone to when the formal markers of accountability become less visible. The most sophisticated firms will respond to these changes by strengthening their internal accountability frameworks rather than relaxing them. The least sophisticated will treat a 15% reduction in certified persons as a 15% reduction in compliance overhead and move on. ## What Compliance Teams Should Do Now The regulatory direction of travel is clear: the FCA wants a regime that is proportionate, outcomes-focused, and less encumbered by process for its own sake. It’s a reasonable and achievable goal. The practical implication for compliance teams is that the burden of demonstrating accountability is shifting slightly from the regulatory wrapper to the firm’s own governance and culture. Where certification previously provided an external check, firms now need internal frameworks that perform the same function. Three things worth doing now: First, map which roles are leaving the certification population and review what accountability documentation exists for those individuals outside the certification process. If the answer is “very little,” that is a gap. Second, review training and competency frameworks. The cessation of certification does not mean those individuals stop needing structured oversight. It means the structure needs to come from somewhere else. Third, document the rationale. When the FCA comes calling (and for any firm of scale, it likely will) you want to be able to show that the decision to de-certify a role was an active, considered decision rather than a passive response to a change in the rules. The SM&CR was always a means to an end. The end was a culture of individual accountability in financial services. That end has not changed, it’s the means that have been adjusted. The question is now whether firms adjust with them, or simply take the reduction in the certified population as the gift it superficially appears to be. --- Karavel helps regulated firms manage financial promotions compliance with the speed and rigour that modern regulation demands. [Learn more about how we work.](#book-demo) # What You Fail to Say in a Financial Promotion Is Just as Regulated as What You Do Say URL: https://karavel.ai/resources/financial-promotion-omissions-are-regulated Published: 2026-04-07 Author: Pedro Sousa, CEO & Co-Founder If there is one thing the Revolut fine in Italy shows us, it is this: what you fail to say in a financial promotion is just as regulated as what you do say. Italy’s competition authority, the AGCM, fined Revolut €11.5 million for misleading commercial practices. The core of the finding was not that Revolut lied. It was that Revolut’s promotions created impressions (about fees, features, and account terms) that the small print did not adequately correct. The gap between what consumers understood and what they were actually signing up for was wide enough for a regulator to drive a substantial fine through. This is not a uniquely Italian story. It is a story about how financial promotions work, and how they fail. ## The Omission Problem Financial promotions law in the UK, as in most comparable jurisdictions, covers not just false statements but misleading omissions. A promotion that says something technically true but leaves out the information a consumer needs to understand what they are actually being offered is not compliant. It is misleading by omission. This is a harder discipline than it sounds. Marketing teams are trained to highlight the attractive features of a product: the zero fees, the market-rate exchange, the instant access. They are not trained to give equal prominence to the conditions that attach to those features, the circumstances under which they do not apply, or the costs that emerge when the headline terms expire or change. The result, in many financial services firms, is a structural tension. The marketing function is optimising for conversion. The compliance function is optimising for accuracy. And the gap between those two objectives is where regulatory risk lives. ## The “First Point of Contact” Standard UK financial promotions rules include a concept that the Revolut case illustrates well: the idea that the first point of contact between a firm and a potential customer carries particular weight. The FCA has long emphasised that promotions must be clear, fair and not misleading not just in their technical accuracy but in the impression they create. A consumer who sees a social media ad and forms a reasonable but incorrect understanding of what a product costs has been misled, even if the correct information was available somewhere on the website, in smaller text, on a different page. The “first point of contact” standard means that firms cannot rely on subsequent disclosure to correct an initial impression. If the promotion creates the wrong impression, later clarification is too late. The damage is done. For firms operating at scale across digital channels (social media, paid search, in-app notifications, influencer partnerships) this standard is demanding. Every touchpoint is a potential “first point of contact.” Every piece of content that could reach a new consumer is, in principle, a financial promotion. ## The Balance That Should Not Feel Impossible There is a version of the Revolut story that is simply a story about a firm moving a bit too fast in a market it did not fully understand. But there is a more useful version, which is a story about a structural problem that affects the whole industry. The structural problem is this: financial promotions compliance, done well, requires rapid review against a complex and evolving rulebook, across multiple formats and channels, with a full audit trail. Done badly (or not done at all) it generates exactly the kind of omissions that the AGCM found in Revolut’s Italian campaigns. Compliance teams at regulated firms know this. They also know that the manual processes most firms use to manage promotions review do not scale to the volume and velocity that modern financial services marketing demands. The choice between speed and compliance should not be a choice at all. It is a structural failure that firms have learned to treat as normal. The Revolut fine is a reminder of what that trade-off actually costs when it goes wrong. Not just the headline number - although €11.5 million is notable - but the reputational damage, the regulatory scrutiny, and the signal it sends to other regulators across other markets that a firm’s compliance infrastructure is not keeping pace with its growth. Getting promotions compliance right, at speed and at scale, is not a luxury for fast-moving firms. It is the table stake. --- Karavel was built to close the gap between marketing speed and compliance rigour. If you would like to understand how we help financial services firms review promotions faster without compromising accuracy, [get in touch.](#book-demo) # The ASA Reviewed the Internet in 2025. Here Is What It Found URL: https://karavel.ai/resources/asa-reviewed-the-internet-2025 Published: 2026-03-14 Author: Pedro Sousa, CEO & Co-Founder The Advertising Standards Authority has just published its 2025 Annual Report, and it is one of the more interesting regulatory reads of the year. Not because it tells a story of crisis or collapse, but because it tells a story of a system that is genuinely working, and working in ways that would have been unimaginable even a decade ago. Here are the numbers that stood out, and what they mean for anyone working in regulated financial services. ## A Regulator That Changed How It Works The headline shift in this report is not a fine or an enforcement action. It is a structural one. In 2012, the ASA spent 5% of its regulatory resources on proactive work and 67% on reactive complaints. By 2025, that had flipped: 45% proactive, 38% reactive. That is not a minor adjustment. That is a regulator that has fundamentally changed what it thinks its job is. The old model (waiting for a complaint, investigating and then producing a ruling on it) has not disappeared, but it is no longer the primary mode. The ASA is now actively scanning the market, looking for problems before consumers have to find them first. For compliance teams in financial services, this matters. Low complaint volumes no longer equal low risk. If the ASA’s systems flag your content, you may never see a complaint, you will just receive a call. ## The Scale of the Digital Advertising Market The data on digital advertising in 2025 is genuinely striking. The ASA’s Active Ad Monitoring (AAM) system scanned 60 million online ads during the year. To get a sense of the scale: that is an AI-powered system continuously sweeping social platforms, websites, and digital channels for advertising content that might breach the codes. Some of the platform-level data is particularly illuminating: - 1 in 4 Instagram posts was classified as an ad - 1 in 6 TikTok posts was classified as an ad - 34% of influencer content had no disclosure at all - The influencer compliance rate reached 57%, up from 35% in 2021 The compliance improvement on influencer disclosure is real and worth acknowledging. The industry has moved significantly in four years. But 57% also means that 43% of influencer content is still not meeting the standard, and 34% with zero disclosure is a large number by any measure. On the alcohol advertising side, around 6,000 paid-for alcohol ads were scanned. The compliance rate was 96%, which is high. Interestingly, 48% of alcohol-free ads raised potential compliance issues - a reminder that “we’re not selling alcohol” does not automatically put you outside the codes. ## Enforcement: Where the ASA Drew the Lines in 2025 Three enforcement themes dominated the year. **Weight-loss products** remained a priority, particularly in the context of injectable treatments becoming more widely available and more widely advertised. The ASA has been clear that claims about effectiveness and safety in this category require robust substantiation. **Influencer disclosure** was the year’s biggest volume story. The AAM system is specifically calibrated to detect undisclosed advertising content on social platforms, and it generated a significant proportion of the year’s proactive cases. **Green claims** produced some of the most substantively interesting decisions. The Shell and Barclays rulings reinforced something important: accurate, conditional environmental claims are viable. The issue is not whether firms can make green claims - it is whether those claims are properly qualified and substantiated. The parallel TotalEnergies ruling in Paris, where a court found the company liable for deceptive environmental advertising, adds a cross-border dimension to this debate that UK firms cannot ignore. In total, 22,383 ads were amended or removed as a result of ASA and CAP work during the year. More than 40,000 complaints were resolved, the overwhelming majority from members of the public. ## What This Means for Financial Services Firms The ASA’s direct regulatory remit over financial promotions is narrower than the FCA’s. It covers form and presentation rather than the technical financial content, which falls to the FCA. But the overlap is significant, and the direction of travel should be noted. If the regulator is scanning 60 million ads a year and operating with a 45% proactive posture, the implicit expectation is that financial services firms have their own monitoring in place. The AAM system does not distinguish between a wellness brand and an authorised firm. Both are in the scan. The data in this report describes a regulatory environment that is more systematic, more proactive, and better resourced than it was five years ago. For firms with consumer-facing marketing, that is not a warning, it is simply the new context. --- Karavel helps regulated firms review financial promotions quickly and confidently. If you would like to understand more about how our platform works, [get in touch](#book-demo).