The UK Financial Conduct Authority (FCA) has published a wide-ranging blueprint for how retail financial services should be regulated as “agentic” AI pushes firms toward near-total automation. In its landmark report, “AI and the future of retail financial services”, FCA executive director Sheldon Mills argues that the industry is moving away from human-led, episodic decisions and toward continuous services delegated to AI systems.

The review, issued as regulators grapple with the speed of generative AI deployment and growing experimentation with blockchain-based finance, frames settlement infrastructure as a key constraint. It suggests that advanced automation will require financial plumbing capable of processing transactions instantly and reliably—something the FCA implicitly contrasts with slower legacy settlement processes.

Key takeaways

• The FCA says retail financial services are shifting from human-led decisions to AI-enabled, continuous and delegated activity.
• The report calls for regulatory work that supports “agentic finance,” including trusted agent protocols.
• It highlights governance and accountability challenges that emerge when AI systems operate autonomously on consumers’ behalf.
• The FCA’s research indicates consumer openness to AI-assisted choices is already material, with 20% of UK adults reportedly willing to let AI make autonomous financial decisions.

An “autonomy spectrum” for retail finance

At the center of the Mills Review is what the FCA describes as an “autonomy spectrum,” reflecting how AI capabilities are evolving from recommending actions to executing them. Mills argues that as models become more independent, humans may increasingly shift into a more passive role—sometimes acting only as observers while AI continuously manages capital.

In the report’s framing, this evolution is not merely about better predictions. Instead, it is about delegation: systems that can be empowered, trained, and authorized to act. That creates a regulatory problem that is different in kind from earlier AI use cases, because the risk no longer rests only with advice or decision support. It rests with operational authority.

The FCA also links the acceleration of this shift to the pace of generative AI. The review notes that more than 20 “frontier models” have been released since late 2025, suggesting firms are moving faster than prior regulatory timetables.

Why settlement speed becomes a regulatory issue

The report connects autonomy to execution. The FCA says that for AI agents to carry out multi-layer transaction strategies smoothly, they require programmable and near-instant settlement mechanisms. By contrast, traditional processes with multi-day latency are described as an operational bottleneck for fully automated finance.

In the FCA’s discussion, systemic stablecoins and tokenized assets are positioned as a potential fit for this need. Because these instruments can be natively issued and transferred on programmable ledger networks, they may enable atomic settlement—where transactions are coordinated in a way that reduces friction and dependency on human clearance.

Importantly, the review does not claim that tokenized settlement is automatically “the answer.” Rather, it highlights a structural challenge: automation at the agent level will stress-test existing financial infrastructure that was built for human workflows and periodic actions.

Accountability, governance, and the “human on the hook” principle

Automation at retail scale also raises questions about who is legally responsible when agents act in unexpected ways. The Mills Review warns that allowing autonomous systems to make and execute decisions introduces severe corporate governance risks, particularly around legal accountability.

The FCA notes industry anxiety about the ambiguity of intention—whether it is possible to reliably distinguish human intent from algorithmic behavior once systems can act continuously and at speed. The report references this concern through industry commentary, including the view that the sector may eventually need something akin to a “Turing test” to separate human intent from machine-driven actions.

In separate remarks, Mills told the Financial Times ahead of the report’s release that accountability must remain anchored to humans. According to his comments, “You need a human on the hook for what they’re doing,” reflecting the FCA’s broader emphasis that operational delegation must not eliminate responsibility.

The Payments Association CEO Emma Banymandhub echoed the governance theme in a statement, saying the FCA’s review “reinforces that firms should treat agentic AI as an accountability and governance issue now,” while maintaining that governance, clear accountability, and consumer trust will determine whether AI’s potential can be realized responsibly.

Recommendations for “agentic finance” and the FCA’s AI capabilities

In its 147-page review, the FCA sets out seven recommendations it says should inform how it responds to the next phase of AI in retail financial services. Among them is an explicit push toward enabling “the foundations for agentic finance”—work intended to support trusted agent protocols that underpin how agentic AI can be used safely.

The report also flags the FCA’s internal capacity, recommending that it consider scaling up its AI Lab to support AI models and system innovation in financial services. The underlying message is that regulatory capability has to evolve alongside the technology it is designed to oversee.

That aligns with the FCA’s earlier steps. The regulator launched the review in January to examine the implications of advanced AI for consumers, retail financial markets, and regulatory oversight (as noted in the FCA’s January press materials).

Still, the report’s most practical impact may be indirect: by framing agentic AI as a continuous, delegated operating model, it implicitly increases pressure on firms to rethink not just their AI tools, but their end-to-end control framework—from authorization and monitoring to dispute handling and accountability trails.

What to watch next

As the FCA turns its recommendations into concrete regulatory expectations, the key uncertainty is how it will balance innovation with enforcement—especially in cases where AI agents execute transactions at speed. Investors and builders should watch for clearer standards on governance, accountability, and how (or whether) tokenized settlement infrastructure fits into the FCA’s definition of safe, trusted automation.

The United Kingdom’s Financial Conduct Authority (FCA) has issued a broad regulatory blueprint for retail financial services, warning that retail financial services are hurtling toward total automation driven by autonomous “agentic AI.”

The landmark report, “AI and the future of retail financial services,” spearheaded by executive director Sheldon Mills, details a structural shift away from periodic, human-led decisions toward continuous, automated financial services that could increasingly rely on programmable financial infrastructure.

“The central shift is from human-led, episodic financial activity towards services that are AI-enabled, continuous and delegated,” Mills wrote. In January, the FCA launched a review into the implications of advanced AI on consumers, retail financial markets and regulators.

The 147-page report comes at an inflection point where generative AI meets institutional crypto adoption. As financial systems transition to autonomous portfolio and cash management, legacy fiat banking rails are seen as structurally incapable of matching machine transaction speeds. This positions systemic stablecoins and tokenized bank deposits as potential settlement infrastructure for AI-driven financial services.

It outlines seven recommendations for the FCA to consider, including enabling “the foundations for agentic finance,” which would support the development of trusted agent protocols that would underpin use of agentic AI and “scaling up the FCA’s AI Lab to support AI models and system innovation in financial services.”

Related: UK plans payments rule changes for stablecoins, tokenized deposits

FCA envisions agents on “autonomy spectrum”

The Mills Report suggests that the catalyst is the rapid evolution of AI from predictive models into independent agents operating on an “autonomy spectrum.” At the far end of this spectrum, humans act as mere “observers” while AI continuously manages capital.

Screenshot of table header that sets out how FCA sees operator activities may change as they move across the AI autonomy spectrum. Source: Financial Conduct Authority.

The acceleration of this shift has outpaced prior regulatory timelines, with more than 20 frontier models released since late 2025 alone.

“Firms are moving from systems that recommend actions to systems empowered and trained to take them, and consumers will soon gain agents that act on their behalf,” Mills said in the report’s foreword. FCA research shows that 20% of UK adults are already open to letting AI make autonomous financial choices.

For these AI agents to execute multi-layered transaction strategies seamlessly, they require programmable, instantaneous settlement mechanisms. Traditional multi-day settlement latency remains an operational bottleneck. Because systemic stablecoins and tokenized assets live natively on programmable ledger networks, they provide the friction-free, atomic settlement needed for automated protocols to move capital instantly without human clearance.

However, this automation introduces severe corporate governance risks regarding legal accountability.

The review highlights growing industry anxiety over this ambiguity, noting that one CEO observed that the financial sector may eventually require a “Turing test” to accurately distinguish between human intent and autonomous algorithmic behavior in the market.

“The FCA’s Mills Review reinforces that firms should treat agentic AI as an accountability and governance issue now, while providing greater confidence to innovate responsibly as AI adoption accelerates,” Emma Banymandhub, CEO of The Payments Association, said in a statement. “AI has enormous potential for financial services, but realising that potential will depend on strong governance, clear accountability and maintaining consumer trust.”

Mills, who is leaving after eight years at the FCA, told The Financial Times ahead of the report’s release that managers would still need to be accountable for the actions of their AI models. “You need a human on the hook for what they’re doing,” he said.

Magazine: AI is banking the unbanked in Africa… faster than crypto

Ethereum co-founder Vitalik Buterin has unveiled a multi-year roadmap that places native privacy, quantum resistance, and protocol simplification at the center of Ethereum’s next major upgrade, describing it as the network’s largest transformation since The Merge.

According to a roadmap published on Strawmap.org and shared by Buterin on X over July 6, the proposed changes are expected to be introduced over the next three to four years following discussions among Ethereum researchers in Berlin.

The document outlines coordinated upgrades spanning nearly every layer of the network and presents what Buterin describes as Ethereum’s third major evolution after its transition to proof-of-stake in 2022.

Native privacy becomes a core protocol feature

Instead of leaving privacy to applications built on Ethereum, the roadmap proposes making it a built-in property of the protocol itself. The document evaluates key components, including Frames, the transaction mempool, and future state designs, according to whether they can support intermediary-free, quantum-safe privacy while keeping computational costs low.

Building on ideas first outlined in May 2026, Buterin’s latest proposal expands an earlier privacy roadmap into a network-wide redesign. What previously focused on incremental improvements has now developed into a long-term architectural plan covering the protocol’s core infrastructure.

Among the document’s strongest statements is Buterin’s observation that “quantum safety has shifted up a LOT in priority.” The roadmap identifies work on quantum-safe blob designs, which support Ethereum’s rollup-based scaling model, as an urgent priority.

According to the proposal, several cryptographic systems currently used by Ethereum, including BLS signatures, KZG commitments, and ECDSA, would eventually be replaced with post-quantum alternatives. The direction aligns with the post-quantum cryptography standards finalized by the U.S. National Institute of Standards and Technology in 2024.

Protocol redesign targets faster verification and smaller overhead

Alongside cryptographic upgrades, the roadmap introduces changes intended to simplify how Ethereum validates transactions. Rather than requiring every node to re-execute every transaction, the proposal recommends recursive STARK-based verification, where one prover performs the intensive computation while the rest of the network verifies a compact cryptographic proof.

The proposal also continues work first discussed by the Ethereum Foundation earlier this year. In February 2026, the Foundation released an initial strawmap examining quantum threats facing Ethereum, while Buterin separately detailed the network’s quantum security risks. The latest roadmap develops those earlier discussions into a more detailed implementation strategy.

Meanwhile, the technical proposal arrives as the Ethereum Foundation continues internal restructuring. The organization has reduced its workforce by roughly 20%, eliminating about 54 positions, while also cutting its budget by a targeted 40%. Recent departures have included protocol contributors Hsiao-Wei Wang, Tomasz Stańczak, Tim Beiko, and Barnabé Monnot.

Community discussion on X has largely focused on the roadmap’s technical detail rather than broad ambitions. Several participants noted that the draft identifies specific signature schemes, cryptographic replacements, and state-size objectives instead of relying on high-level goals.

For now, the roadmap remains a working draft rather than a finalized implementation schedule. According to the document, the upcoming Hegotá fork is expected to be the final major network upgrade before Ethereum enters what Buterin describes as the Lean Ethereum era, where privacy, scalability, and quantum resistance are treated as core protocol requirements rather than optional additions.

Ben-Sasson also welcomed Buterin’s decision to make privacy and quantum-resistant cryptography top priorities.

“Quantum safety—excellent,” he wrote on X. “Glad to see this as a high priority.”

But he argued Ethereum shouldn’t wait three to four years to get there.

“‘3-4 years’ as the timeline is way too long,” Ben-Sasson said. “Especially for quantum readiness.”

Former Ethereum Foundation researcher Dankrad Feist struck a similar tone. Calling the roadmap’s vision “really cool,” Feist said on X that features like near-instant transaction finality and dramatically higher throughput could transform the network.

His biggest concern, however, was speed. “But 3-4 years is very slow,” Feist wrote. “I think we should be ambitious and get it done in ~1 year.”

Feist even suggested recent advances in AI tools, including large language models, could help accelerate development.

Not every discussion centered on timing. Some researchers dug into the roadmap’s technical details.

Ben-Sasson questioned one of Buterin’s proposals to introduce new types of blockchain “state,” essentially the data Ethereum stores about accounts, balances and smart contracts.

“New kinds of state: what does that mean? Who is affected by it?” he asked, calling for more explanation.

Meanwhile, Ethereum Foundation researcher Barnabé Monnot focused on how the roadmap had changed from an earlier version released in February.

You think you research crypto brands on Forbes or CoinDesk. You don’t. You ask ChatGPT. And ChatGPT already picked its favorites.

The Gatekeeper Nobody Noticed

For twenty years, if you wanted to know whether a financial brand was trustworthy, you went to established media. Forbes. The Wall Street Journal. Bloomberg. CoinDesk for crypto specifically.

Those outlets shaped perception. Their coverage decided who was credible and who wasn’t. A positive Forbes profile moved markets. A negative Bloomberg investigation destroyed reputations.

That era just ended.

On July 2, 2026, PR firm 5W released The Crypto Trust Index, research scoring how ChatGPT, Claude, Perplexity, Gemini, and Google AI answer when a first-time buyer asks whether a crypto brand is safe.

The finding: AI engines do not stay neutral. They answer with a verdict, recommend, hedge, or warn.

There is no neutral tier.

The gatekeeper didn’t disappear. It just moved. And almost nobody noticed.

What The Research Actually Found

5W ran 60+ first-time-buyer prompts across six question types, five times per engine, across 25 crypto exchanges and brands.

The results are unambiguous:

Coinbase Trust Score 94. The default first-time-buyer recommendation across all five AI engines.

Kraken 87. Cited for proof-of-reserves and clean operating record.

Fidelity Crypto 82. Inherited trust from a traditional-finance brand the engines extend without re-litigation.

Gemini 77. Recommended on US regulatory posture.

And then there’s everyone else. Hedged. Warned about. Or simply absent from AI recommendations entirely.

If your brand doesn’t appear in an AI recommendation, you don’t exist for the majority of first-time buyers. Not because you’re not listed on Google. Because the AI didn’t mention you.

Why This Changes Everything About Crypto Marketing

Traditional crypto marketing operated on a simple premise: get coverage, build awareness, convert buyers.

The funnel looked like this:

• User hears about crypto
• User Googles crypto brand
• User reads reviews on media sites
• User decides to buy or not

AI just collapsed that funnel into a single step:

• User asks ChatGPT “is [brand] safe?”
• ChatGPT answers with a verdict
• User acts on the verdict

The media layer, the reviews, the coverage, the PR campaigns, still exists. But it now feeds AI training data rather than reaching users directly.

Your brand’s reputation isn’t built on the Forbes article anymore. It’s built on what Forbes said that the AI learned from.

And you can’t see what the AI learned. You can only see the verdict it delivers.

The Uncomfortable Finding AI Inherited Traditional Finance Bias

One of the most significant findings from the 5W research: brokerage-backed crypto brands inherit trust the engines extend on day one.

Fidelity Crypto scored 82, not because Fidelity has been in crypto longer or built better technology than native crypto companies. But because Fidelity has decades of traditional finance credibility, and AI engines absorbed that credibility without question.

This is a structural bias baked into how AI systems were trained. They learned from the same financial media, regulatory filings, and institutional coverage that already favored traditional finance brands.

So when a first-time buyer asks ChatGPT whether Fidelity Crypto is safe, ChatGPT answers with the confidence that comes from decades of institutional reputation, even if Fidelity’s crypto product launched last year.

Meanwhile, a crypto-native brand that’s been operating for ten years, has proof-of-reserves, and has never been hacked starts from zero in AI trust scores if it doesn’t have extensive mainstream media coverage.

The playing field isn’t level. The AI decided who starts with an advantage.

The New Rules Of Crypto Marketing

The 5W research identified exactly what moves AI trust scores:

What works:

• Proof-of-reserves (auditable, citable, verifiable)
• Clean regulatory record (no enforcement actions, no major incidents)
• Traditional finance association (legacy credibility transfers automatically)
• Mainstream media coverage (feeds the training data that shapes AI responses)

What doesn’t work:

• Marketing campaigns (AI doesn’t factor in ad spend)
• Social media presence (follower counts don’t move trust scores)
• Community building (Discord size is invisible to AI trust evaluation)
• Sponsored content (AI learns to discount promotional material)

This is a fundamental inversion of how crypto brands have marketed themselves for the last decade.

The entire playbook, build community, drive social engagement, get influencer coverage, sponsor events, doesn’t register in AI trust evaluations.

A clean audit moves your score. A viral tweet doesn’t.

Why “The On-Ramp Moved” Is The Most Important Sentence In Crypto Marketing Right Now

5W’s research includes one line that should be required reading for every crypto marketer:

“The on-ramp moved. The first trust decision now happens inside an AI answer, before a buyer reaches any site.”

Let that land.

Before a first-time buyer visits your website, reads your whitepaper, checks your social proof, or sees your ad, they’ve already formed an opinion based on what an AI told them.

If the AI recommended you, they arrive pre-convinced.

If the AI hedged on you, they arrive skeptical.

If the AI warned about you, they don’t arrive at all.

Your entire marketing infrastructure, your website, your content, your community, your influencer partnerships, is operating on users who’ve already been filtered by an AI they didn’t realize was making a judgment call.

The Winners And Losers This Creates

Established brands with clean records win.

Coinbase at 94 doesn’t need to convince AI engines. Years of regulatory engagement, mainstream media coverage, and transparent operations built a reputation that AI absorbed and now distributes to every first-time buyer who asks.

Traditional finance entrants win immediately.

Fidelity at 82 on day one of its crypto product. The institutional credibility transfers. No crypto track record required.

Crypto-native brands without mainstream coverage lose.

If your brand built its reputation on crypto Twitter, crypto influencer networks, and community Discord servers, none of that feeds AI training data in a way that builds trust scores. You’re invisible to the new gatekeeper.

Brands with any regulatory history lose badly.

The AI doesn’t forget. A 2019 enforcement action, a 2021 hack, a 2023 regulatory warning, all of it is in the training data. All of it surfaces when a first-time buyer asks if you’re safe.

The AI doesn’t distinguish between “resolved issue from five years ago” and “current problem.” It just knows it exists.

What This Means For Crypto’s Future

The trust infrastructure of crypto just got centralized, not by a government, not by a regulator, but by AI systems trained on data those systems chose.

This is ironic in the deepest possible way. Crypto was supposed to be decentralized. Trustless. Permissionless. No gatekeeper deciding who’s legitimate.

But now five AI engines, ChatGPT, Claude, Perplexity, Gemini, Google AI, are making trust judgments about 25 crypto brands, and their answers are shaping where billions of dollars flow.

That’s not decentralization. That’s a new kind of centralization that’s harder to see because it’s embedded in a conversational interface rather than a regulatory filing.

And unlike a regulator, you can’t appeal an AI trust score.

The Marketing Question Nobody’s Asking Yet

If AI engines are the new gatekeepers of crypto trust, and those engines are trained on data that favors traditional finance brands and mainstream media coverage, what happens to the next wave of crypto innovation?

A genuinely novel DeFi protocol, a new blockchain architecture, an innovative crypto product that launches without institutional backing or mainstream media coverage, starts from zero in AI trust scores.

Not because it’s less safe. Not because it’s less innovative. But because the AI hasn’t seen enough reliable coverage of it to form a verdict.

The early-stage crypto project that needs trust the most, to attract first-time users, to build legitimacy, to grow, is the one that AI trust evaluation helps the least.

That creates a structural disadvantage for innovation and a structural advantage for incumbents.

Which is exactly what crypto was supposed to fix in finance.

What Comes Next

Every crypto brand needs to answer a new strategic question: What does ChatGPT say about us when a first-time buyer asks if we’re safe?

Not “what does our marketing say.” Not “what do our users say on Twitter.” What does the AI say?

Because that’s the first answer the next billion users are going to get.

And the brands that understand this shift, that start building for AI trust evaluation rather than traditional marketing metrics, will have an enormous advantage over the next three years.

The gatekeeper moved. Most crypto brands are still marketing to the old one.

Ask ChatGPT right now if your favorite crypto brand is safe. Then tell me what it said, because that answer is shaping more buying decisions than any ad campaign you’ve ever seen.

Bitcoin (BTC) saw flash volatility into Monday’s Wall Street open as markets reacted to tech company Strategy’s new BTC sales.

Key points:

• Bitcoin reacts sharply to news that Strategy had sold nearly 3,600 BTC.
• A rebound during the US trading session failed to recoup more than half of the day’s losses.
• Strategy may reveal a compensatory BTC buy, an analyst suggests.

Bitcoin erases holiday gains on Strategy sale

Data from TradingView showed BTC/USD dropping to near $61,000, sparking daily losses of more than 4%.

BTC/USD four-hour chart. Source: Cointelegraph/TradingView

A rebound at the start of the US session pushed the price higher before settling around the $62,000 mark at the time of writing.

Strategy revealed that it sold 3,588 BTC through July 5 to fund preferred stock dividend payments and replenish cash reserves.

Commenting on the latest BTC price moves, X commentator Exitpump suggested that the Strategy news was the catalyst for an already weakening market.

“Bearish signs were there, posted about it yesterday, news about Saylor selling just triggered more dump,” they wrote. 

“Funding is still pretty positive. That was it i guess. Short term bounce from 61.2k and then more dump imo.”

Exitpump referred to funding rates across exchanges, with a post on Sunday eyeing a buyer entity using a time-weighted average price (TWAP) method to add exposure.

“Once the TWAP buyer backs off, I wouldn’t be surprised to see a fast flush lower,” they wrote, anticipating a price ceiling at $64,000.

BTC chart with funding rate data. Source: Exitpump/X

Trader and analyst Rekt Capital appeared unsurprised by the behavior, reiterating similarities between current price action and the latter portion of the 2022 bear market.

“Generally, Bitcoin is doing the same exact thing now as it was doing in the Summer of 2022,” he told X followers.

An accompanying chart showed the 50-month exponential moving average (EMA) trend line potentially becoming new resistance, just like four years ago.

BTC/USD one-month chart with 21, 50EMA. Source: Rekt Capital/X

Analyst: Strategy may reveal more BTC buys

Others remained upbeat, with trader Jelle eyeing bullish divergences on weekly time frames on the BTC/USD relative strength index (RSI).

Related: $60.4K Becomes ‘most important area’: Five things to know in Bitcoin this week

“I have seen the $BTC chart look much worse than this over the years,” he argued.

BTC/USDT one-week chart with RSI data. Source: Jelle/X

As Cointelegraph continues to report, various onchain indicators have printed reversal signals absent since late 2022.

Crypto trader and analyst Michaël van de Poppe, meanwhile, suggested that Strategy itself could end up delivering a market rebound.

“The markets are reacting with a shock response to this news. $BTC drops, and it’s clearly valuing the potential impact that Strategy can continue to sell Bitcoin going forward,” he wrote on X. 

“However, I wouldn’t be surprised to see a message in the coming days that they’ve been buying more $BTC than they’ve sold.”