| |

This Week in AI in Marketing & Management (1st June)

ByAnn Stanley

Anthropic Launched Opus 4.8 and Raises $65bn, as the AI Jobs Debate Evolves

The week belonged to Anthropic, which launched Claude Opus 4.8 alongside a new Dynamic Workflows tool for Claude Code, then raised a record $65bn at a $965bn valuation to become the most valuable AI company in the world ahead of an expected IPO. Apple was reported to be shrinking Google’s Gemini model to run a new Siri on the iPhone, and OpenAI moved into performance advertising with conversion-focused ads inside ChatGPT. On the management side, the mood shifted: Nvidia’s Jensen Huang and OpenAI’s Sam Altman publicly softened their warnings about AI job losses, even as Groupon, Cloudflare and others announced AI-linked redundancies. Agentic commerce kept accelerating, with Amazon selling its AI shopping tools to rival retailers, Alipay launching an AI wallet, and Visa and Mastercard staking out the payment rails. Here is the full briefing for the week ending 31 May 2026.

Table of Contents

AI News, Tech & Tools

AI in Marketing

AI in Management

AI in E-commerce, Retail and Agentic Commerce

AI for Other Sectors and Industries

Key Takeaways

Frequently Asked Questions

Conclusion

AI News, Tech & Tools

Anthropic launches Claude Opus 4.8

Sources: anthropic.com, techcrunch.com, zdnet.com, forbes.com | 24-28 May 2026

Anthropic released Claude Opus 4.8 on 28 May, an upgrade to its top model that improves on Opus 4.7 across coding, agentic tasks (work the AI carries out by itself, step by step, rather than waiting for a prompt at each turn) and professional knowledge work, at the same price as the previous version. The headline change for working teams is effort control. Users on claude.ai can now decide how much effort Claude puts into a task, turning it up for hard problems where accuracy matters and turning it down when speed matters more. That single dial gives a marketing team a practical way to balance cost, speed and depth on the same model. It launched alongside new agent tooling for Claude Code, which we cover next. In a related move aimed squarely at smaller firms, Forbes reported that Anthropic has introduced “Claude for Small Business”, which connects Claude directly into everyday tools such as QuickBooks, PayPal and HubSpot rather than leaving it as a separate chat window.

Anthropic also made the fast version of Opus 4.8, which runs at around 2.5 times the speed, three times cheaper than fast mode on earlier models, and doubled the usage limits on Claude Code so teams can run more work in a given period. ZDNET led its coverage on a different angle. The model is being positioned as more honest and more willing to admit uncertainty, which the publication framed as the standout feature for anyone relying on AI for serious analysis rather than a quick draft.

Why it matters
For business owners, the change that earns its keep is not the benchmark score, it is the combination of effort control, a cheaper and faster mode, and a model that says when it is unsure. Those three together make the same AI usable for both a fast first draft and a careful piece of analysis without paying top rates for everything. At Anicca this matters directly: we build client audits, reporting dashboards and content pipelines on multi-agent workflows, where several AI agents work at the same time on different parts of a job, and a model that admits uncertainty is exactly what keeps a confident but wrong answer from one agent quietly distorting a whole report.

Claude Code’s Dynamic Workflows can run tens of agents on one job

Sources: claude.com, code.claude.com, techcrunch.com | 28 May 2026

On 28 May 2026 Anthropic released Dynamic Workflows in Claude Code, a research preview launched alongside its new Opus 4.8 model. In plain terms, it is a coordinating script that reads your instruction, plans the work itself, and then sets many subagents (smaller AI helpers, each handling one part of the task) to work in the background. The subagents run in parallel, meaning at the same time rather than one after another, while your main session stays free for you to keep working. It is built for large, long-running jobs that previously took weeks, and it requires Claude Code version 2.1.154 or later.

The scale is the notable part. A single run can have up to 16 subagents working at once, and up to 1,000 in total across the whole job. The method is deliberately adversarial: some agents tackle the problem from independent angles, while others try to refute or check what the first group found, and the run keeps repeating until the answers converge on a reliable result. Anthropic names codebase audits, large software migrations and cross-checked research as early use cases. In one example, a full rewrite of the Bun software project produced roughly 750,000 lines of code in eleven days. It is available on all paid plans, the Anthropic API, Amazon Bedrock, Google Cloud Vertex AI and Microsoft Foundry.

Why it matters
This is the same pattern Anicca already uses to build client audits, reporting dashboards and content pipelines: many AI agents splitting a large job between them and checking each other’s work. Dynamic Workflows makes that approach faster and more reliable, because the refute-and-converge method reduces the risk of a single agent presenting a confident but wrong answer, which is the main concern when AI handles real client data. For business leaders, the lesson is that big, repetitive analysis tasks no longer need a large team or several weeks. Pick one job that currently ties up people for days, such as a full content review or a competitor audit, and ask your technical team whether it could be run this way instead.

Anthropic raises $65bn at a $965bn valuation to become the most valuable AI company

Sources: anthropic.com, techcrunch.com, axios.com, fortune.com | 28 May 2026

Anthropic announced a $65bn Series H funding round on 28 May at a $965bn post-money valuation, the largest private funding round on record and enough to make it the most valuable AI startup in the world, ahead of OpenAI’s most recent $730bn mark. The round was led by Altimeter, Dragoneer, Greenoaks and Sequoia, and co-led by a long list including Capital Group, Coatue, GIC and ICONIQ. It also folds in $15bn of previously committed investment from cloud providers, including $5bn from Amazon.

The company said its run-rate revenue crossed $47bn earlier in the month, a figure that has grown sharply as enterprise adoption of Claude has spread. Anthropic said the money will go toward safety and interpretability research, expanding the compute needed to meet demand for Claude, and scaling the products and partnerships its customers depend on. Coverage from Fortune and Axios framed the round as Anthropic nudging toward a $1tn valuation ahead of a likely public listing.

Why it matters
For business leaders the number itself is less important than what it signals: the model behind a growing share of corporate AI work is now one of the most valuable companies on the planet, and it is being funded to keep scaling for years. That stability matters when you are deciding which AI provider to build on. It also sharpens the strategic question for every business: if Claude, ChatGPT and Gemini are all backed to expand aggressively, the cost of standing still rises each quarter. The agencies and in-house teams that have already built working AI into their operations are pulling further ahead of those still treating it as an experiment.

Apple is trying to shrink Google’s Gemini to power a new Siri on the iPhone

Sources: arstechnica.com, techcrunch.com | 28 May 2026

Ahead of Apple’s WWDC developer conference in June, reports said Apple is working to compress Google’s very large Gemini model so that a meaningfully smarter Siri can run on the iPhone itself, with a cloud component still likely for the heaviest requests. Separately, Bloomberg published leaked renders of a planned standalone Siri app that looks built to take on ChatGPT directly, with a conversational interface rather than the old command-and-response model.

The reports point to Apple leaning on Google’s model rather than a fully home-grown one, a pragmatic move after a difficult couple of years for Apple’s own AI efforts. Running a distilled version on-device would let Apple keep more requests private and fast, while reserving the cloud for complex tasks.

Why it matters
If a far more capable Siri ships to a billion iPhones, a large slice of everyday search, shopping and messaging will start to flow through an on-device assistant before a person ever opens a browser or an app. For marketers that is the same shift Gemini and ChatGPT are already driving, now reaching the most loyal mobile audience in the world. The practical task does not change: make sure your brand, products and offers are clear, structured and machine-readable, because an assistant will increasingly be the first thing that reads them on the customer’s behalf.

The Economist launches the first ChatGPT app from a major consumer news publisher

Source: niemanlab.org | 27 May 2026

The Economist launched its own app inside ChatGPT, the first of its kind from a major consumer news publisher, according to Nieman Lab. Called “The Economist – Graphs”, it runs natively in ChatGPT and lets users interact with the publication’s data visualisations. At launch it focuses on US polling data, letting people ask questions about The Economist’s approval-rating tracker and see the answers broken down by state, demographic and issue.

The move is a notable bet on distribution inside an AI assistant rather than only on a publisher’s own site or app. Instead of waiting for ChatGPT to summarise its journalism second-hand, The Economist is putting a controlled, branded experience directly where readers are increasingly asking their questions.

Why it matters
This is an early template for any brand that owns useful data or tools, not just publishers. As people spend more time inside ChatGPT, Gemini and Claude, the brands that build a genuine app or connector into those assistants get a branded presence in the conversation rather than being reduced to an unattributed answer. If your business has its own data, calculators or product tools, it is worth asking now whether they should live inside an AI assistant as well as on your website. First movers in each category will set the expectation.

AI in Marketing

OpenAI is preparing conversion-focused ads for ChatGPT

Source: searchengineland.com | Anu Adegbola | 26 May 2026

OpenAI is moving beyond brand-awareness advertising into performance marketing, with conversion-focused ad formats built to drive measurable actions, according to reporting in The Information relayed by Search Engine Land. The company has reportedly been telling advertisers and ad-tech firms that it wants to attract smaller local businesses, naming examples such as dry cleaners, car washes and appointment-based services. The planned formats are designed to encourage purchases, appointment bookings and contact-form submissions.

Crucially, advertisers testing the formats would only pay when those actions happen, bringing ChatGPT advertising close to the pay-for-results model that defines Google and Meta performance campaigns. That positions OpenAI to compete directly for the small and mid-sized advertiser budgets that have funded Google Ads for two decades.

Why it matters
A pay-per-action ad system inside ChatGPT would be a genuine new performance channel, not an experiment. For local and small businesses in particular, it could become a serious alternative to Google Ads for leads, bookings and sales. Marketers should start watching for the beta, prepare clean conversion tracking, and think about how their offers would read inside a conversation rather than a search results page. The brands that learn the new format early, as they did with early Google Ads, will get the cheapest results before competition bids the cost up.

Google folds Display Ads into Demand Gen

Source: blog.google | 26 May 2026

Google announced that Display Ads now have a new home inside Demand Gen campaigns. Advertisers can still serve ads exclusively on the Google Display Network if they want to, but the management of those campaigns now runs through the more streamlined Demand Gen interface, which Google has been pushing as its AI-driven, visually rich format across YouTube, Discover, Gmail and the Display Network.

The change continues Google’s steady consolidation of its older, manually managed ad types into fewer AI-optimised campaign formats. For advertisers used to running standalone Display campaigns, it means learning the Demand Gen controls, audience signals and creative requirements, while keeping the option to limit delivery to Display inventory only.

Why it matters
This is part of a clear pattern: Google is retiring granular manual controls in favour of AI-led campaign types, and Demand Gen is now the home for upper-funnel visual advertising. Marketers should not treat this as a simple rename. The targeting, bidding and creative best practices inside Demand Gen are different from classic Display, so audit any Display budget, move it deliberately rather than by default, and test creative built for the format rather than recycling old banners. Treat it as a chance to refresh tired Display activity, not just a forced migration.

Google brings Preferred Sources into AI search, and warns against buying brand mentions

Sources: martech.org, seroundtable.com | 28-29 May 2026

Google is bringing Preferred Sources into its AI search experience, giving people a way to tell Google which publishers they want to see more of, and giving those publishers a new route to build loyalty-driven visibility inside AI Overviews and AI Mode, according to MarTech. The week’s Search Engine Roundtable recap added an important warning: Google has strongly cautioned against buying brand mentions or otherwise trying to manipulate how often a brand appears in AI responses, alongside a status update on the slow-moving May 2026 core update.

Together the two stories show Google trying to shape behaviour as AI search matures. Preferred Sources rewards genuine audience loyalty, while the warning against paid mentions signals that Google will treat attempts to game AI citations the same way it treats manipulative link building.

Why it matters
The takeaway for brands chasing visibility in AI answers is to build it honestly. Encourage your real audience to set you as a preferred source where that option exists, and invest in the content quality and authority that earns genuine citations, rather than paying for brand mentions that Google has now explicitly warned against. Any agency selling guaranteed AI mentions is selling something Google intends to penalise. The durable approach is the same as good SEO: be the genuinely useful, authoritative answer.

How to use digital PR and B2B PR to shape what AI recommends

Sources: martech.org, searchenginejournal.com, semrush.com | 25-29 May 2026

Two complementary pieces this week argued that public relations is becoming one of the most effective ways to influence what AI systems say about a brand. MarTech explained that AI answers often show only a handful of vendors, and that PR can shape not just whether a brand appears but how it is framed, compared and recommended during the buying process. Search Engine Journal made the wider point that digital PR has not fundamentally changed, AI search has simply made the fundamentals, earned coverage, credible third-party mentions and authority, more important than ever.

The shared message is that AI models draw heavily on what reputable sources say about a company. Brands that earn coverage in trusted publications, build a consistent factual presence, and are described accurately across the web give the models better material to quote. Those that rely only on their own website give the models little to work with. The platforms are responding too: Semrush has unified its enterprise SEO and AI-search content tools into shared workflows, a sign that earning visibility in classic search and in AI answers is increasingly one job rather than two.

Why it matters
This reframes PR as a direct input into AI visibility, not just a reputation activity. When a buyer asks ChatGPT or Gemini to recommend a supplier, the answer is assembled partly from third-party coverage and mentions. Marketing and PR should now work as one: pursue earned coverage, keep your facts consistent everywhere they appear, and make sure independent sources describe your strengths accurately. The work that built brand reputation for humans now also trains the AI tools that increasingly make the shortlist.

AI in Management

AI leaders soften their warnings on job losses

Source: france24.com | 28 May 2026

Some of the most prominent figures in AI are stepping back from earlier warnings about mass unemployment, as the industry faces growing public hostility over how AI is changing work. Nvidia chief executive Jensen Huang and OpenAI chief executive Sam Altman, whose past comments fuelled much of the anxiety, are now arguing that the doom-laden predictions were overblown. Speaking to Channel News Asia, Huang took direct aim at executives blaming AI for redundancies: “The narrative that connects AI to job loss, for many of the CEOs that are doing it, it is just too lazy. AI has just arrived. How is it possible they are already losing jobs?”

Huang has long argued that AI will create roughly as many jobs as it displaces, and his comments push back on the idea that current redundancies are genuinely AI-driven rather than ordinary cost-cutting wearing an AI label. The softening tone marks a notable shift from the apocalyptic framing that dominated much of the previous year.

Why it matters
For business leaders, this is a useful corrective to the headline panic. The most credible voices in AI are now saying that mass overnight job losses are not what is actually happening, and that many companies citing AI for redundancies are using it as cover. The honest management position sits in the middle: AI is genuinely changing what work looks like, but the near-term story is role redesign, not wholesale replacement. Plan for your teams to do different work with AI, resist the temptation to use AI as an excuse for cuts you cannot evidence, and be ready to explain your own AI-and-headcount decisions credibly.

Groupon, Cloudflare and the growing list of AI-linked layoffs

Sources: businessinsider.com, technologymagazine.com, peoplemattersglobal.com | 27-28 May 2026

Even as AI leaders softened their warnings, the redundancy announcements kept coming. Business Insider’s running list now tracks 14 companies, including Wix and Coinbase, that have publicly tied job cuts to AI. Groupon confirmed it is cutting up to 400 jobs worldwide across HR, customer service, software engineering and operations as it rebuilds the business around automation and AI-driven workflows. Cloudflare announced reductions affecting 20% of staff, with chief executive Matthew Prince saying in a Wall Street Journal op-ed that the “vast majority” of those affected were what he called “measurers”, people in middle management, finance, legal and internal audit.

Notably, Prince framed the Cloudflare cuts as being about how a high-growth company creates value in the agentic AI era rather than as simple cost-cutting, and said the company is still hiring “builders and sellers”. US job-cut data listed AI as the leading stated cause of reductions in both March and April, though, as Huang argued, the real drivers are often harder to separate from ordinary restructuring.

Why it matters
The pattern worth noting is which roles are being cut. Cloudflare’s “measurers” framing, middle management and oversight functions, points to where leaders believe AI removes coordination overhead first. Whether or not you agree, the message to managers is clear: roles that mainly monitor, summarise and pass information along are the most exposed, while roles that build, sell and make judgement calls are being protected. If your team’s work is largely coordination, the priority is to move it up the value chain, toward the judgement and creativity that AI still cannot replace.

How not to announce AI-driven redundancies

Sources: hcamag.com, insurancebusinessmag.com | 25-26 May 2026

Human Resources Director and Insurance Business both examined a string of corporate crises in which AI-driven redundancy announcements went badly wrong, including one where a chief executive who announced large job cuts later faced threats to his own family. The coverage frames these as a masterclass in how not to handle AI-linked restructuring: cuts announced bluntly, justified vaguely with “AI made us do it”, and communicated without dignity or a credible plan for the people affected.

The analysis argues that genuine AI transformation in the first 12 to 18 months tends to look like role redesign rather than mass headcount reduction, with displacement and new hiring happening at the same time across different functions. Where leaders treat AI as a one-line excuse for cuts, they damage trust internally and invite a public backlash that can dwarf any saving.

Why it matters
Whatever the scale of your own AI plans, how you communicate them is now a management risk in its own right. Staff, customers and the press are primed to be sceptical of “AI made us do it” framing. If AI genuinely changes roles in your business, say specifically what is changing and why, show the redesign rather than just the reduction, and treat affected people with care and transparency. A clumsy AI-redundancy announcement can cost more in reputation and morale than it saves in payroll.

KPMG: CEOs bet on human-AI collaboration, not replacement

Source: kpmg.com | 27 May 2026

KPMG’s CEO Outlook research, highlighted this week in its insurance-sector analysis, found that chief executives remain confident about growth and frame AI primarily as a route to smarter risk management, operational efficiency and human-AI collaboration rather than straightforward staff replacement. The framing is a useful counterweight to the layoffs headlines: at board level, the dominant story is using AI to make existing teams more effective, not to empty the building.

The insurance findings echo a broader theme across KPMG’s work, that the leaders seeing the strongest returns are pairing AI with human judgement in regulated, high-stakes processes rather than handing those processes over wholesale. Confidence in growth is tied to disciplined adoption, not to cutting people.

Why it matters
This is the version of the AI workforce story that most businesses should actually plan around. The realistic near-term gain is not a smaller team, it is a more capable one, with AI handling the repetitive analytical work and people focusing on judgement, relationships and decisions. For marketing leaders, that means investing in your team’s AI fluency and redesigning workflows so AI and people each do what they are best at. The companies treating AI as a collaboration tool are quietly outperforming those treating it mainly as a way to cut costs.

AI in E-commerce, Retail and Agentic Commerce

Amazon starts selling its AI shopping tools to rival retailers

Source: modernretail.co, practicalecommerce.com | 28 May 2026

Amazon has begun selling the technology behind its AI shopping assistant to other retailers, according to Modern Retail, opening up the AI tools it built for its own marketplace as a product that rival shops can buy. The move turns Amazon’s heavy internal investment in conversational shopping, product discovery and recommendation into a new revenue line, much as it did years ago when it turned its internal infrastructure into Amazon Web Services.

For retailers, the appeal is obvious: rather than build an AI shopping assistant from scratch, they can license a proven one from the company that has processed more shopping behaviour than anyone. The catch is equally clear, handing a competitor visibility and influence over your customer experience, and deepening your dependence on Amazon’s ecosystem.

Why it matters
Retailers now face a genuine make-or-buy decision on AI shopping tools, and Amazon has made “buy” much easier. The trade-off is strategic, not just technical: licensing Amazon’s assistant gets you capable AI quickly, but it embeds a major competitor inside your customer journey and your data. Smaller retailers on platforms such as Shopify and Magento may be better served by the AI tools their own platform offers, which keep more of the customer relationship in their hands. Practical Ecommerce makes a useful related point: AI is changing how customers discover products, not how loyal they stay, so the priority is being findable by AI shopping tools without surrendering the relationship that keeps customers coming back. Weigh speed against control before signing up to a rival’s technology.

A field guide to the new agentic advertising and commerce protocols

Source: adweek.com, retail-week.com | 26-28 May 2026

Adweek published a useful guide to the fast-growing maze of technical standards that now govern how AI agents buy ads, make payments and complete transactions, names such as MCP, A2A, AdCP, UCP and TAP. The piece notes that as the leading AI labs’ reasoning models have grown better at following instructions and completing multi-step processes, demand for autonomous agents has surged, and a handful of these protocols are becoming the de-facto standards for how bots transact on a brand’s behalf.

The scale of the shift is striking: Adweek cites research showing automated traffic grew 23.5% last year, around eight times the rate of human traffic growth. As agents take on more of the buying journey, the protocols that let them read offers, verify merchants and complete payment are quietly becoming critical infrastructure that most marketers have never heard of. This is no longer theoretical: Retail Week reported the same week that Asos has connected its catalogue to ChatGPT, a concrete example of a named high-street brand putting its products in front of an AI assistant.

Why it matters
You do not need to master every acronym, but you do need to know that an agent-readable layer of commerce is being built right now, and that being absent from it will cost visibility. The practical move is to make sure your product data, pricing and merchant credentials are structured and verifiable, so that when an agent evaluates options on a customer’s behalf, your brand is one it can read, trust and transact with. This is the plumbing behind agentic commerce, and the brands that get their data house in order early will be the ones agents can actually buy from.

Alipay launches an AI wallet and Token Pay for the agentic economy

Sources: afp.com, cfotech.asia | 27 May 2026

Alipay introduced a full-stack AI payment system aimed at powering the agentic economy, including a new AI Wallet and a feature called Token Pay. The tools are designed to let AI agents make payments on a user’s behalf while keeping the human firmly in control, letting people approve and track agent-led spending rather than handing over a blank cheque. Alipay is positioning the infrastructure for partners across industries, from AI companies to traditional retailers.

The launch is part of a wider race among payment providers to solve the trust and authorisation problem at the heart of agentic commerce: how to let software spend your money safely. Token Pay and the AI Wallet are Alipay’s answer, giving agents a controlled, auditable way to transact within limits a user has set.

Why it matters
Payment is the step that makes agentic commerce real, and the providers solving it first will shape how it works. For merchants, the message is that the rails for agent-led spending are being laid now, with built-in controls and tracking that customers will come to expect. As these wallets spread, make sure your checkout and payment setup can accept agent-initiated, tokenised transactions, and watch which providers your customers adopt. The businesses ready to accept payment from a trusted agent will capture sales that others cannot complete.

Visa and Mastercard race to own the trust rails of agentic commerce

Sources: mastercard.com, retailtechinnovationhub.com | 26-28 May 2026

Mastercard set out its vision for trusted agentic commerce, describing a future of AI agents transacting through secure payments and trusted technology, with the card network positioning itself as the layer that verifies identity and keeps agent-led transactions safe. In a clear signal of how seriously the payment giants are taking this, Visa hired Ines Clark, founder of the AI retail personalisation startup MartyAI, into a new Agentic Commerce Strategy and go-to-market role. Clark said discovery and checkout were the two areas where the right choices could “unlock a completely different world for consumer journeys”, and that all roads led back to Visa’s agentic commerce work.

Both moves underline that the card networks intend to be the trusted middle layer of agentic commerce, the part that proves an agent is acting legitimately for a real customer. With Alipay, Google and others building competing rails, the question of whose system agents use to verify and pay is becoming one of the most valuable contests in commerce.

Why it matters
The payment networks are betting that trust and verification, not just moving money, will be the prize in agentic commerce. For retailers, the practical implication is that the major networks will offer agent-verification and secure-checkout services you will want to plug into, much as you rely on them for fraud protection today. Keep an eye on what Visa and Mastercard launch in this space, because adopting their agentic-commerce tools early will likely be the simplest way to let trusted agents buy from you without taking on the fraud risk yourself.

AI for Other Sectors and Industries

HEALTHCARE: CHAI publishes AI governance guidance as the market heads past $25bn

Sources: healthcaredive.com, finance.yahoo.com | 26-29 May 2026

The Coalition for Health AI (CHAI) released new governance guidance to help health systems adopt AI responsibly, offering a flexible framework intended to work regardless of an organisation’s size or resources, according to Healthcare Dive. The guidance arrives as the commercial momentum builds: a separate research report projects the generative-AI-in-healthcare market to grow by more than $25bn between 2025 and 2033, driven by the rising complexity of medical data, better diagnostic accuracy and workflow efficiency.

The pairing is telling. As investment and product launches accelerate, the sector is moving quickly to put governance frameworks in place, recognising that healthcare AI cannot scale safely without clear rules on data quality, accountability and clinical oversight.

Why it matters
For anyone marketing to or working within healthcare, governance is becoming the gate that AI products must pass through. Buyers will increasingly ask how a tool fits frameworks like CHAI’s before they ask about its features. The lesson generalises beyond healthcare: in any regulated sector, demonstrating responsible AI governance, clean data, clear accountability and human oversight, is becoming a precondition for adoption, not an afterthought. Lead with how your AI is governed, not just what it does.

Sources: gartner.com, artificiallawyer.com | 26 May 2026

Gartner predicted that legal-technology budgets will double by 2028 as legal AI use expands, a sign of how fast the profession is moving from cautious pilots to real investment. The forecast follows a busy spring for legal AI, including Anthropic’s expansion of Claude for Legal with role-specific tools for tasks such as due diligence and document drafting, adopted on live matters by firms including Freshfields and Quinn Emanuel, and reports that OpenAI is planning its own legal offering.

The combination of a credible analyst forecast and concrete product launches suggests legal AI has reached an inflection point. Specialised platforms such as Harvey, Legora and Thomson Reuters CoCounsel are delivering measurable productivity gains, and budgets are following.

Why it matters
Even traditionally conservative, risk-averse sectors are now committing real money to AI, which tells every other industry something about the pace of change. For professional-services firms of any kind, the message is that clients will increasingly expect AI-assisted efficiency and pricing. For marketers serving the legal sector, AI capability is fast becoming a buying criterion. If a profession built on precedent and caution is doubling its AI budgets, the question for your own sector is not whether to invest but how quickly.

FINANCE: Most financial advisers now use AI, and banks cut fraud losses with it

Sources: dwealth.news, commbank.com.au | 26-29 May 2026

Adoption of AI in financial services has crossed into the mainstream. An InspereX Spring 2026 survey found that 70% of financial advisers now use at least one AI tool, rising to 84% among advisers managing more than $351m in assets, a clear sign that AI has moved from curiosity to standard kit among professionals who are paid to be careful with money. On the defensive side, Commonwealth Bank reported that its agentic AI fraud-detection system, which spots emerging scam patterns and generates the rules to intercept them, helped cut fraud losses by more than 20% year on year.

The two data points capture both sides of AI in finance: it is now a productivity tool that most advisers rely on, and a security tool that is measurably reducing losses, even as fraudsters use the same technology to scale their attacks.

Why it matters
When most professionals in a cautious, heavily regulated field are already using AI daily, “wait and see” is no longer a credible strategy in any sector. The Commonwealth Bank example also shows the pattern to aim for: AI applied to a specific, measurable problem with a clear return, in this case a 20% reduction in fraud losses. For your own business, the lesson is to pick concrete problems where AI can show a number, rather than adopting it in the abstract. Measurable wins build the confidence and the budget for everything that follows.

Key Takeaways

  • Anthropic launched Claude Opus 4.8 with a new Dynamic Workflows tool for Claude Code that coordinates work across tens to hundreds of background agents, plus user-controlled effort levels and a faster, cheaper fast mode.
  • Anthropic raised a record $65bn at a $965bn valuation, overtaking OpenAI to become the most valuable AI company, with run-rate revenue past $47bn and an IPO expected.
  • OpenAI is preparing conversion-focused, pay-per-action ads inside ChatGPT aimed at small and local businesses, a genuine new performance channel to rival Google and Meta.
  • Apple is reportedly shrinking Google’s Gemini to power a smarter, more conversational Siri on the iPhone, ahead of WWDC in June.
  • AI leaders including Jensen Huang and Sam Altman softened their job-loss warnings, with Huang calling the AI-equals-layoffs narrative “too lazy”, even as Groupon, Cloudflare, Wix and Coinbase announced AI-linked cuts.
  • Cloudflare’s 20% cuts hit “measurers” in middle management and oversight functions while it keeps hiring “builders and sellers”, a signal of which roles leaders see AI replacing first.
  • Agentic commerce hardened into infrastructure: Amazon is selling its AI shopping tools to rivals, Alipay launched an AI wallet, and Visa and Mastercard moved to own the trust and verification layer.
  • Even cautious sectors are committing: Gartner says legal-tech budgets will double by 2028, and 70% of financial advisers now use at least one AI tool.

Frequently Asked Questions

What is Claude Opus 4.8 and should our team use it?

Claude Opus 4.8 is Anthropic’s latest top-tier model, released on 28 May 2026, with improvements across coding, agentic tasks and knowledge work at the same price as the previous version. The two features most relevant to marketing and management teams are user-controlled effort levels on claude.ai, so you only pay for deep reasoning when a task needs it, and a Dynamic Workflows tool in Claude Code that can run a whole project across many background agents rather than one step at a time. At Anicca we build client audits, reporting dashboards and content pipelines on exactly this kind of multi-agent workflow. If you would like help working out where AI agents could replace a manual, multi-step process in your business, get in touch with the Anicca team.

How should we prepare for AI shopping agents and agentic commerce?

The starting point is to make your product data, pricing and merchant information structured, accurate and machine-readable, so that AI agents can read, trust and transact with your brand on a customer’s behalf. New standards such as the Model Context Protocol (MCP) and the Universal Commerce Protocol (UCP) are becoming the plumbing of agent-led shopping, and payment providers including Alipay, Visa and Mastercard are building the trust and verification rails on top. If you are on Shopify or Magento, both platforms already offer plugins that handle much of the feed structure and agent compatibility, so the job is configuring and validating them. For brands on bespoke or older platforms, Anicca runs a dedicated agentic-commerce readiness audit covering feed coverage, structured data, identity signals and your margin model when an agent sits between you and the customer. Contact us to run this for your brand.

Is AI really causing the job cuts in the news?

Often, not in the simple way the headlines suggest. AI leaders including Nvidia’s Jensen Huang have publicly argued that many companies blaming AI for redundancies are using it as convenient cover for cost-cutting that was happening anyway, with Huang calling the AI-equals-layoffs narrative “too lazy”. The more accurate picture, supported by HR analysis this week, is that genuine AI transformation in the first 12 to 18 months looks like role redesign rather than mass replacement, with displacement and new hiring happening at once. The practical management advice is to plan for your teams to do different work alongside AI, protect the roles that build, sell and exercise judgement, and never use AI as an excuse for cuts you cannot evidence.

Conclusion

This was the week the AI industry’s own narrative matured. Anthropic became the most valuable AI company in the world and launched tools that let a single instruction run a whole project across many agents, while its rivals raced to put AI advertising, shopping and payments in front of every business and consumer. At the same time, the loudest voices in AI quietly stepped back from predicting mass unemployment, reframing the near-term story as redesign rather than replacement. Three actions for senior leaders this quarter: first, find one manual, multi-step process in your business and test whether an AI agent workflow could run it. Second, get your product and brand data structured and machine-readable, because agents are increasingly the first thing that reads them. Third, if AI is changing roles in your team, plan and communicate that change with evidence and care, because how you handle it is now a reputational risk in its own right.

Need help adapting your AI marketing strategy? Contact the Anicca team for expert guidance.

FREE GUIDE

Free download

New Edition (3) of The Complete
Claude Code Implementation Guide

Ann Stanley
Ann StanleyFounder & CTO, Anicca Digital

Updated May 2026 – 88 pages for marketers, managers and business leaders, now with a brand new chapter on Company AI & AI operating systems. No coding required.

Why it matters Understanding Claude Code Getting set up Working with Claude Skills Business use cases Rollout & governance Company AI & AI operating systems
Download → anicca.co.uk/claude-code-guide
The Complete Claude Code Implementation Guide, Edition 3 cover
The Thursday AI Club logo

Thursday AI Club

A hands-on AI club for marketers and managers. Fortnightly 3-hour sessions: Hour 1 is an open Q&A on any AI question; Hours 2-3 split into a workshop track (for newbies) and an advanced track (live demos and deeper questions).

Led by Ann Stanley, Darren Wynn, James Allen.

Workshop recordings library
Skills and prompts to copy
Private community channel
Secret Agents community access
Full-day hackathon every quarter
Cancel any time

Membership: £40/month or £400/year

Join the Thursday AI Club →

This roundup is compiled from publicly available sources using AI-assisted research. While we review every article for accuracy, our analysis reflects our interpretation of the original reporting. We strongly encourage readers to click through to the original sources linked throughout this post for full context and detail. If you spot anything that needs correcting, please let us know.

Similar Posts