The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.
A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.
Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail
AI for creating, distributing, and measuring content across channels. The most mature creative domain: SEO, copywriting, email, and social media management are established practice. Personalisation at scale and sentiment-driven strategy are advancing but unevenly adopted. Content authenticity and deepfake detection remain bleeding-edge.
The headline: Almost every marketing team now uses AI, and almost none can prove it made money. The winners this fortnight were the ones who spent on review processes, not on tools.
Marketing has adopted AI faster than any other business function — 88 to 97 percent of marketers use it daily, and roughly three-quarters of new web pages now contain machine-generated text. Almost nobody is getting paid for it: MIT research puts the share of company AI pilots showing no measurable financial return at 95 percent. Everyone uses AI, so using it has stopped being a differentiator. The small group pulling ahead has invested in the unglamorous layer — documented brand rules, review checkpoints, honest measurement — and is posting real numbers. The rest are producing more content that search engines rank lower and customers trust less.
Europe's AI labeling rules went live on August 2. AI-generated images, video, and audio must now carry machine-readable markings and be disclosed to users, with fines up to 15 million euros or 3 percent of global turnover. Written text gets an exemption where a named human has editorially reviewed it — video, voice, and synthetic imagery do not. If you advertise or publish into Europe, the practical answer is to name an accountable reviewer for text and add provenance marking to everything else.
Starbucks Korea's CEO was fired over a single AI-generated campaign. The creative unintentionally evoked the 1980 Gwangju massacre; criminal charges followed, card volume dropped 26 percent in a week, and all 2,160 Korean stores went through mandatory retraining. The copy was grammatically perfect — which is exactly the danger, because fluency now hides cultural errors that used to be obvious. Any campaign running in a market whose history your team doesn't know needs a native reviewer before launch, not after.
A 500-million-impression study found AI ads win clicks but lose sales. Ad tech firm Taboola measured AI-generated creative at 0.76 percent click-through against 0.65 percent for human-made — but 8 to 14 percent fewer conversions on higher-priced products. Cheap AI creative is a reasonable bet on low-value, high-volume items and a poor one above roughly $200 a basket. Check which side of that line your products sit on before scaling variant production.
LinkedIn now demotes AI-sounding posts, and disclosure labels cut engagement further. Peer-reviewed analysis of 56,005 posts found a 4 to 7 percent reach penalty for templated AI phrasing, and separate research found that labeling content as AI-made suppresses engagement, worst on emotional posts. Transparency and performance now conflict, so budget for the editing time that makes output not read as machine-written.
AI-written summaries now appear on 30 percent of searches, double January's rate. Meanwhile 83 percent of the content those summaries cite does not rank in the top ten search results, and about a third of what your analytics calls "direct traffic" is actually people arriving from AI tools. Your search reporting is understating a growing share of demand; ask your team what percentage of pipeline they genuinely cannot attribute.
Enforcement of the European rules is the next shoe to drop. No fines have been issued yet, and global spending on AI governance tooling sits at only $492 million — small relative to the exposure. Expect the first enforcement actions within 12 months; get an inventory of where your business publishes synthetic media, and who signs off on it, done before then.
Vendors are splitting into "control" and "speed" camps. Jasper has repositioned as an enterprise governance platform ($180 million projected revenue, customers including Boeing, UPS, and Accenture) while rivals such as Writer lead on enforcement of brand rules. Your renewal decision over the next year is really a decision about which risk you care about more — slow output or off-brand output.
Synthetic market research is being marked down. Two rigorous studies this month found AI-generated "consumer panels" overweight demographic stereotypes by a factor of 40 to 67 and point targeting at the wrong segment in half to three-quarters of test cases. Only 8 percent of researchers use them regularly. Treat synthetic panels as a way to rank options, never as a source of the numbers you put in a business case.
Governance lags production by a wide margin. Eighty-one percent of organizations ship off-brand content despite having written brand guidelines, 46 percent of marketers skip fact-checking entirely, and 65 percent have no AI policy at all. Meanwhile 34 percent of enterprise teams now run agents (software that acts on its own without being prompted) in production, and 29 percent abandon them within 90 days — usually because nobody defined what success looked like.
AI is good at average and bad at excellent. A peer-reviewed Ipsos and Syracuse University study of 3,000 consumers found AI-made ads score five points below the norm while human-made ads score eleven above — even though 87 percent of viewers could not tell which was which. Separate research comparing 100,000 people against the leading models found AI matches the average person and loses badly to the top ten percent. For anything carrying your brand's reputation, the AI is a drafting tool, not the author.
You can no longer measure what you used to measure. Sixty-eight percent of US searches now end without a click, and 76 percent of demand-generation leaders say they don't trust their attribution models — while 89 percent use them for budget decisions anyway. Rebuilding measurement around mix modeling and controlled tests is slow and expensive, and the alternative is allocating spend on numbers you privately know are wrong.
Go deeper: the full Content & Marketing briefing — the longer analytical write-up, plus every practice we track in this domain with its maturity rating, the tools to consider, and the evidence behind our assessment.