Q3 in one sentence: The experimentation phase ended in Q2. Q3 is where UK businesses that built consistent AI habits pull ahead of those still exploring. Three signals from the last 24 hours each give you a specific, practical move to make before Friday.
Signal 1 — Claude Sonnet 5 is now the default for every Claude plan
As of today, Anthropic has made Claude Sonnet 5 the default model for Free, Pro, Max, Team, and Enterprise plans. If you pay for Claude, you are now using Sonnet 5 without doing anything. If you use Claude free, you are also now on Sonnet 5.
Why this matters for UK operators: Sonnet 5 performs close to Opus 4.8 — the most capable model in Anthropic's range — but costs dramatically less through the API. The promotional price of £2 per million input tokens (approximately) through 31 August makes it the best price-to-performance option currently available from any major provider. For a typical small business running 50-100 AI tasks per day, the economics are significantly more favourable than six months ago.
The practical read: if you have been holding off on AI tools because the capable models felt expensive, that barrier has just dropped substantially. The floor for near-frontier performance is now the free tier.
Log into claude.ai and run your hardest recurring task today. Document the output. This is your baseline for August. If you are on the API, update your model parameter to claude-sonnet-5 before the price rises on 1 September.
Signal 2 — Gemini 3.5 Pro enters July without launching
Google's most-anticipated model missed its June 30 target and entered July 2026 still in limited enterprise preview on Vertex AI. The 2-million-token context window and Deep Think reasoning mode remain inaccessible to most UK businesses.
This is not necessarily bad news for operators. Gemini 1.5 Flash and Gemini 2.0 Flash are both available today, both fast, and both capable of handling most routine business tasks — email drafting, document summarisation, customer query handling — adequately. The question of whether to wait for Gemini 3.5 Pro only makes sense if your specific use case genuinely needs the 2M-token context (processing a 1,500-page contract, for instance) or Deep Think reasoning for complex decision support.
For most UK service businesses — chimney sweeps, tradespeople, retailers, local services — Gemini Flash is sufficient right now and costs less than waiting for Pro.
Set a 15 July calendar reminder to evaluate Gemini 3.5 Pro when it ships. In the meantime, if you use Google Workspace, explore the Gemini sidebar in Gmail and Docs — it is live now and costs nothing extra on eligible plans. Do not delay platform decisions waiting for Pro.
Signal 3 — The UK government just launched formal AI tools for employers
The UK Department for Science, Innovation and Technology (DSIT) has published the SKAI (Skills and Knowledge for AI) programme alongside the PRIMES framework — structured guidance for employers on how to upskill their workforce for AI and measure the resulting productivity gains. This follows BCC data showing 54% of UK businesses now use AI but fewer than 25% have it embedded in core processes.
The gap between using AI and gaining from AI is now officially a government concern. The PRIMES framework in particular gives small employers a structured self-assessment tool — something that was previously only accessible through expensive consultancy or enterprise software.
Search "DSIT PRIMES AI framework" and run the self-assessment for your business. It takes under an hour and tells you exactly where you are in AI adoption. This is free, government-backed, and gives you a benchmark you can track quarterly.
The Q3 operator read
Three moves before the end of this week
Move 1 — Upgrade your Claude use today. Sonnet 5 is live and free or cheap depending on your plan. Run your hardest task. Document the output. This is the baseline for measuring Q3 gains.
Move 2 — Set a Gemini 3.5 Pro evaluation date. 15 July. Put it in the calendar. Do not delay other AI decisions waiting for it to ship.
Move 3 — Run the PRIMES self-assessment. Under an hour. Free. Gives you a Q3 baseline for AI adoption maturity in your business.
Q3 is not about discovering new AI tools. It is about squeezing more value from the tools you already have access to — and making that use consistent. Sonnet 5 landing as the default on every plan today removes the price excuse. The only variable left is consistency.