September–October 2026 was one of the busiest stretches the AI world has seen. Within a few weeks, OpenAI, Google and Anthropic introduced new generations of models and work tools. This post skips the benchmark numbers and focuses on the real question: what does this change for an ordinary business?
What was launched?
- OpenAI: GPT-6. In early September OpenAI began a staged rollout of the GPT-6 family, first to selected business customers, then to paid ChatGPT plans and the API. The emphasis is stronger reasoning and the ability to carry out multi-step tasks on its own.
- Google: the Gemini agent. On 8 October, at its "Gemini at Work" event, Google introduced a single Gemini agent that works inside Gmail, Drive, Docs, Sheets and Calendar and also integrates with Slack and Microsoft 365. It is currently in a limited preview for enterprise customers.
- Anthropic: the Claude 5.5 family. The flagship Claude Opus 5.5 arrived in September, followed by the faster and cheaper Sonnet 5.5 and Haiku 5.5. The key difference is that strong model capability is now available at a much lower price than before.
4 key takeaways for business
1. The "chatbot" era is ending and the "agent" era is beginning
All three companies send the same message: AI no longer just answers your question, it does the task. It writes and sends the email, fills in the spreadsheet, prepares the report. That means a large share of everyday repetitive office work can be automated.
2. Prices are falling, so AI is within reach of small businesses too
With lightweight models getting cheaper, an AI operator answering customers on WhatsApp, or a system that writes product descriptions automatically, is no longer a luxury reserved for large companies.
3. AI now lives inside work tools
Gemini moving into Gmail and Docs, and its rivals into their own office suites, shows that AI is no longer a separate "program". The key question becomes not "which AI should I pick?" but "are my processes ready to work with AI?"
4. Data and rules matter more
The more capable the agent, the more sensitive the permissions and data you give it. Decide in advance what information it can access, which decisions AI makes on its own, and where a human needs to approve.
So what should you do now?
- List the three repetitive tasks that take up the most of your team's time each week.
- Try one of them with AI, as a small, measurable project.
- Measure the result: hours saved, response speed, number of errors.
- Scale what works and drop what doesn't.
Models will be updated every few months. The winner won't be whoever adopts the newest model first, but the company that fits AI into its processes best.
Want to work out together where AI will bring real value to your business? See our AI chatbot and business automation services.