Gemini for Google Workspace in a UK business: what is included, what it costs, and when to add Claude or ChatGPT anyway.

Google folded Gemini into Workspace’s Business plans in 2025 and raised the plan prices at the same time, so a firm on Business Standard already pays for it. In the UK, Business Standard is listed at £11.80 a user a month on an annual commitment or £14.00 flexible, before VAT, and includes Gemini across Gmail, Docs, Sheets, Slides and Meet, the standalone Gemini app, and NotebookLM for questions over uploaded sources. Business Plus and Enterprise add more. Check the pricing page for your plan, because the inclusions differ by tier and the figures move.

Inclusions by plan from Google Workspace’s pricing page, read September 2026.

Google states that Gemini for Workspace does not use customer prompts, content or generated responses to train its models, that interactions stay within the Workspace boundary, and that they are covered by the Workspace terms, data processing addendum and existing certifications. That is the same category of answer as Claude Team and ChatGPT Business, and it is materially better than the consumer Gemini app on a personal account, which is where staff drift if the firm does not switch the Workspace feature on. The admin console controls which users have Gemini and which connectors it may use.

For most Google firms, most of the time. Drafting and rewriting in the documents the firm already uses; summarising threads and meetings; questions over a Drive folder through NotebookLM; the Sheets analysis that used to wait for someone who knew formulas. The advantage is that it is where the work already is, with the firm’s permissions, at no additional cost, and with no second tool to govern. Switch it on, adopt the one-page policy, run a session per team on their own documents, and measure a first process before deciding anything else.

Three cases. When the work is long-document reasoning, contracts, technical reports, complex proposals, and a side-by-side test on the firm’s own documents shows a clear quality gap. When the firm needs a coding agent for its own software, which is Claude Code’s ground. And when the team wants a shared project library with standing instructions and templates that it lives in, which is what Claude Team’s projects and ChatGPT’s custom GPTs provide. The test is the same as always: the ten tasks the team does most, run on each tool, scored blind. Pay for the second seat only where the gap is real and the task is frequent.

Leave Gemini switched off in the console while staff use the consumer app on personal accounts, which is the shadow-AI position with the tool already paid for. Buy a second assistant for everyone before testing whether Gemini covers the tasks. Turn on Meet note-taking without the consent and notice rules in place. And build a knowledge assistant on a third-party platform when NotebookLM over a well-kept Drive folder would answer the first hundred questions.