The reseller email lands on a Tuesday. Your firm already pays for Google Workspace Business Standard, the AI drafting buttons have been appearing in Gmail and Docs for months without anyone asking for them, and now someone wants to sell you an add-on called AI Expanded Access. The pitch implies it is the obvious next step for any business taking AI seriously. Whether that holds for your business depends on what your team actually does all day, and for many owner-managed firms the honest answer is that the AI you already pay for covers it.
What is AI Expanded Access in Google Workspace?
AI Expanded Access is a per-user add-on licence for Google Workspace Business Standard, Business Plus and the two Enterprise tiers. It raises the monthly usage caps on the heavier generative features, advanced video, avatar, image and music generation, high-volume AI work in Sheets, and real-time speech translation. The everyday AI, drafting in Gmail and Docs, summarising documents, meeting notes in Meet, is already included in your core plan.
The context matters here. In April 2025 Google folded its Gemini features into the core Business editions. Before that change, a firm on Business Standard that wanted serious AI capability paid for the plan plus a separate Gemini Business add-on, around 32 US dollars per user per month in total. Afterwards, the same firm pays 14 dollars, which Google was keen to point out is only 2 dollars more than Workspace cost without any AI at all.
So the old question, should we buy AI for the business, has largely been answered on your behalf. The live question is narrower and more useful. Do specific people in your firm need more of it than the standard caps allow? Expanded Access exists purely to answer yes for those people. Users who hold the licence see an Expanded badge in the Gemini app and get materially higher quotas, while everyone else carries on exactly as before.
What price should you expect to pay?
Google does not publish a standard list price for AI Expanded Access. Its documentation says pricing depends on your local currency and the length of your subscription, and directs you to the admin console or your reseller for an exact figure. What you can pin down is the base layer underneath it, because the add-on is charged on top of your existing per-user plan cost.
The base layer looks like this. Business Standard costs 14 US dollars per user per month on an annual commitment, or 16.80 on the month-to-month Flexible plan. Business Plus sits at 22 and 26.40 respectively. UK buying guidance benchmarks Gemini-inclusive Workspace at roughly £11.50 per user per month, against £24.70 for Microsoft 365 Copilot and about £19.50 for ChatGPT Business (formerly ChatGPT Team), which makes Google the mid-price option if your firm already lives in its tools.
Two quirks affect the quote itself. If your organisation has never paid Google for anything, you may be asked for a one-off early payment of 30 dollars before you can buy Expanded Access licences at all. And the billing plans behave differently. The Flexible plan pro-rates to the day, so a licence added on the 1st and removed on the 15th costs half a month, while the Annual plan locks your licence count in until renewal. That difference matters more than it looks, because the sensible way to buy this add-on is as a time-boxed pilot rather than a blanket commitment.
Where will you actually meet the usage limits?
Almost nowhere, if your firm’s AI use is text-first. The standard Business plans include up to 300 Gemini Thinking prompts per user per day, and the monthly caps on media features are generous for ordinary use. You meet the limits when the work itself is generative media or heavy data processing, produced at volume, month after month.
The numbers make the boundary concrete. On Business Standard or Plus, video generation in Vids is capped at 50 videos per user per month, rising to 200 with Expanded Access. Avatar generation goes from 25 to 100. The Nano Banana Pro image model in Slides and Vids goes from 30 uses per month to 300. The AI Function in Sheets moves from 5,000 calls per month to 25,000, and audio overviews of PDFs double from 20 per day to 40.
Two details shape how those figures play out in practice. Quotas are per user, they reset daily or monthly at midnight Pacific Time, and unused capacity cannot be pooled or carried forward, so a heavy user cannot borrow a colleague’s allowance. And some features currently feel limitless because Google has not started enforcing their caps yet. Match writing style in Docs, for instance, is on promotional access with limits due to bite from July 2026, so today’s experience is an unreliable guide to next year’s.
When should you buy it, and when should you leave it alone?
Buy it for named roles with a measurable media or data workload. A marketing lead producing client videos in Vids, a training business generating avatar-led content, an analyst pushing thousands of AI calls through Sheets each month. Leave it alone if your team mainly drafts, summarises and searches, because the embedded quotas already cover that kind of work comfortably.
The buying pattern that fits an owner-managed business is a small pilot. Assign Expanded Access to two or three people on the Flexible plan for a quarter, and watch how close they come to the caps. The admin console reports AI usage by feature, user and app, so this is measurable rather than a matter of impression. If the quotas are genuinely being hit and the output shows up in client work, convert those seats to Annual pricing. If they are barely touched, cancel, and you have spent very little finding out.
There are also firms for whom the answer is no regardless of workload. If your governance stance keeps client data out of AI prompts, as it should in parts of accountancy, law and regulated advice, higher quotas add little. If your firm runs on Microsoft 365 or a multi-vendor stack, deeper Google-only AI duplicates capability you already pay for elsewhere. And Google’s own conditions say the licence should not be assigned to anyone under 18, which occasionally matters for training and education businesses with junior staff.
What else should you check before you decide?
Governance, mainly. Higher quotas widen what staff can produce, so the purchase should travel with an acceptable use policy, a simple data classification rule and admin settings that keep AI activity inside the protected Workspace environment. UK regulation is moving towards meaningful human involvement in consequential automated decisions, and the ICO expects policies that reflect what your team actually does.
Start with what is already happening. Microsoft research cited in UK guidance found 71 per cent of employees using AI tools their organisation did not know about, and that pattern will include your firm until you ask. A short conversation about which tools people already reach for tells you whether Expanded Access would regularise existing behaviour or simply add cost on top of it.
Then set the technical side to match. Workspace admins can turn on Gemini with added data protection, under which Google states that chats are not reviewed, not used to train its models and not shared beyond your organisation. Pair that with a policy under three pages that names the approved tools, defines public, internal and restricted data, and states plainly that client personal or financial information stays out of prompts. The ICO’s AI guidance and the NCSC’s advice on using AI securely were written for exactly this exercise, and neither requires a consultant to apply.
The decision itself is smaller than the pitch makes it sound. The AI is already in your subscription. Expanded Access is a volume purchase for specific roles, bought best as a measured pilot with a governance wrapper around it. If you want a second pair of eyes on where AI spend genuinely earns its keep in your business, book a conversation.
