Bundling moved the AI decision from opt in to opt out
Google moved Gemini from a paid add on into the core Workspace Business and Enterprise editions and raised list prices to cover it. That reversed the buyer's position: before, you opted in seat by seat; now you opt out only by changing editions or negotiating the uplift away. Most first quotes carry it unchallenged, because an edition repricing reads as non negotiable. It is negotiable.
Prepared by Redress Compliance · August 10, 2026 · Google advisory. Based on 20 to 25 Workspace and Gemini procurement reviews, 2024 to 2025.
Executive summary
Active Gemini use sat at 20 to 35 percent of licensed seats a quarter after rollout. The uplift prices every seat for AI while telemetry says most seats never invoke it, and that gap is the entire negotiation file.
In roughly 12 of the 20 plus estates we reviewed, the bundled uplift priced AI across a population where two thirds of seats never touched the features. Most Gemini overspend is invisible in the invoice precisely because it sits in that gap rather than in a line item.
Renewal uplifts opened at 15 to 25 percent and settled near single digits once usage data was on the table.
The evidence is a 90 day Gemini usage report from the Admin console showing active AI users against licensed seats, pulled before the renewal conversation starts rather than requested during it.
Sellers know what the report says, which is why tabling it changes the tone of the meeting rather than merely the arithmetic.
Edition re mapping cut total Workspace spend 10 to 18 percent, more than a flat discount does. A tiered mix beats a uniform edition in almost every estate above 500 seats, and the saving comes from moving light users down rather than from discounting the top tier.
Heavy AI and meeting users belong on Enterprise Plus, core knowledge workers on Business or Enterprise Standard, light and mobile first users on Frontline or Starter, and shared or service accounts on archive or nothing at all.
Three levers move the uplift, and the order matters. Usage first, opening with the telemetry gap and pricing the uplift against active users rather than seats. Then a competitor anchor, because a current written Copilot quote resets the AI premium conversation where a verbal mention does nothing.
Then term structure, trading an annual or two year commitment for a written cap on the next renewal uplift. Usage data earns credibility, the anchor sets the range, and the cap locks the win in writing.
The edition mix, by user profile
| User profile | Right edition | Common overbuy |
|---|---|---|
| Heavy AI and meeting users | Enterprise Plus | Correct fit |
| Core knowledge workers | Business or Enterprise Standard | Enterprise Plus |
| Light and mobile first users | Frontline or Starter | Business Standard |
| Shared and service accounts | Archive or none | Full paid seat |
The mix conversation usually stalls on administration rather than on economics, and the finance answer is short. IT teams resist mixed editions because admin overhead rises slightly with more tiers to manage, provision, and audit.
A 10 to 18 percent spend cut pays for a considerable amount of administrative time, so model both sides rather than conceding uniformity as a default.
The dormant account cut is cheaper still: licensed accounts with no Workspace activity at all are the least contested reduction available, and they require no argument about edition fit or feature entitlement.
What each edition includes after the bundling is set out in the Gemini in Workspace licensing guide.
The usage audit that changes the meeting
- Pull at least 90 days of Gemini usage from the Admin console before any renewal conversation opens, since the report is worth far less produced in reaction to a quote.
- Measure active Gemini users, meaning unique users invoking AI features monthly rather than cumulative trial counts, which flatter the number substantially.
- Test edition fit: identify users on Enterprise Plus whose measured behaviour fits Business Standard, which is the largest single re mapping population in most estates.
- Quantify dormant seats: licensed accounts with no Workspace activity at all, the cheapest cut available and the one nobody defends.
- Collect a written competitor quote rather than mentioning one, because Google and Microsoft price the AI premium against each other and a documented quote moves the range.
The Gemini Workspace procurement kit
The seat usage worksheet, the edition mix model, and the uplift cap language that survives Google's redlines.
Get the white paper →Running the three levers in order
The sequence is what makes the levers work, because each one earns the right to the next. Usage comes first: open with the telemetry gap and price the uplift against active users rather than licensed seats, which converts an abstract argument about value into an arithmetic one about population.
That is also the point at which credibility is established, because the seller can verify every figure and knows the report exists whether or not you produce it.
The competitor anchor comes second, once the population question is settled, and it has to be a current documented quote rather than a reference to market conditions: Google and Microsoft price the AI premium against each other.
So a written Copilot number resets the range in a way that no assertion does.
Term structure comes last, because a commitment traded before the price is settled buys nothing.
Annual and multi year plans carry lower effective rates than flexible monthly billing, and they are the natural trade for a written cap on the next renewal uplift.
But commit only after the edition mix is fixed, since committing on a population you are about to re map locks in the overbuy you were trying to remove.
Run them out of order and each lever loses most of its force. The competitive picture across the assistants sits in the three assistant comparison, and the Microsoft seat economics in the Copilot pricing guide.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What we saw across Workspace engagements, 2024 to 2025
The standard reseller advice is to accept the bundled uplift because AI is now table stakes and the per seat increase looks small. We disagree, because the increase prices every seat for a capability most seats never invoke:
Licensed seats showing genuine AI use in admin telemetry a quarter after rollout, against an uplift applied to the entire base.
Total Workspace spend removed by matching users to the cheapest edition their measured behaviour supports, before any discount.
Three patterns recurred: active Gemini usage in admin telemetry sitting at 20 to 35 percent of licensed seats a quarter after rollout, estates that re mapped users across editions before renewal cutting total Workspace spend 10 to 18 percent.
And renewal uplifts opening at 15 to 25 percent then settling near single digits once usage data sat on the table.
The buyer side move is to price the uplift against measured active users, re map editions before renewal rather than after, cut dormant and shared accounts from the licensed base, and take the cap in writing.
Paying list for bundled ambition is a choice rather than a default, and the telemetry is what makes that visible.
Your first five moves
- Pull 90 days of Gemini and Workspace usage from the Admin console before any renewal conversation, because the report produced in reaction to a quote carries far less weight.
- Map every user to the cheapest edition their measured behaviour supports, which cut total spend 10 to 18 percent in our file, more than a flat discount achieves.
- Quantify dormant and shared accounts and cut them from the licensed base, the least contested reduction available and the one requiring no argument about entitlement.
- Collect a current written competitor quote to anchor the AI premium, since a documented Copilot number moves the range and a verbal mention does not.
- Open with usage data, then trade term length only for a written uplift cap, and commit only after the edition mix is fixed. The Google practice runs the audit and the renewal with you.
Frequently asked questions
Is Gemini now included in Google Workspace?
Yes. Google folded Gemini AI features into the Workspace Business and Enterprise editions and raised list prices to cover them, retiring the separate add on SKUs for most editions.
The cost now arrives through the edition price, which is why the renewal conversation is an edition repricing rather than an AI SKU decision.
Can you opt out of paying for Gemini in Workspace?
Not directly on the main editions. You manage the cost by edition mix instead: moving light users to cheaper tiers, cutting dormant and shared accounts from the licensed base, and negotiating the uplift against measured usage.
Bundling moved the decision from opt in to opt out, so the levers are commercial rather than a configuration setting.
How much did Workspace prices rise with the bundling?
List increases varied by edition and typically ran in the low double digits in percentage terms, but the number that matters is your own renewal uplift. In the estates we reviewed those opened at 15 to 25 percent before negotiation and settled near single digits once usage telemetry was tabled.
What usage share justifies pushing back on the uplift?
Anything under 50 percent active use is a strong pushback position. Our reviews measured 20 to 35 percent active Gemini use a quarter after rollout, which moved renewal quotes materially.
The measurement has to be unique users invoking AI features monthly rather than cumulative trials, which overstate adoption considerably.
Does an edition mix really beat a flat discount?
In almost every estate above 500 seats, yes. Re mapping users across editions cut total Workspace spend 10 to 18 percent, and the saving comes from moving light users down rather than from discounting the top tier.
IT teams resist the admin overhead, but a cut of that size pays for a great deal of administrative time.
Does a Microsoft 365 quote move a Google renewal?
Yes, and it is the single most effective anchor, because Google and Microsoft price the AI premium against each other. The requirement is a current documented quote rather than a mention: a verbal reference does very little, while a written number resets the range the seller is working within.
Should we sign an annual or multi year Workspace plan?
Annual and multi year plans carry lower effective rates than flexible monthly billing, and they are the natural trade for a written cap on the next renewal uplift.
Commit only after the edition mix is fixed, though, because committing on a population you are about to re map locks in the overbuy you were trying to remove.
