Google’s AI "Intern" in Workspace Is Really a New Boss Layer
Google doesn’t want to replace your job – it wants to sit between you and your job. The latest Workspace updates, unveiled at Google Cloud Next, push AI deeper into email, documents and spreadsheets, promising to kill busywork and turn Gemini into something like a tireless office intern. But the real game is bigger: data lock‑in, control over how knowledge work is done, and a quiet reshaping of office power dynamics. In this piece, we’ll unpack what Google actually shipped, who wins and who should be worried – especially in Europe.
The news in brief
According to TechCrunch’s reporting from Google Cloud Next, Google is rolling out a major AI upgrade to Workspace, its subscription productivity suite.
The headline feature is Workspace Intelligence, an AI layer that sits across Gmail, Calendar, Chat and Drive (Docs, Slides, Sheets). It can access a user’s Workspace data to assist with tasks, with admin controls to restrict which services it can see. Google’s message is clear: the more data you let it ingest, the more helpful it becomes.
In Sheets, Gemini can now generate spreadsheet structures via natural‑language prompts, handle formatting and data retrieval, and perform “prompt‑based” filling of data. Google claims this can populate sheets up to nine times faster than manual entry, and can also turn unstructured information into organized tables.
In Docs, new writing tools let Gemini generate, rewrite and refine text based on content from Drive, Gmail and Chat, plus the web. Users can ask it to help write or to mimic their existing writing style.
Why this matters
On the surface, this is about productivity: fewer hours formatting spreadsheets, writing routine emails or wrestling with poor documentation. For individual workers, that’s attractive. If Gemini can build a project tracker in Sheets, summarize a week of email and produce a first draft of a client update, that’s real time saved.
But the strategic stakes are elsewhere.
For Google, this is a retention and expansion play. Workspace is already entrenched in startups, universities and many enterprises. By weaving AI into every corner of the suite, Google makes it harder for companies to switch away: your workflows, templates and even your “organizational memory” start to live inside Gemini’s prompts and suggestions.
For Microsoft, this is confirmation that Office vs. Workspace is now Copilot vs. Gemini. The core apps are mature; the new competitive frontier is the AI layer that orchestrates them. Google’s advantage is that many younger companies “grew up” on Workspace; this release is an attempt to solidify that generation as long‑term customers.
The losers could be mid‑level knowledge workers whose value lies heavily in coordination, formatting and synthesis of information produced by others. When AI handles the drudge work of organizing, drafting and summarizing, the premium shifts further toward domain expertise, interpersonal skills and original thinking.
There’s also a governance problem: giving an automated assistant access to mailboxes, calendars and documents creates an enormous surface area for mistakes, bias and quiet surveillance. IT departments get powerful new knobs to turn, but also new responsibilities and regulatory risk.
The bigger picture
These updates sit squarely in a broader trend: the rise of the "AI orchestration layer" for office work.
In the last two years, Microsoft has pushed Copilot across 365; start‑ups like Notion, Coda and Linear have embedded generative models; and Google itself previously experimented with Duet AI before rebranding around Gemini. The pattern is clear: productivity tools are no longer just containers for human work – they are becoming semi‑autonomous actors in the workflow.
We’ve been here before, in a way. GUI office suites in the 1990s abstracted away command lines and macros. The SaaS wave in the 2010s abstracted away local IT. Now, generative AI abstracts away much of the UI itself: instead of learning where features live, users talk to a conversational layer that figures out which API calls to make in the background.
That has consequences:
- Vendor power increases. If your team relies on Gemini prompts that touch mail, docs and spreadsheets simultaneously, it’s harder to migrate to a rival. Exporting data is not enough; you’d need to rebuild workflows and “prompt recipes” elsewhere.
- Interoperability shrinks. AI features are deeply tied to proprietary data schemas and logs. Open standards like ODF or simple CSV exports don’t capture the behavior of an assistant that “knows” your organization.
- Skill distribution changes. Power users who already build complex Sheets or write good specs gain a force multiplier. Others may become more dependent on opaque AI suggestions they don’t fully understand.
Compared with startups, Google and Microsoft enjoy a massive edge: decades of telemetry about how people actually work. Every keystroke in Gmail or Sheets is potential training signal (within privacy and contractual limits). Workspace Intelligence is an attempt to cash in on that behavioral data – not by selling it, but by using it to entrench Google as the default environment for desk jobs.
The European and regional angle
For European organizations, the question is not just "does this boost productivity?" but "can we deploy this without violating half the EU rulebook?"
By design, Workspace Intelligence touches personal data (emails, calendars, chats) and can shape decisions about employees and customers. That brings it into the crosshairs of GDPR and, soon, the EU AI Act. Depending on how it is configured, an AI that summarizes performance feedback or suggests who to invite to a meeting could fall into the category of workplace‑related AI systems that face stricter obligations.
European data protection authorities have already scrutinized cloud productivity suites over telemetry and data transfer practices. An AI layer that “sees” even more could trigger new waves of DPIAs, contractual addendums and, in some cases, resistance from workers’ councils and unions.
At the same time, the EU’s Digital Markets Act (DMA) is pushing large gatekeepers to open up and avoid anti‑competitive bundling. There’s an open question: when AI becomes a core part of office software, will regulators treat assistants like Gemini as a tied add‑on that disadvantages smaller competitors in the European SaaS and open‑source ecosystem (OnlyOffice, Nextcloud Office, Collabora and others)?
For European SMEs and public bodies that already use Workspace, these tools are tempting – but they may need strict configuration: limited scopes for data access, clear logging of what the AI did, and local policies on when AI‑generated content is acceptable (and when a human must review or rewrite).
Looking ahead
Expect Google’s AI "intern" to become a lot less optional over time.
In the near term, we’ll likely see:
- Deeper automation: multi‑step workflows like “turn this email thread into a project plan, schedule the kick‑off and draft a status report every Friday” executed mostly by Gemini.
- More proactive behavior: instead of waiting for prompts, Workspace Intelligence will surface suggested emails, documents or follow‑ups based on your activity.
- Tighter admin levers: finer‑grained controls, audit logs, and regional data handling to calm European regulators and big‑enterprise CISOs.
The open questions:
- Pricing and accessibility: will full AI features remain a premium add‑on? If so, we may see a productivity divide between teams that can afford AI seats and those that can’t.
- Quality and accountability: hallucinations in customer emails or financial spreadsheets are not just embarrassing; they are legal and financial risks. How will Google let organizations trace which outputs were AI‑generated, and under what context?
- Labor relations: as AI takes over more coordination and reporting, will European works councils treat it as a change in working conditions requiring negotiation? We’ve already seen early versions of this debate around monitoring software.
For readers, the practical move is twofold: start experimenting personally to understand the capabilities and limits – and, if you influence IT policy, push for clear internal rules before the "AI intern" quietly becomes an invisible boss.
The bottom line
Google’s new Workspace AI is less a helpful intern and more a new infrastructural layer for knowledge work. It will likely make many tasks faster and more tolerable, but it also deepens dependence on Google, raises fresh privacy and regulatory questions in Europe, and accelerates a shift in which skills are rewarded at work. The real decision for organizations is not whether to use it, but on whose terms. If your company adopts an "AI intern", what guardrails – technical and cultural – will you insist on first?
