News from Google Next 2026: All new features Google is planning to launch within Google Workspace Studio, new success stories and questions
Table of Contents
Summary
Google made Workspace Studio generally available at Google Next 2026, citing 3.5 million monthly active users and 170 million tasks automated in a single month — a 700% growth in three months. The announcements included agentic loops, Gemini agent integration, MCP servers, and a governance dashboard that is, notably, still on the roadmap rather than in the product.
This article covers what was announced, what Karcher’s real-world rollout actually looked like, and what Google didn’t say on stage: that Studio cannot send emails automatically, cannot loop over data, cannot assign tasks to other users, cannot trigger from external systems, and caps most paying customers at 400 flow runs per month. Against that backdrop, the announced “agentic loops” — described as fanning out across a hundred parallel tasks — raise an obvious architectural question that went unasked at the keynote.
Google Workspace Studio Goes GA
At Google Next 2026, Google made it official: “Workspace Studio is now generally available for everyone, everywhere“. First announced in January 2026, the tool has moved out of early access and into the hands of Google Workspace’s three billion users — a platform that now counts over 13 million paying customers and a marketplace with more than six billion apps installed.
The pitch is straightforward: AI-powered automation for people who don’t write code. Powered by Gemini 1.5, Workspace Studio sits deeply embedded inside Gmail, Google Calendar, and Google Chat — giving users the ability to build automated workflows that tap directly into their emails, labels, and meeting schedules. Build something useful, and you can share it with your entire organisation.
Google Workspace Studio Adoption By the Numbers
3.5 million
monthly active users
70 million tasks
automated in a single month
700% growth
in automated task volume since January
These are not small numbers. Whether you think Workspace Studio is the right tool for enterprise automation or not, the adoption curve is real.
What's Actually New In Google Workspace Studio In 2026
Beyond the GA announcement, Google outlined a roadmap that signals where Workspace Studio is heading.
1. Agentic Capabilities
The most significant shift is the move toward “agentic workflows”. Users will be able to `@mention` specialised Gemini Enterprise agents directly inside the Gemini side panel — for example, pulling in an agent to triage an inbox, run a feasibility check, or manage a procurement request. These agents will work across Google Chat, Gmail, and Workspace Studio itself, with interoperability being a stated priority.
Google also announced “agentic loops” — the ability to “fan out” and execute high-volume parallel tasks, such as researching a hundred contacts simultaneously. We will return to this claim shortly, because it deserves significant scrutiny.
2. New Integration Categories
Two new integration categories are coming to Workspace Studio:
- NotebookLM — allowing agents to tap personal or catalogued data as context
- Gemini Enterprise agents — enabling custom, specialised agents to be called within any flow
Additionally, Google confirmed work on connecting to SAP, Jira, and other back-end enterprise systems — which, as we’ll discuss, is where things get complicated.
3. Admin and Governance Tools
Admins are getting a centralised Agent Management Dashboard to track usage at scale. Upcoming security features include runtime Data Loss Prevention (DLP), user confirmation for sensitive agent actions, and built-in abuse prevention.
Google is also launching official Workspace MCP servers, a CLI, and plans to make AppSheet a core Workspace service later this year — bringing enterprise-grade auditing and reliability to app development within the ecosystem.
Your Ultimate Tool For Google Workspace AI Agents & Workflows
Zenphi is your professional-grade alternative to Google Workspace Studio . It allows you to design. deploy, monitor and manage your AI agents and AI workflows withing Google Workspace and beyond. Book a call to learn more.
Karcher's Story: The Human Side of Rollout
To ground the announcements in reality, Google brought Morena, a Workspace AI Enablement Lead at Karcher — a German manufacturer of cleaning and catering technology with 17,000 employees — to share what actually happened when they handed the tool to 400 early adopters.
The results were genuinely impressive in places. One user saved 70% of manual processing time on business-related tasks. Across the board, users reported saving one to two hours per week on repetitive administrative work. The showcase flow — an automated feature request pipeline that took ideas from Chat, through a brainstorming agent, a technical feasibility agent, and a UX agent, before delivering a draft user story — reduced what was hours of manual work to under two minutes, an estimated 90% time saving.
Users described the experience as “building with toy blocks.” The no-code interface lowers the barrier to entry significantly: no need to understand automation logic deeply, just know your business processes and match them to the tool’s capabilities.
Two Hurdles Karcher Identified
Morena was candid about where things got difficult. Two patterns emerged among users:
Karcher’s response was a two-track rollout: a Core Track (one intro session, one workshop — done) and a Flex Track (an open calendar of best practice sessions, workshops, and peer-hosted demonstrations for those who want to go deeper).
The Technical Reality: What Workspace Studio Actually Can't Do
Before evaluating Google’s announcements honestly, it’s worth establishing the baseline. Independent testing of Workspace Studio against real operational workflows has revealed a consistent pattern of capability gaps that the Google Next keynote did not address.
These limits are not edge cases. Processing a busy support inbox, running invoice extraction across a month’s documents, automating employee onboarding across dozens of new starters — any of these can exhaust monthly quotas at the Business Standard tier within days, not weeks. And when you don’t know what the daily limit is, planning around it is educated guesswork.
The Agentic Loops Question Nobody Asked on Stage
Here is where the Google Next announcements and the technical reality diverge most sharply, and where a direct question needs to be asked.
Google announced agentic loops — the capability to “fan out” and perform high-volume tasks in parallel, with the illustrative example of researching a hundred different people simultaneously.
Let’s hold that against the current architecture:
- Studio has no native looping or iteration logic in its current build
- Monthly flow run limits top out at 10,000 for AI Ultra Access — the most expensive tier available
- Each flow is capped at 20 steps
- There is an undisclosed daily execution ceiling that Google declines to quantify
If researching a hundred contacts simultaneously requires a hundred flow executions, an organisation on Business Standard has just consumed 25% of their monthly quota in a single agentic task. An organisation on the entry tier has already exceeded it entirely. And that assumes the looping architecture needed to support fan-out is even built — which, as of current independent testing, it is not.
The gap between “agentic loops that fan out across a hundred tasks” and “20 steps per flow, 400 runs per month” is not a minor implementation detail. It is a foundational architectural question. How does Google plan to deliver agentic scale on a platform with these structural constraints? The session at Google Next 2026 did not offer an answer. It deserves one.
On Governance: Arriving Fashionably Late
Google spent meaningful time at this session talking about upcoming governance features: admin dashboards, agent attribution controls, DLP at runtime, and security standards that developers are now being encouraged to build toward.
This is progress. But it’s worth pausing to ask: why is enterprise-grade governance still upcoming for a tool now in general availability?
Zenphi, on the other hand, has offered exactly this — admin-controlled automation, full auditability, governed deployment, and centralised visibility — since they launched. Governed automation wasn’t a feature Zenphi added once it reached scale; it was foundational. The irony of Google announcing governance tooling after general availability, while positioning Workspace Studio as enterprise-ready, is not subtle.
Enterprises don’t typically treat governance as a Phase 2 nice-to-have. It tends to be the reason procurement signs off in the first place.
The "Over-Ambition" Problem Deserves a Closer Look
The framing of over-ambition as a user problem is worth interrogating. Karcher’s example of the worst-case over-ambitious user was someone who tried to build a full SAP system inside Workspace Studio. Fair enough — that is a mismatch of tool and task.
But here is the thing: wanting to automate a procurement approval workflow, or an HR onboarding process, or a multi-step finance exception handling flow, is not over-ambition. It is the actual point of enterprise automation.
The irony is substantial. Karcher’s own most celebrated demonstration — the feature request pipeline — involved four sequential AI agents, feasibility checks, and human review gates. That is not a simple flow. And yet when users try to apply that same logic to processes that touch ERP systems, HR platforms, or compliance workflows, they’re counselled to right-size their ideas.
It is also worth noting that Studio cannot assign tasks to users other than the person running the workflow, cannot trigger from external systems, and cannot iterate over a list. So a user who wants to route a leave request to their manager for approval is not being over-ambitious. They’re being told by the platform’s architecture — quietly, through what it simply cannot do — that they’ve asked for too much. Calling that “over-ambition” is a generous framing.
170 Million Tasks — But What Kind?
The 170 million monthly automated tasks figure is attention-grabbing. But the nature of those tasks matters. Summarising an email. Preparing a meeting brief. Drafting a reply. These are genuinely useful — nobody is dismissing the value of lightening the cognitive load of individual knowledge workers.
But they are not the same as automating a multi-system approval chain, synchronising records between a CRM and a billing platform, or triggering actions in a legacy ERP based on business rules. Studio’s deep Gmail and Calendar integration is its genuine strength — and, at present, its ceiling.
For the 170 million tasks automated, it would be instructive to know how many crossed a system boundary, involved a second user, exceeded 20 steps, or required any form of iteration. The number almost certainly tells a story about personal productivity assistance, not enterprise process automation. And there’s nothing wrong with personal productivity assistance — as long as that’s what’s being sold.
What to Watch
The roadmap is substantive. MCP servers, CLI access, AppSheet as a core service, expanded enterprise integrations, and governance tooling suggest Google is playing a long game. If even half of the announced capabilities ship on time and function as described, Workspace Studio will be meaningfully more capable by the end of 2026 than it is today.
But “generally available” implies a level of enterprise readiness the tool hasn’t yet earned — particularly for organisations whose automation needs extend beyond the walls of Google Workspace, require interaction between multiple users, or depend on processes longer than 20 steps.
The agentic loop announcement is exciting. The platform it was announced for, today, structurally cannot support it. That is the gap worth watching — and worth asking about directly, rather than waiting for the next keynote to find out.
This article is based on the notes from Google Next 2026 and a session recordings at the workshop Automate Work with Google Workspace Studio: https://www.googlecloudevents.com/next-vegas/session-library?session_id=3924724&name=automate-work-with-google-workspace-studio
Vahid Taslimi
Vahid Taslimi is a software engineer turned SaaS executive and the co-founder and CEO of Zenphi, a no-code workflow automation platform built for Google Workspace. With over two decades of experience in software development and product leadership, Vahid has dedicated his career to making enterprise-grade automation accessible to IT and operations teams
