← back to writing
// writing /07 · google workspace studio · low-code ai · advisor model

Google just commoditized half of my services business. I am teaching my clients how to do it themselves.

Google Workspace Studio turned a paid engagement into a Tuesday afternoon. I used to charge real money to wire up these automations. Now anyone in the company can. My job is no longer building them - it is telling you which ones are worth building.

By · PrincipalDec 16, 2025· 5 min readUpdated: June 2026
Google Workspaceautomationlow-code AIGeminiadvisor modelprocess design

A short answer first. Google shipped Workspace Studio, and a service I used to sell for real money - building chained, AI-augmented automations inside an organization - just became a drag-and-drop feature inside the product your team already pays for. As recently as the end of 2025, this work needed an engineer who could read API docs, wire up auth, and stitch the steps together. Now any operations lead with a coffee can build a flow that takes a Google Form, posts to Chat, opens an Asana task, fires an email, and routes the attachment to Drive. The build is no longer the hard part. Picking the right process to automate is.

So instead of hiding the tool from my clients, I am teaching them how to use it. My job moved up the stack. That is fine. That is the only honest reaction.


What actually changed

Until Workspace Studio, an "AI automation" inside a mid-sized company looked like this: someone described a manual process, an engineer mapped it to APIs, wrote glue code, hooked it to a model, added a queue, debugged auth for two days, and shipped a fragile chain that broke the next time Google rotated a scope. It was a real engagement. People paid for it because it was hard.

Workspace Studio collapses that into a visual builder, sitting inside Workspace, with native connectors to Gmail, Drive, Forms, Sheets, Chat, Calendar, and Gemini 3. The auth is already done. The connectors are already done. The model is one node in a flow you drew with a mouse.

It is, painfully, very good. By Google Cloud Next 2026 it was already past three million monthly active users - this is not a beta anyone can ignore. It is also the right move - this work was never the moat. The moat was knowing which process to point it at.


Three automations you can ship today

You do not need me, or anyone like me, to build these. You need an afternoon and a willingness to draw boxes.

1. The auto-recruiter intake

A candidate fills out a Google Form. The flow fires: a Chat message lands in your hiring channel with the highlights, an Asana task is opened against the assigned recruiter with a link to the application, and the candidate gets a personalized "we received your application" email. Zero human hands. Time from application to first internal touch drops from "next business day" to "ninety seconds."

2. The Gemini-powered manager brief

Every morning at 7am, a flow scans your unread inbox, passes the threads to Gemini with a "what does my manager need to know in two sentences each" prompt, and pushes a single digest to your phone. You walk into the day already knowing which three threads matter. The other thirty-seven can wait.

3. The smart attachment archive

Invoices and statements arrive in email all month. A flow watches the inbox, identifies the attachment type, and files each PDF into a Drive folder structured by year and month. On the 14th, when your accountant asks for "everything from March," you send a link instead of spending two hours hunting.

4. Two more I would add to the list

Customer support triage: a Gemini node reads incoming support emails, classifies them by urgency and topic, drafts a first response in the right tone, and queues it for a human approval click. Weekly client status report: a flow pulls last week's activity from Sheets and Calendar, asks Gemini to write the narrative, and drops a draft into a Doc that the account lead just edits.

None of these need an engineer. They need someone who has watched their own workflow long enough to know where the friction lives.


Why I am giving this away for free

Because the value moved. Building the chain used to be the engagement. Now the chain takes an afternoon. What is left, and what is genuinely hard, is the diagnosis: of all the manual work happening inside your company right now, which five processes are worth automating, in what order, with what guardrails, and what is the second-order effect on the team whose job changes when you do it.

That is the part you cannot draw with a mouse. That is the part that needs someone who has watched a hundred of these go right and a hundred go wrong. And that is the part I will keep selling, happily, for as long as anyone wants to buy it.

If you can build the automation yourself in an afternoon, you should. The honest thing for me to do is tell you that and then help you pick the right one.


The honest take

Every services business has a layer that gets eaten by the platform eventually. The smart move is to notice when it happens and walk one rung up the ladder before the rung you are standing on disappears. Workspace Studio just ate the bottom rung of the AI-automation services market. I am not sad about it. I was tired of charging for plumbing anyway.

The new job is process design. Which workflow is worth automating, which one will collapse if you do, where the human stays in the loop, what KPI you are actually trying to move. The tooling is finally good enough that the conversation can skip "can we build it" and land on "should we, and what happens to the team when we do." That is a better conversation.

Go build the three flows above this week. You will save your team hours by Friday. When you hit the fifth flow and you realize you are automating the wrong thing, call someone who has seen this movie before.


FAQ

Q
Should I cancel my current automation vendor?
Not on Monday morning. Audit what they actually do for you first. If they are wiring up Workspace-native flows that Studio can now build in an afternoon, the answer is yes, eventually, and you should be using the savings on process design rather than on a different vendor. If they are running cross-system pipelines that touch Salesforce, custom databases, or non-Google SaaS at depth, Studio will not replace them yet. The honest test: ask them to show you the last flow they shipped and rebuild it in Studio. If they can do it with you in an hour, you have your answer.
Q
What can Workspace Studio not do yet?
Deep integration with non-Google systems. Anything with real-time SLAs that needs custom error handling and retries. Workflows that need to write back to a database you own. Anything that needs a real engineering review for compliance, audit logging, or PII handling at scale. If your automation has the word "production" in front of it and a regulator behind it, you still want an engineer in the room. For the other ninety percent of internal ops work, Studio is enough.
Q
Why are you giving this away for free?
Because the value is no longer in the build. It is in knowing which process to point it at, which one to leave alone, and what happens to the team when the process changes. I get paid to answer the second question, not the first. Hiding the tool would only make me look slower than my clients eventually figure out, and that is a worse business than just telling the truth.
Q
Who in my company should be building these flows?
Whoever owns the process. Not IT. Not "the AI champion." The person who currently does the manual work or manages the person who does. They have the context to know what good output looks like, where the edge cases live, and when the flow has drifted. Give them an afternoon, a Studio license, and permission to break things in a sandbox folder.
Q
What is the biggest mistake you see when teams start using tools like Studio?
Automating the wrong process first. The most-automatable process is rarely the most-valuable one. Teams pick "send an auto-email when a form is submitted" because it is easy, and ignore the four-hour weekly reconciliation that everyone hates. The right first flow is usually the one your best person has been complaining about for six months. Start there.
Q
How do I know when I actually need outside help versus building it myself?
Three signals. One: you have automated three or four flows and you cannot tell which one is moving a real number. Two: a flow you shipped is now load-bearing and you realize nobody owns it. Three: the next automation on your list touches a system Workspace Studio cannot reach, or it touches data with compliance weight. Any one of those, call someone. The other times, build it yourself.

Written by Michael Fleicher, Principal at Bina Labs. Two-time CTO. We used to build these automations for clients. Now we help clients decide which ones to build, in what order, and what to do with the time they win back. If you have a list of processes and you are not sure which one to automate first, start here.

// next step

Want a build, not a roadmap?

We embed senior AI engineers into B2B teams. 4-week paid pilot. Working code in your environment by week four, or we walk.

Start a conversation →Read the manifesto
// more writing
Feb 10, 2026 · 8 min read
Salesforce promised 1 billion AI agents. They sold 3,000.
Oct 18, 2025 · 7 min read
GladBites made $1.6M in 90 days with no kitchen.
Sep 28, 2025 · 4 min read
An AI agent that has to fund its own tokens.