Opal
“An AI app builder, strong at automation” — a workshop that builds tools when you ask
What is Opal?
Opal is an experimental tool from Google Labs. Describe what you want in plain words, with no coding, and it builds a small multi-step AI app that you can share with colleagues by link.
- 1Describe it in words and a multi-step AI app is built
- 2Share the finished app as a link and colleagues use it as is
- 3It is an experimental Google Labs product, so it is free for now
What sets it apart
These are the differences we found by running the same material through several tools in our training sessions.
Just describe it and the app appears
Describe it in plain words, such as “a tool that takes a company name, finds its recent news, and outputs a five-line summary and three meeting questions,” and that flow is built on screen. The key is that you get a multi-step tool without writing any code.
You can share what you build
Share a finished mini app by link and colleagues can use it as is. Unlike prompts copied around in chat, recipients only fill in the input fields, so the same quality spreads across the team.
You can see the flowchart
Each step sits in a block, and you can follow on screen what flows where. Unlike a long instruction dropped into a chat window, you can pinpoint which step went wrong and fix it, which makes this good practice for anyone new to automation.
⚖️ The weaknesses are clear, too
It is an experimental Google Labs product, so features and availability may change without notice, and no pricing policy has been announced yet. Dedicated automation tools connect to business systems far more broadly, so it is too early to build core business processes on it.
💡 In plain EnglishInstead of retyping a prompt every time, think of it as a workshop that hardens that prompt into a small machine with a button.
The last hurdle in AI training is that nobody can reuse the instructor’s good prompts on their own. Turn a prompt into an app and share it as a link, and teammates only fill in the input fields — no explanation needed.
Same job, different approach
The prompts we teach in class don’t get used on your team
The old way
- Polished prompts get shared with the team over a messenger app
- Everyone copies it and edits it a little differently
- Result quality varies from person to person
- Two weeks later, no one is using it
With Opal
- Turn that prompt into an app with two input fields
- Send the team just one link
- Team members just fill in the company name and purpose, then run it
- Results share one format, so they are easy to combine
Common uses
Client Pre-Meeting Research Tool
Enter a company name and the app finds recent news and the company’s business, organizes it in a set format, and even suggests questions to ask in the meeting. You then share it with your sales team.
Match internal document formats
Build an app that turns pasted meeting notes into a draft in your company’s report format. No need to explain the format every time.
Small tools for training exercises
Build the workflow you learned in class into an app on the spot and share it with participants as a link. No installation or sign-up instructions needed.
Who it’s for
Best for
- ✓Training or planning managers who need to spread good prompts across their team
- ✓Anyone who wants to build their own small tool without coding
- ✓Anyone who already works with a Google account
Better served by other tools
- –Anyone who needs full-scale automation that connects to internal systems
- –Anyone who wants to rely on it for ongoing work that can’t tolerate feature changes
- –Environments where every screen must be in Korean
Can we use it at work?
The questions that adoption leads and security teams ask first.
To confirm: check Google’s official terms for how this experimental tool handles data, whether data is used for training, and where it can be used on organization accounts.
10-minute quick start
- 1
Go to opal.google and sign in with your Google account. Another company’s product has a similar name, so check the address before you enter.
- 2
Open one of the example apps already available and look through the steps it is built from.
- 3
Pick a prompt you use over and over, describe it in plain words, and turn it into an app.
- 4
Rename the input fields in words your colleagues will understand, then send the link to just one person first.
Copy-and-paste prompts
Replace only the [bracketed] parts, then paste it in.
Build me an app that takes a company name and meeting purpose in input fields → searches the web for that company’s recent news and business → and outputs, in order, a five-line summary, three points where we could make a proposal, and three questions to ask in the meeting.
Build me an app where I paste meeting notes into an input field → it sorts them into a table of three parts: decisions made, who does what by when, and what is still undecided → and gathers any item with an unclear owner separately under “Needs confirmation.”
Build me an app that takes the client’s industry, the training audience, and the number of hours in input fields → sets three training objectives that fit those conditions → drafts a schedule with a hands-on exercise idea for each hour → and finally writes the text of a proposal email.
Build me an app that takes a customer inquiry and [our course list] in input fields → classifies the inquiry as quote, schedule, or other → drafts a reply that fits the category → and adds a line at the end saying we reply within 24 hours.
Build me an app that takes two different proposals in two input fields → builds a table comparing them on four axes: price, timeline, scope included, and what is missing → and ends with three “questions we should still ask.”
Using it with other tools
Linking two or three tools changes your work more than any single tool does.
Prompt → team tool
Research app → Proposal draft
Before you start
- Another company has a product with the same name. Google Opal opens at opal.google, so check the address before you sign up.
- This is an experimental product. Features and availability may change, so don’t build the backbone of your everyday operations on it.
- Instructions inside a finished app can be seen by anyone who receives the link. Do not put confidential information in them.
- If you put too many steps into one app, it becomes hard to find where things went wrong. Build short apps and chain them together.
Pricing
For individual use
Sign in with a Google account to use it. No separate paid plan has been announced.
Checked 21 Sep 2026 · Opal is an experimental Google Labs product, so it has no dedicated official pricing page. Neither the official site nor Google’s official announcement shows a price, so we list it as “free.” Pricing can change without notice — check the official page before you pay. Opal official site ↗ · Google announcement — Opal expands to 160+ countries ↗
Agentic AI — the version that does the work itself
In February 2026, Opal added an “agent step” that picks its own path instead of following a fixed flowchart.
Agent step
Google’s official announcement explains that “the agent understands the goal and chooses the tools and models it needs to reach it.” It calls web search when it needs material, calls a video generation model when it needs video, and if information is missing, it **asks first to fill the gap before proceeding**.
- No installation needed — when building a generation step in Opal, choose an agent instead of a model.
Setup steps and conditions are based on the official documentation. Google announcement — Opal agent step ↗
Where Opal falls short
We don’t just list the positives. Here is what fell short when we used it.
- Many people share the same name, which makes it hard to explain
Search for the name and other companies’ products with the same name show up first. Every time you recommend it to a colleague, you have to send the Google Labs address separately, which makes it tedious to roll out to many people.
- No one knows how long it will stay this way
Because it is experimental, neither pricing nor usage limits have been announced. It is free for now, but conditions could appear one day, which makes you hesitate to rely on an app you worked hard to build.
- Editing one step breaks others that were working fine
Because you edit by describing changes in words, fixing just one spot can sometimes change other steps that were working fine. You end up needing to build in short steps and check often.
Tools to compare
GeminiGoogle’s all-purpose AI→
LovableEasy vibe-coding helper→
ZapierThe world’s largest workflow automation tool→
n8nAutomation and AI integration→FAQ
Last updated 22 Sep 2026 · The ratings on this page come from John Kim, CEO of Nanum Consulting, who has used these tools firsthand in training and real work.
Opal as an organizational capability
Send us a message and we’ll reply within 24 hours. We’ll propose a curriculum tailored to your industry and job roles.
Courses in this fieldAI Agents & Work Automation: Hands-On Training