Google’s AI features are very confusing

Follow ZDNET: Add us as a favorite resource on Google.

Highlights taken by ZDNET

• Many of Google’s new AI features span multiple workspaces.
• Smart features have replaced the overall development of the Gemini.
• Multiple touch points with a single function can get confusing.

At this point in the AI ​​race, many AI labs have discovered that the real money lies in business use cases: the large, agent-based features that most affect the way large companies move and operate. Google is one of those labs, which could explain why many of the company’s consumer-forward AI features unveiled at I/O on Tuesday feel…

Also: Google’s new Omni AI tool will let you create video for yourself – I’m amazed (and worried)

Not only that, but they are incredibly different from the rest of Gemini’s already multi-layered environment. This I/O was mainly Google’s chance to make everyday AI appealing to users, especially those who doubt how hard the company is pushing it. While these new features offer some convenience, Google has packaged them in ways that may undermine their value and usability.

Many different lives

Google Workspace has received new AI capabilities in I/O, the main two being Docs Live and Gmail Live. And yes, they work the same way.

Like Gemini Live, which lets you use voice to interact with Gemini, Docs Live and Gmail Live let you query each program by voice.

In a demo, a Google employee used the first one to create a Google document on his phone, pieced together from different notes, presentations, and other information about the situation he was talking about. To demonstrate the latter, he asked Gmail questions such as, “What’s going on at school this week?” Gmail Live scanned his inbox and replied that there was a field trip to prepare for.

Also: I compared how Gemini, ChatGPT, and Claude can analyze videos – this model wins

Both features handled complex follow-up questions about scheduling conflicts and changed topics by appearing without confusing related words, such as “field trip” and “trip to Detroit.” But this is the effectiveness of Gemini Live, used in new applications. Why give them different names and place them in certain places? Why not just extend the Gemini Live integration?

None of the Google reps I asked on site could really answer this.

It’s a strange way to make Workspace accessible with AI when, until now, Google has been putting Gemini in everything. Like it or not, it’s a neat, authority-building approach that gives users a one-stop shop for these subtle but simple AI things (that is, if they work). Besides, Gemini on Workspace already covers Gmail, Docs, and Spreadsheets in various capacities — why not put them under this umbrella, at least?

Also: Your Android phone gets the power of an agent with Gemini Intelligence – here’s how and when

By splitting nearly identical functions into separate features, Google risks confusing or overwhelming the very person it needs to convince: a long-time Google customer who is hesitant about AI or unsure if it’s important to them. That’s a lot easier to do because Gemini gives you information from several apps by voice rather than creating a laundry list of standalone product titles.

Google is not alone in this: Microsoft has included Copilot in many existing products, leaving users to explore the differences between Copilot Chat, Microsoft 365 Copilot, and other overlapping names. OpenAI ran the risk of similar confusion when it launched ChatGPT Apps, which are simply an integration of ChatGPT with other apps — not a new product or app creation platform as the name might suggest.

In addition, Docs Live and Gmail Live are being rolled out to AI Pro and Ultra subscribers — who pay $20 and at least $100 per month respectively — and will be in preview for Workspace business users this summer. It’s not clear if they will eventually roll out to all users or to the Gemini itself.

These are the most used cases of private launch, not to mention that users will need to pay for them. Both seem to be available only on mobile, at least for now, focusing on making “on-the-go” tasks seamless.

A very short Daily Brief

Derived from an experiment called CC, announced late last year, the new Daily Brief feature receives information from your email, calendar, and other connected apps to give you a summary of what your day looks like. For those who are concerned about the meetings associated with ideas, it also includes links where in your ecosystem it has released each item of the agenda.

I say feature, but Google called the Daily Brief a “new agent.” Not to bury us in semantics, but unless I’m missing something, this feature doesn’t strike me as living up to the agent standard OpenClaw has set this year.

Also: I tested ChatGPT Plus vs. Gemini Pro to see which is better – and if it’s worth switching

To be fair, the Daily Brief shows a certain level of deep thinking: We haven’t been able to test it ourselves yet, but in two separate demos, it’s gone beyond today to flag up charges coming later in the week. It also determined that the test user was about to renovate his kitchen, suggesting he set up a makeshift cooking station.

But that’s where the “agent” stopped. When I asked if the feature could take action on any of those things, a Google employee told me that the user had to start a conversation with Gemini or log into Spark, Google’s new “agent” that lives (again, confusingly) inside Gemini.

Not to sound like a brat, but … that’s it?

Also: How Google recently revamped Gemini Enterprise in the engineering era – here’s what’s new

Daily information is the product of the agent’s summary work. But is the brief itself an agent? For Google’s definition, I’m not sure (although manual testing may reveal additional agent capabilities). While updating data from your connected apps requires agent capabilities, Daily Brief is the smallest app for those capabilities.

In the grand scheme of AI expectations, Google is also somewhat late.

“One of the things that OpenClaw is known for is its ability to provide you with daily information. People also use Claude Cowork for it,” said David Gewirtz of ZDNET. “It’s certainly not a new skill in the AI ​​world. People have been talking about it in daily forums since agents first came into use.”

Given that context, I have the same question I had about Many Lives: Why not just include the Daily Brief in Gemini as a new (probably long overdue) feature?

AI tools need clarity

The various levels of Gemini’s ability are already different in communication, they are only distinguished by slightly different names. Gemini Intelligence, the Android-specific agent arm, launched earlier this month with agent-based, multitasking capabilities. That’s different from Gemini Personal Intelligence, by the way, which customizes the answers to your question based on the data you give it access to.

Then there’s Gemini in Search and the standalone Gemini app.

Also: This powerful Gemini setting made my AI results more personal and accurate

This setup risks making what Gemini can do so invisible that users simply uninstall. Speaking of technology, Google may separate each arm of Gemini based on how it works in different areas: For a developer or product manager, it may make sense to separate the Android version of the Gemini product.

But from a marketing perspective, maybe that doesn’t need to be seen by the consumer.

Even though AI labs are now following Anthropic’s success and focusing on business, the field of consumer AI tools is still significant — and crowded. Making it difficult for users to keep your products straight is not helpful.