- Practicaly AI
- Posts
- 🧠 Gemini Embeddings, AI Docs, and Notion Dashboards
🧠 Gemini Embeddings, AI Docs, and Notion Dashboards
Today in AI: Google upgrades its productivity suite with Gemini, launches a multimodal embedding model, and Notion introduces a new way to visualize your databases.
👋 Hello hello,
Every productivity tool now seems to be asking the same question: what happens when AI actually understands the context of your work?
This week’s updates push that idea forward. Google is embedding Gemini directly into its core apps; its new embedding model understands multiple media types at once, and Notion is making it easier to visualize entire databases in one place.
In short, the tools we use every day are getting much better at understanding our data.
Let’s get into it.
🔥🔥🔥 Three Highly Curated AI Stories
Google is rolling out a new Gemini-powered experience across Docs, Sheets, Slides, and Drive.
The update introduces AI overviews, AI-generated slides that remain fully editable, and improved grounding sources to make writing more context-aware. The goal is to help the AI understand the content inside your files rather than treating each prompt as a standalone request.
This means users can generate presentations, summarize documents, and write content while the system references existing work across their files.
The new experience is now available to Gemini 1 Pro and Ultra users.
Notion just announced a new feature called Dashboards, designed to provide a high-level view of database information.
Dashboards combine multiple views such as boards, tables, charts, and timelines into a single interface. The goal is to make it easier to monitor important data without switching between different views.
For teams that rely heavily on Notion databases, this creates a more visual way to track projects, metrics, and workflows in one place.
Google also introduced Gemini Embedding 2, its first embedding model designed to work natively across multiple types of media.
Embeddings are the technology used to convert information into numerical representations so machines can search, compare, and understand data more effectively. This new model expands that capability beyond text.
Gemini Embedding 2 can process text, images, video, audio, and documents in a single embedding space, which allows developers to build systems that retrieve and analyze different media types together.
The model supports several formats, including images, short videos, audio, and PDFs. It also captures semantic meaning across 100+ languages, enabling applications like semantic search, retrieval systems, and large-scale data analysis.
The model is currently available in public preview through the Gemini API and Vertex AI.
🔥🔥 Two AI Tools Worth Knowing
Claude now supports connectors that allow it to work with external tools like Figma.
If you enable the Figma connector, you can ask Claude to generate diagrams for topics you are trying to understand. Claude then uses the connector to create the diagram directly inside the workflow instead of describing it in text.
This is especially useful when learning complex topics that benefit from visual explanations.
Perplexity’s Computer tool can automate data collection and reporting tasks across different platforms.
One example workflow uses it to build a social media performance dashboard. The system pulls engagement data from multiple channels, analyzes which posts perform well, and categorizes content into top-performing and underperforming groups.
It can also analyze transcripts, track audience engagement trends, and update reports automatically every few hours.
FYI: Perplexity Computer is currently a paid feature, but similar workflows can also be built using tools like Claude Code.
🔥 Things You Didn’t Know You Can Do With AI
Claude now has an Excel add-in that allows the model to interact directly with your spreadsheet data.
Instead of copying data into a chatbot, Claude reads your workbook, analyzes cells and formulas, and performs tasks directly within the file.
Here are a few things you can do with it.
Ask Claude to clean messy datasets, standardize formats, remove duplicates, and highlight errors.
Request insights like top performing categories, anomalies, or trends without building formulas manually.
Select any cell and ask Claude to explain or fix formulas in plain English.
Ask Claude to build models or forecasts directly in the spreadsheet structure.
I wrote a detailed guide breaking this down step-by-step. You can read it here.
Before you go, did today's newsletter help you stay ahead? |
💬 Quick poll: What’s one task you’d want AI to run automatically for you?
Until next time,
Kushank @PracticalyAI
Reply