by Agmo Studio
Lark Base looks like a spreadsheet, but its AI tools can write, translate, summarise, tag and extract data inside the table itself. This guide explains what Lark documents, how to try it on your own data, and where teams in different industries could use it.
Lark Base AI AI field shortcuts Smart toolbar
What Lark Base AI can do inside a table
Lark Base is Lark's table-style database. Rather than copying text into a separate AI chat and pasting the answer back, you can ask the AI to work on the cell or record in front of you.
Guides from a Lark partner describe a smart toolbar that opens when you click a text cell in grid view. It offers actions such as Fill, Polish, Translate and Summarise, and a menu with more options such as Extract Info, Smart Tags and Quick Ask. Lark's own Base AI page also describes one-click summaries of text.
- Fill: the AI writes content based on the other information in the same record.
- Polish: improve writing, fix grammar, make the text longer or make it shorter.
- Translate: a partner guide lists 12 target languages, including English, Simplified Chinese, Spanish, Japanese and German.
- Summarise: a short summary of the cell, with insights drawn from the record and the wider table.
- Extract Info and Smart Tags: pull out key details and generate tags from a record's content.
- Quick Ask: type a question or instruction and the AI answers using the record's information.
From one-off prompts to AI fields
The toolbar is handy for occasional use. For work you repeat on every row, Lark describes AI field shortcuts: preset fields built for common business scenarios. According to a Lark blog, these fields can classify data, translate text and pull key information out of images or PDFs. The same blog gives a property-management example where tenants are tagged automatically as commercial, residential or VIP.
Lark's developer changelog also describes a configurable AI field extension. It says an existing text, single-select or number field can be set up to generate, summarise, classify, translate or extract content from other fields in the same record. That matters because the AI result then lives in a normal column that you can filter, sort and chart like any other.
Ideas for putting AI columns to work
These are suggestions for how a team could apply the features above, not results Lark has published. Always test on a small sample first.
- Customer feedback: keep reviews and survey comments in one table, add a translated column for multilingual comments and a summary column for the key points.
- Supplier paperwork: attach invoice PDFs or photos and use extraction to fill amount, date and supplier fields, then check them against the original.
- Product or menu copy: use Fill and Polish to draft descriptions from a record's attributes, then translate them for other markets.
- Enquiry handling: tag incoming enquiries by type so the right person sees them first.
- Meeting follow-ups: summarise long notes into a short status column for a weekly review.
Good habits before you rely on AI columns
AI output can be wrong or incomplete, so keep a person in the loop for anything that affects customers, money or compliance. Spot-check a sample of rows, keep the original text in its own column, and avoid putting sensitive personal data into a table unless your company policy allows it.
Features, plans and language coverage differ by region and Lark version, and some menu names may have changed since the guides above were written. Ask your Lark admin which AI features are switched on for your workspace.
Try it: step by step
- Open a table in Lark Base and switch to grid view.
- Click a cell in a text field to open the smart toolbar.
- Choose an action such as Fill, Polish, Translate or Summarise, and read the result.
- Insert or replace the cell content, or regenerate if the answer is not right.
- Open the menu for more options, such as Extract Info or Smart Tags, and try them on a few rows.
- For work you repeat on every row, look in the field type options for AI field shortcuts and set one up on a test table.
- Spot-check the results against the original data before you share the table.
Lark features, plans, languages and availability change and can differ by region and plan, so check your own Lark admin settings and Lark's Help Center before you rely on a feature. Review AI output before you use it. This article is compiled with AI assistance from Lark's published material and written in our own words.
Want help rolling out Lark in your business? Agmo Studio is a Lark partner in Malaysia: read about our Lark Malaysia partner services, or book a demo. More from the Lark AI Corner.