Field types that check, link, and compute your data

Give each column in a table the type that its values need, such as text, currency, a relation, a formula, or a rollup, so that Tables checks every value as your team enters it. Beryl Lens can also fill in AI columns that sort your records into categories or score them.

Get early accessView all products

Field types for each kind of record

Text, number, date, currency, attachment, and select fields hold the basic facts about each record. To connect one table to another, you can add relations and lookups, while formulas and rollups calculate values from the linked records. For values that call for a judgment about each record's content, Beryl Lens fills in AI fields that categorize records, summarize text, detect sentiment, and predict scores. To set a priority score for one support ticket, Beryl Lens compares that ticket with every other ticket in the table.

Add fields that enforce structure and compute answers

The steps below go from a typed field that checks each entry to an AI column that Beryl Lens fills from the whole table, with relations and formulas in between.

The field types that Tables offers

Standard field types for structured data

A spreadsheet cell can accept any text that someone types into it, so a date column can end up holding 'next Tuesday', '2024-03-15', and 'March 15th' side by side. Anyone who sorts or totals that column will first have to clean it by hand.

Tables checks every value against the type of its field as someone enters it, so a number field accepts only numbers and a date field stores every date in one format from its date picker. Select and multi-select fields offer only the options that you define, while checkbox, URL, email, phone, percent, currency, and attachment fields each check their own format.

Relation fields for connected data

If your team keeps a customer's orders, invoices, and support tickets on separate sheets, someone will have to type the customer's name on each sheet and correct every copy whenever the name changes. A lookup formula that points at another sheet can also break when someone inserts or moves a column.

A relation field links a customer's row to the rows for their orders, invoices, and support tickets, each of which opens with a click. A lookup field then shows chosen values from the linked row, such as the customer's name on each order, from the single copy that Tables keeps. Rollup fields summarize the linked rows with a count, a sum, an average, a minimum, a maximum, or a formula that you write.

Formula fields for computed values

A spreadsheet formula refers to other cells by position, such as B4 or C12. Once a formula grows long, the colleague who writes it may be the only person who can tell what each reference means.

A formula field refers to other fields by name, so you can write a margin formula as 'revenue - cost', which gives the same result wherever those columns appear in the table. You can use formulas for arithmetic, text, dates, conditional logic, and totals across related records. As you type, the formula editor suggests field names, so you can build an expression by picking fields from the list.

AI fields from Beryl Lens

When a support team tags 200 tickets a day by urgency, category, and sentiment by hand, its agents spend part of each shift sorting the queue before they can answer it. An AI column that scores each ticket on its own can do some of that sorting, but someone will still have to compare each score with the rest of the queue.

Describe in plain language what an AI column should hold, such as 'categorize this support ticket by urgency', 'summarize this feedback entry in one sentence', or 'score this lead based on engagement signals'. Beryl Lens then compares each record with the rest of the table as it works out the value, so it sets the priority of a ticket against every other ticket in the database. When the source data changes, Beryl Lens recomputes the field and marks any record that receives a new category or score.

Attachment fields with file management illustration

Attachment fields with file management

If your team pastes file links into a spreadsheet, each link can stop working when someone moves the file. If your team keeps the files in a separate drive, someone will have to match each file to its record by hand.

An attachment field stores documents, images, PDFs, and videos on the row itself, where your team can preview each file in Tables. When you connect OneCloud, Tables syncs those attachments with your OneCloud storage, so the same files open from the table and from the OneCloud file manager. Gallery view also uses the attachments as card covers, so your team can browse records by their images.

What typed fields change for your team

Clean data from the first entry

Tables checks each value against its field type as someone enters it, so whoever types a wrong value can correct it on the spot. Your date columns therefore contain only dates, while your select columns contain only the options that you define.

Connected data without duplication

Tables stores each value once and shows it on every linked record through lookups and rollups, so a correction to a customer's name appears on every linked order, invoice, and ticket.

AI fields that use the whole table

Before Beryl Lens assigns a category or a score, or flags an anomaly, it compares the record with every other record in the table, so it can pick out the patterns that appear only across many records.

Formulas written with field names

Because each formula refers to fields by name, a teammate can read a shared formula and see what it calculates. The formula also returns the same result when someone inserts a column.

Fields in practice

300 tickets a day, sorted by an AI field

A support team receives 300 tickets a day. Through an AI field on the tickets table, Beryl Lens assigns each ticket a category, such as billing, technical, account, or feature request, and an urgency score from 1 to 5, based on its subject line and description. Because Beryl Lens weighs each ticket against the full ticket history, it gives a 'can't log in' ticket from an enterprise customer a higher score than the same ticket from a trial user. The team's kanban view groups the tickets by category and sorts each column by urgency.

Result

Beryl Lens sorts each ticket as soon as a customer submits it, so the support agents stop tagging tickets by hand and open their kanban view to the most urgent tickets in their category. Beryl Lens places an urgent ticket from an enterprise customer at the top of its column, so an agent can pick it up before the rest of the queue.

300 tickets a day, sorted by an AI field

Property manager watches spend per building

A property management firm links every maintenance work order to the row for its building. A rollup field on each building's row totals the quarter's spending, while a formula field calculates the difference from that building's repair budget. When a building's spending exceeds its budget by more than 15%, Beryl Lens flags the building in an AI field and summarizes the three work orders that account for most of the overage.

Result

Tables recalculates the rollup as the team logs each work order, so the portfolio director sees an overrun on a filtered grid of flagged buildings before the monthly close. The operations team starts each investigation from the three work orders that Beryl Lens names in its summary.

Property manager watches spend per building

Sales ops scores rep performance with formulas

A sales operations team tracks its deals in Tables, with fields for deal size, close date, stage, and outcome. On each rep's row, a rollup totals that rep's closed deals, averages their deal size, and sums their revenue, while a formula field divides their won deals by their closed deals to give a win rate. Beryl Lens also writes a one-sentence summary of each deal in an AI field for the monthly pipeline report.

Result

The revenue lead builds the monthly performance report from the reps' rollups, so they stop compiling figures from separate systems and a spreadsheet. Since Beryl Lens summarizes each deal, the lead can fill in the report's deal notes without rereading every closed deal.

Sales ops scores rep performance with formulas

Insurance adjuster keeps every claim file on the row

An auto insurer stores its claims in Tables, with attachment fields for damage photos, repair estimates, and police reports. Relation fields link each claim to the assigned adjuster and the scheduled inspection, while a lookup field shows the adjuster's phone number on the claim row. Tables syncs the attachments to OneCloud, so the table and the OneCloud file manager show the same documents.

Result

Each adjuster opens the damage photos and estimates from the claim row, so they stop searching a shared drive for them. When an adjuster assembles a settlement packet, they can take every document from the claim's row, while OneCloud stores the same files for the insurer's claims portal.

Insurance adjuster keeps every claim file on the row

Explore more of Tables

Databases

Design your tables in a visual editor while a managed Postgres or SQLite engine stores every row.

Learn more

Views

Open the same table as a grid, a kanban board, a calendar, a gallery, a timeline, or a form.

Learn more

Sync

Several colleagues can edit the same table at once, with a live cursor for each person.

Learn more
View all Tables features

Choose a type for every column, including AI

Add typed fields, relations, formulas, and AI columns that Beryl Lens fills from the whole table.

Get early accessView all products