Tables are only available on PRO plans or higher, not CORE. Running table analysis costs tokens.
What Tables are for
Use Tables when you need to ask the same question across many project sources and compare the answers side by side. In &AI, each row represents a document or product, and each column represents a prompt, search, or other analysis instruction that produces a result for every row.
This workflow is useful when you want to:
compare multiple prior art references against the same criteria
extract product attributes across a product set
organize answers into a reviewable grid instead of reading each source one at a time
verify, flag, filter, and export a structured result set after the AI run finishes
Open the Tables workspace
Open the Tables page inside a project to see every table already created for that project. The page includes:
a New Table menu with options for New Table and From Template
a search box for filtering the table list by name or description
a sortable list of existing tables
an availability message if your workspace does not include table analysis
If you already have a table, select it from the list to open its detail page.
Table structure
Tables are designed to feel like a spreadsheet, but the cells are driven by project context and AI analysis instead of manual formulas.
Rows are sources. A row can represent a document or a product from the current project.
Columns are analysis instructions. A column can use a custom prompt, a saved prompt from the library, or a search-based execution mode.
Cells are outputs. Each cell stores the result for one row and one column, along with status, confidence, reasoning, and review state.
Depending on the column design, a result can be plain text, an enum value, a boolean, a number, a float, or a date. Tables can also include citations, confidence indicators, flagged states, and verified states that help you review the output at scale.
Start a new table
You can start in one of two ways:
Create a blank table
Choose New Table, enter a required table name, optionally add a description, and create the table. &AI opens the new table immediately after creation so you can begin adding sources and columns.
Create from a template
Choose From Template if you already have a reusable table structure. In the template picker, search for a template, select it, and create the table from that saved configuration. This is the faster option when your team repeats the same analysis pattern across projects.
Work inside a table
On the table detail page, you typically move through the workflow in this order:
Add sources
Use Add Sources to add documents and products from the current project. The add-sources modal supports both categories in the same action, so one table can mix document-based and product-based rows when that is useful.
Only project documents that have successfully processed are available for selection. Existing rows are excluded from the picker so you do not accidentally add duplicates.
Add columns
Use Add Column to define what &AI should answer for each row. You can:
create a new prompt
choose a saved prompt from the library
run a search-based column such as prior art or product search
When you create a new prompt-backed column, you can define the output type, output constraints, and whether web search or deep research should be enabled for that column.
Run or resume analysis
When the table has unfinished work, the header shows Run All or Resume All. Running the table applies the configured column logic across the current rows and consumes tokens.
Review the results
After results appear, use the grid and cell details to inspect confidence, reasoning, and citations. You can verify cells, flag cells that need attention, edit values manually when needed, and use filters or sorts to focus the review.
Use controls that help you review at scale
The table actions header exposes several controls that are important during analysis and review:
Sorts to order the table by source or by structured output columns
Filters to create structured filters or AI filters
Run All or Resume All to process incomplete work
Cancel while a table run or pipeline run is still active
overflow actions for Highlight Verified, Highlight Flagged, Show Hidden Columns, Show Hidden Rows, Wrap Text, export, and template-saving actions
These controls are especially useful once a table grows beyond a small number of rows and columns.