AI commands in the Prompts panel

The Prompts panel lets you edit your list or ask questions about it by typing plain instructions instead of clicking through menus. Open it via More → Prompts in the list header, then type into the Ask about your data... field.

What you can do with it#

Edit your data by instruction. Select rows by a condition ("select personal emails"), tag rows ("mark decision makers as key-contact"), set a status the list has ("set status of selected to Qualified"), hand rows to a pipeline ("send selected to HubSpot"), update or clean values ("trim whitespace in company names"), unify inconsistent spellings ("normalize country names" — this one shows you the proposed groupings to approve before applying them), or fill columns from existing enrichment data ("use the industry field from enrichment data"). It also understands enrichment requests ("find emails for selected", "enrich company data from domains"): it starts the same job as the selection bar's Enrich, which picks each record's cheapest route and fills empty cells as results arrive. Requests to re-enrich filled cells are declined — see "Which Enrich button do I use?"

Values a command writes into cells are suggestions at first: the cells are highlighted and the list header shows N suggestions with Apply · N and Discard. Nothing is written to the list until you click Apply (see "Saving is automatic: undo and AI suggestions"). Commands that hand records to a pipeline, set a status, mark rows ignored or start an enrichment first show a dialog Run this action? — nothing happens until you confirm it.

Ask questions about your data. Ask things like "how many contacts are from Germany," "what countries are there," or "show statistics by country," and it answers directly rather than changing anything in the grid.

For a longer list of example sentences in English and German, see "AI command examples (English & German)."

What it sees#

The Prompts panel sends your instruction, your column names and types, your total row count, the names of the list's statuses and of your pipelines, and — only for commands that need to judge actual values — the distinct values of the relevant column(s). Your individual cell values aren't sent just to understand a request. See "What the AI sees of your data" for the full breakdown.

Cost per reply#

Every AI command is billed by usage, not a flat per-command price — it depends on how much the request needed to process. Each reply ends with a small note showing the exact number of credits it used, so you always see the cost after the fact. Typical requests cost a small fraction of a credit (see "Price list: what every action costs" for how usage-based AI pricing compares to other actions); a request that also triggers a paid enrichment adds that enrichment's own credits on top. In a workspace, the credits come from the workspace owner's balance (see "Credits, plan and limits in a workspace"). If the balance is too low for the request, you'll see an insufficient-credits message before anything runs.

Which AI model answers?#

A model chosen by ListPlus answers every request; you can't pick or change it, and there's no model selector. The small note under each reply shows which model handled it and the credits it used. ListPlus may change the model over time, and how you use the panel stays the same.

Limits#

• One list at a time: a command only ever reads or changes the list you're currently in, never your other lists.
• Commands work on your list as it is in the grid right now, including open suggestions, and only on the rows your current filter or sublist shows. Clear the filter first if you want a command to cover the whole list.
• There's a fair-use cap of 1,000 AI requests per day and 20,000 per month. It counts per person, not per workspace, so colleagues in the same workspace don't use up your requests; hitting it shows a "Limit reached" message.
• You can type in English, German or other languages — the Prompts panel understands the request either way.
• Add "for selected" to a command to target only the rows you've selected in the grid, and combine multiple conditions with "and" or "or" (for example "select where country = France and city = Paris").

Related topics: ai-command-examples · what-the-ai-sees · which-enrich-button · price-list · saving-is-manual.

FAQ#

What is the Prompts panel and how do I open it?#

It's a field where you type plain instructions or questions about your list instead of clicking through menus. Open it with More → Prompts in the list header, then type into Ask about your data....

Can I edit my list by just typing what I want?#

Yes — requests like "select personal emails," "tag decision makers as key-contact," or "normalize country names" are understood as edit instructions and applied to matching rows, the same way clicking through the grid's own menus would. Commands that hand records to a pipeline, set a status, mark rows ignored or start an enrichment first ask "Run this action?" and do nothing until you confirm.

Can I ask questions instead of giving instructions?#

Yes — questions like "how many contacts are from Germany" or "what countries are there" are answered directly without changing your data.

Where do I see how much an AI command cost?#

In a small note right after the reply itself — every AI command is billed by usage, not a fixed price, so the exact number of credits only appears once it's done. See "How much does an AI command cost?" in the price list for typical amounts.

Which AI model does ListPlus use, and can I choose it?#

A model chosen by ListPlus; you can't pick it. The note under each reply names the model that answered and the credits used, but there's no setting to switch models.

Does the Prompts panel see my other lists?#

No. Every request is scoped to the one list you're working in.

Are changes from a Prompts command written to my list right away?#

No — cell values a command changes are shown as highlighted suggestions, and the list header shows "N suggestions" with Apply · N and Discard. Apply writes them to the list (one undo step with ⌘Z / Ctrl+Z); Discard puts the cells back to their previous values.

Does a command cover rows hidden by my filter?#

No. Commands act only on the rows currently shown in the grid, so an active filter or sublist narrows what they touch — clear the filter to include the whole list. Open AI suggestions are included: a command sees the grid as it is now.

Is there a limit to how many AI commands I can send?#

Yes, a fair-use cap of 1,000 AI requests per day and 20,000 per month, counted per person rather than per workspace, so colleagues don't use up your share. If you ever hit it, wait a while or contact support@listplus.ai. The credits for each request still come from the workspace owner's balance.

Can I ask the AI to enrich my list?#

Yes. Requests like "find emails for selected" or "enrich company data from domains" start the same job as the selection bar's Enrich: each record gets its cheapest route, empty cells fill as results arrive, and lookups charge only on a hit. You confirm "Run this action?" first, and the AI usage fee comes on top.

Why does the AI refuse to re-enrich cells that already have values?#

Enrichment fills empty cells only, so the AI answers "re-enriching filled cells is not available". Clear the cells you want looked up again and ask once more; only an AI Analysis from the Enrich menu re-analyses filled cells.

What is the "Large operation" question before an AI command?#

It is a confirmation that appears when a command would process more than 500 records ("This will process N records"). Click Proceed to run it or Cancel to stop.

Where do I see the result of an enrichment I asked the AI for?#

In the task indicator at the top of the screen, and in the grid as the cells fill. The chat only confirms the start ("Enrichment started for N records — the cells fill as results arrive").

Will the AI look up a company from a Gmail address?#

No. Free-mail domains such as gmail.com are never sent to a company lookup, because they would only return the mail provider, not the person's employer.