Skip to main content
The AI chat is the fastest way to start a new Workspace table. Instead of clicking through filters, you simply describe who you want to reach in plain English - and Lusha builds a results table for you instantly.
This feature is under a rollout and will be released to all users by May 2026.

How to access the chat

There are three entry points: “Create new” button - In the left navigation bar, click Create new. This opens a blank Workspace page with the AI chat ready on the left side. Homepage chat input - At the top of your Dashboard home screen, there is a “What table do you want to build?” input. Type your request and click Build. ”+” button in Recents - On the homepage, click the + button next to your Recents section to start a fresh table from the chat.
Create new button, homepage chat input, and plus button entry points for building a table

Step 1: Describe your target audience

Type a plain-language description of who you want to reach. The AI reads job title, seniority, industry, company size, location, technology use, and more from a single sentence. Examples that work well: Tips for better prompts:
  • Be specific about seniority (VP, Director, Manager) when it matters to your ICP
  • Combine industry + location + title for precision: “Marketing VPs at SaaS companies with 50–200 employees in London”
  • Mention technology or funding stage if relevant: “companies that recently raised Series A”
  • You don’t need to be grammatically perfect - the AI understands intent

Step 2: Review the preview table

After you type your request, the AI builds a preview table on the right side of the screen. This is a live view of matching contacts or companies from Lusha’s database. You can scroll through, review results, and decide whether to refine or save. The preview table shows default columns: Contact name, Job title, LinkedIn profile, Company name, Work email, Mobile, Phone, Direct phone, Industry, and Sub industry.
AI chat preview table showing default contact and company columns

Step 3: Refine through conversation

Keep the dialogue going in the left chat box to adjust your results at any time:
  • “Show only companies that have raised funding in the last year”
  • “Narrow to companies with 50–200 employees”
  • “Remove contacts from the UK”
  • “Add a column for company headcount”
  • “Sort by most recent job change”
The AI updates your preview table in real time or gives you step-by-step guidance for more complex adjustments. You can go back and forth as many times as you like - you are only charged for saving, not for chatting.

Step 4: Save your table

When you’re happy with the results, click “Save first X contacts” at the top of the preview table. You can choose to save 50, 150, 500, or 1,000 contacts. Once saved, your table becomes a permanent Workspace table and appears in your left navigation under tables.
Save first X contacts button on the AI chat preview table

Understanding what the AI can and cannot do

When the AI cannot fulfill a request directly, it will tell you and suggest using the manual filter panel instead.

Credit costs

Exploring and refining through conversation is intentionally low-cost so you can experiment freely before committing to a table.

FAQs

Can I combine chat and filters? Yes. Start with a chat prompt to get a rough audience, then click “Filter contacts” or “Filter companies” in the toolbar to apply additional filter criteria on top of your chat results. Can I add rows to a saved table using filters? Yes. In addition to AI Chat and adding similar contacts, you can now expand a saved table using filter criteria. Open your saved table, click Add rows in the toolbar, and choose Use filters to apply filter criteria and add matching contacts or companies. Learn more here. What happens if my prompt returns zero results? The AI will let you know and suggest how to broaden your criteria - for example, expanding the location, removing a seniority filter, or widening the industry. Is there a character limit for prompts? No hard limit. Short, clear sentences work best. If your prompt is very complex, break it into multiple follow-up messages.