Skip to main content

How Laziza AI Uses Job Posting Context to Generate Replies

context, job posting, proposal, how it works, Laziza accuracy, get_proposal

Written by Iryna Grytsaenko

Laziza doesn't generate replies from scratch — she builds them from context. Understanding what data she reads, and what she can't access, helps you get more accurate and useful outputs.


What Laziza Reads Every Time She's Triggered

When you mention Laziza or a wake-up fires, she automatically receives:

  • The full conversation history — every client message and internal note from the beginning of the thread, not just the most recent message

  • Internal team comments — any notes your team has left

  • Client information — the client's name and profile details visible in the room

  • Your team information — team name, the user who triggered the action

  • Dynamic variables — freelancer timezone, location, and room title if available

  • Your Custom Prompt — the tone, rules, and instructions you've configured in Settings → CRM AI Configuration


How She Pulls Job Posting Data

When a reply depends on scope, pricing, or deliverables, Laziza automatically calls the get_proposal tool. This retrieves:

  • The client's full job posting text

  • Your submitted cover letter

  • Your screening question answers

This is why Laziza can reference specific requirements from a job post even if they weren't mentioned in the chat itself.

You can also ask her to pull it explicitly — type @, select Laziza, and write:

summarize the original job posting for this lead


When Laziza Lacks Detail

Laziza can only work with what's in the conversation or the job posting. If the client's message is vague — no budget mentioned, no timeline given, no clear requirements — Laziza will flag the gaps rather than invent details.

Example: if you type @, select Laziza, and write "draft a proposal for this client" but the job post says only "Need a developer" with no further detail, she'll ask you to provide more context before drafting.

How to feed Laziza more context manually:

Add a note before triggering her — switch to Note mode and write directly (no need to tag Laziza for this):

"Note for Laziza: the client told me on a call that their budget is $2,000 and they need this done in 3 weeks. They specifically want a mobile-first design."

Then type @, select Laziza, and ask for the draft. She'll incorporate the note into her response.


What She Does Not Have Access To

  • Your Upwork profile statistics or ratings

  • Previous conversations with the same client from before the CRM was connected

  • External URLs, files, or documents unless their content is pasted into a note

  • Real-time Upwork job market data


Tips for Better Outputs

  • The more context in the conversation, the better Laziza's responses. Don't rely solely on the job post — add notes from calls, emails, or other interactions.

  • If Laziza's draft misses something, don't just edit it — add a note explaining what was wrong, then ask her to regenerate. This gives her a feedback loop within the conversation.

  • For complex proposals, ask Laziza to summarize what she knows first before asking for a draft: type @, select Laziza, and write

what do you know about this project so far?

Then fill any gaps before drafting.


Related Articles

Did this answer your question?