atlas-chat

Atlas Chat — Workforce Intelligence

Conversational workforce intelligence powered by live tools. Get salary benchmarks, talent supply/demand data, competitor intel, and hiring advice.

Workflow

Step 1: Determine Complexity

Choose the right tool based on the question:

Simple hiring questions → atlas_advisor_chat (2 credits, fast)

• "Should I include equity in this offer?"
• "What's a reasonable notice period for a VP?"
• "How should I structure this compensation package?"

Data-heavy workforce questions → atlas_chat (3 credits, live tools)

• "What's the talent supply for React developers in Berlin?"
• "Compare salaries for data scientists across NYC, London, and Singapore"
• "Which competitors are hiring aggressively for ML engineers?"

Step 2: Set Up Context (FREE, optional)

Call atlas_list_contexts to check for existing hiring contexts. If one exists, pass context_id to the chat for context-aware responses.

Step 3: Start Conversation

If the user provided a topic as $ARGUMENTS, use it as the first message. Otherwise, ask what they want to know.

For simple questions: Call atlas_advisor_chat with message. Fast response, no live tool execution.

For data questions: Call atlas_chat with message and optional context_id. May take 2-3 minutes as it executes live tools (salary APIs, talent supply data, competitor intel).

Optional: Include conversation_id from a previous response to continue a thread.

Step 4: Follow-Up Questions

For follow-ups, always pass conversation_id to maintain context. The AI retains the thread history.

The user can:

• Drill deeper: "Break that down by seniority level"
• Compare: "How does that compare to the European market?"
• Get actionable: "Based on this data, what should our offer look like?"

Step 5: When to Escalate

If the user needs more than chat can provide:

• Full research report: Direct to /atlas-report (15 credits) for a comprehensive PDF
• Candidate evaluation: Direct to /atlas-gem, /atlas-fit, or /atlas-shortlist
• JD analysis: Direct to /atlas-jd-review for job description quality check

Credit Cost

Action	Cost
Advisor chat (lightweight)	2 credits/message
Full workforce chat (live tools)	3 credits/message

Typical session: 3-5 messages = 6-15 credits

Tips

• atlas_advisor_chat is faster and cheaper — use for quick hiring questions that don't need data
• atlas_chat has live tool access (salary APIs, talent supply data) — use for data-heavy questions
• atlas_chat may take 2-3 minutes for complex queries due to live tool execution — set expectations
• Pass conversation_id for follow-up questions to maintain context
• For comprehensive research that you'll reference later, suggest /atlas-report instead (one-time 15 credits vs multiple chat messages)
• If you have a hiring context, pass context_id for more relevant responses