Cross-Company Queries in AI Analyst Chat

New in AI Analyst Chat and Deep Research Agents: cross-company queries. Ask about any company mid-conversation, even if your chat or agent is focused on a different one. No new thread, no context switching. Marvin spins up a specialist agent behind the scenes and brings the answer back inline.

How it works

Your main chat is an expert on the company you're analyzing. It has the full context: filings, earnings calls, press releases, management commentary. But it isn't an expert on every other company you might want to reference mid-conversation.

When you mention a different company, Marvin doesn't try to answer from general knowledge. It starts a sub-discussion with another agent that has the same depth of context. It's like asking a colleague who is an expert on that name. You don't need to pick up the full context yourself. You ask, and they respond with the depth you'd expect from someone who's read everything that company has published.

If the first answer isn't enough, the main agent follows up. It can ask the sub-agent to clarify, go deeper on a specific point, or approach from a different angle. This isn't a single lookup. It's a back-and-forth between agents, and Marvin decides how far to take it based on what your question needs.

You don't always need to name the company

You can trigger a cross-company query by mentioning a ticker or company name directly. But you can also ask questions like "What are the margins of this company's competitors?" without naming anyone specific.

That works because cross-company queries build on company relationships. Marvin already knows who the competitors, suppliers, and customers are. Ask about "competitors" and it figures out which names to query.

The two features feed each other: relationship mapping tells Marvin which companies matter, and cross-company queries let you pull detailed context from each of them without leaving your thread.

Parallel queries across multiple companies

Ask a question that spans three or four names. Marvin queries each specialist in parallel.

• "How do Microsoft, Google, and Amazon's capex guidance compare for the next fiscal year?"
• "Which of my semiconductor names mentioned pricing pressure in their latest call?"
• "Compare Nike and Adidas management commentary on inventory levels"

Each sub-agent pulls from that company's actual filings and earnings calls. Every answer is grounded in primary documents.

Cross-company context without losing your thread

Analysts covering 40-60 companies regularly need context from adjacent names. Maybe a supplier keeps coming up in an earnings call, or you need a quick peer comparison for a client note.

Before, the options were: open a new thread, lose your train of thought, and stitch the context together manually. Or settle for a shallow answer from an AI that hasn't actually read that company's filings.

Now you stay in your conversation and ask. You're midway through analyzing a company's margin trajectory, you want to check what a competitor said about input costs, and you keep going. Each sub-agent interaction shows up as a step in the chat so you can see which company was queried and what it found.