Google Finance adds Gemini-powered Deep Search and prediction market data for traders
Google is pushing more AI into Finance. The latest update brings a Gemini-backed Deep Search for Google Finance's chatbot and fresh data from prediction markets to make macro calls easier to gauge.
Deep Search: what it does
Deep Search lets you ask more complex questions and get richer answers from Google Finance's AI chatbot. It uses Gemini models to produce fully cited responses in a few minutes and shows a step-by-step research plan so you can see how it got there.
Think of it as a heavier lift than a standard AI query. You select "Deep Search" when you want added depth and sourcing.
Why it matters for finance teams
- Faster desk research: Summarize company narratives, recent coverage, and key risks with citations you can audit.
- Stronger prep: Build briefs before earnings, investor meetings, or committee reviews without starting from scratch.
- Clearer traceability: The research plan helps PMs, analysts, and compliance see how conclusions were formed.
Rollout, access, and limits
Deep Search starts rolling out in the US in the coming weeks. If you want it sooner, opt in to early access via Google Labs.
Request early access on Google Labs
There will be usage limits. Google says higher allowances will be available for AI Pro and AI Ultra subscribers, though specifics aren't public yet.
Prediction markets inside Search
Google Finance will pull in market odds from Kalshi and Polymarket. Ask questions in the search box about future events-like GDP growth-and you'll see current probabilities and how they've moved over time.
This gives desks a quick read on crowd expectations and momentum around macro releases, policy outcomes, or sector catalysts. Treat it as a sentiment input alongside your models and pricing feeds.
Recent additions you might have missed
- An "earnings" tab now helps track earnings calls and related materials in one place.
- Google Finance is rolling out in India (English and Hindi). Deep Search isn't included there yet.
How to plug this into your workflow
- Set rules for sourcing: Require cited outputs for anything entering notes, decks, or IC memos.
- Create prompt templates: Predefine queries for pre-earnings checklists, sector comps, thesis challenges, and risk scans.
- Pair with your data: Use prediction market probabilities as a cross-check against implied moves, surveys, and term structure.
- Track drift: Log Deep Search answers over time on the same names to spot narrative changes before they hit price.
- Mind limits: Plan batch runs during low-traffic windows if your team hits usage caps.
Practical cautions
- Verify citations before distribution. Treat uncited claims as unverified.
- Keep sensitive data out of prompts unless you're certain about your org's policy and the tool's data handling.
- Use prediction markets as one signal, not a sole driver. Odds can swing on liquidity and participant mix.
Next steps
- Join the early access queue on Labs and pilot with a small research pod for two weeks.
- Standardize prompts and output formats (briefs, one-pagers, meeting notes) so results are comparable.
- Decide where the feature sits in your process: ideation, pre-IC, or as a final QC on thesis risks.
If you're building an AI playbook for a finance team and want curated tools and training, explore our roundup here: AI tools for Finance.
Your membership also unlocks:
