Ask what your own data says about your pricing
A new pricing analysis reads your workspace's own records and says, in plain sentences with the numbers behind them, what is ready, what is stuck and what to do next. Free, for people and agents alike.
Your workspace now has a pricing analyst that runs on your own records, with no AI model and at no cost. Open Settings, Learned pricing, or ask it from the API or from your agent, and it answers with findings, most urgent first:
- A learned estimate is ready but learned pricing is off. It names the estimate, its 95% interval and how many outcomes it came from.
- Learning cannot be used yet for a product, and exactly what it needs. For example "12 more quotes with a reported outcome", "3 more rejected quotes reported", or prices that vary by more than 1%.
- Too few quotes have a reported outcome, with the share that do.
- Your guardrails held back most of a product's prices, with how often and which rule, and whether those were your own rules or the defaults.
- Approvals waiting, or about to expire within 48 hours.
Every finding states its numbers and links to where you act on it. Nothing is changed for you: you, or your agent, decide.
Agents get the same answer from the get_pricing_analysis tool on the MCP server, since an agent often runs the workspace. From the API, call GET /api/pricing-analyst/findings with a key that can read routing settings.
Written briefs and answers to free-form questions ("why did this price change?") come next.