Create a Product Area
Availability
Wave is a paid offering for organizations on Growth and Enterprise plans. To get access, talk to a Wave expert.
A Product Area gives Amplitude's agents context on where to focus based on your product surfaces, workflows, and goals. It also lets you customize how agents identify, prioritize, and spec new Opportunities for that area.
A Product Area defines both where the system looks and how it decides what matters. Product Areas don't have to be small. They can cover a large surface or workflow, but they should have clear scope, goals, and boundaries. The more specific those boundaries are, the sharper the resulting Opportunities.
Start with clear boundaries
Amplitude recommends starting with two or three well-scoped Product Areas. A Product Area can be large, but it should have explicit boundaries. "Checkout" is easier to tune than "The whole web app" because the included and excluded work is clearer.
What a Product Area contains
Each Product Area captures the context the discovery agents use to find and prioritize opportunities.
| Setting | What it does |
|---|---|
| Overview | The name, owner, and a plain-language description of what the area covers and what success looks like, for example "Checkout covers the path from starting payment to a paid order. Success is a completed purchase without needing support." State the boundaries here so discovery stays on-topic and excludes near-neighbor areas. |
| Key Metrics | The metrics that define success. Each metric has a role: primary (the main success metric), secondary (diagnostic), or guardrail (must not regress). |
| Opportunity Types | The ranked types of work Wave should focus on: Strategic Bet, Quality Improvement, Instrumentation, Bug Fix, and New Feature. Turn off any type Wave should skip. |
| Sources | Which data sources discovery should analyze for this area, including competitive analysis and MCP external sources. For the full list, go to Connect context. |
| Investigation Instructions | Three durable instruction lanes: Investigation (where agents should look), Plan (solution approaches, constraints, and specification requirements), and Coding agent (how coding agents should implement). |
| Investigation Schedule | How often Wave investigates the area, for example weekly. |
Repositories are set for the whole workspace, not per Product Area. Connect GitHub in Wave settings > Codebase, then choose the repositories Wave indexes in Wave settings > Repository Index.
To edit these settings later, open the Product Area and click Configure.
Instructions are durable, not one-off
Use Product Area plan, investigation, and coding agent instructions for lasting preferences, such as "Require evidence from at least two sources" or "Prefer reversible changes for mature flows." For temporary investigations, run the Opportunity Manager against a specific scope, page, or signal category. Those one-off runs can explore a focused question without changing the durable Product Area settings. For how to write each instruction lane, go to Tune Wave discovery.
Create a Product Area
In Wave, open Product Areas and click Create product area. Wave opens the setup agent in Wave Chat.
The setup agent is a conversational onboarding partner that scans your project, lists existing Product Areas, reviews product context and event names, then asks which product area you want to focus on. From there it investigates the area, watches session replays when they're available, proposes target metrics and goals grounded in your real data, and presents a complete draft for you to confirm or edit.
The setup agent pre-populates every field from your data, so you review and adjust rather than fill in a blank form. You can always edit and tune the Product Area settings afterward.
Setting up a Product Area with the setup agent
AI
I looked through your project: 3 dashboards and 47 events in the
purchase.*namespace relate to checkout. Which product area or workflow do you want to focus on first?You
Let's focus on checkout. We're seeing drop-off before payment.
- Agent action
Runs a focused scan of checkout dashboards, events, and metrics, then reviews 3 representative session replays.
AI
Here's a draft Product Area for Checkout:
- Scope:
checkout_started→payment_submitted→checkout_completed - Primary metric: checkout completion rate
- Guardrail: checkout latency
- Goals: Conversion, Revenue
Want to adjust the metrics or scope before I create it?
- Scope:
You
That's right. Create it.
- Agent action
Presents an editable settings draft for review.
AI
Your Checkout draft is ready. Review its settings, then click Activate to create the Product Area and start its first discovery run.
After setup
Click Activate after reviewing the draft to create the Product Area and trigger its initial discovery cycle. Investigation time depends on scope and connected sources. After that, Wave investigates the Product Area on the cadence set in Investigation Schedule. To run discovery between scheduled runs, open the Product Area and choose Create > Generate opportunities. To update scope, metrics, goals, or sources as priorities change, open the Product Area and click Configure. To change the schedule, owner, or language, or to archive the area, go to Manage Product Areas.
Was this helpful?
For AI agents: a documentation index is available at /docs/llms.txt. Append .md to any page URL for markdown, or send Accept: text/markdown.