The Agent is your AI partner inside Lane. Ask it questions about your customers, features, and roadmap. Let it help you think through what to build next and shape that thinking into a Plan. It knows what’s in your product because it can read your Features, Insights, Customers, and Objectives directly.Documentation Index
Fetch the complete documentation index at: https://docs.laneapp.co/llms.txt
Use this file to discover all available pages before exploring further.
What you can do with the Agent
Find what matters
Ask broad questions across your workspace and let the Agent surface what’s relevant.- “What are the top 5 Signals from enterprise customers this quarter?”
- “Which features have the highest revenue impact?”
Think through a problem
Start here when you don’t know the answer yet — the Agent helps you explore before you commit to a direction.- “What are the recurring pain points around onboarding?”
- “How are customers reacting to the new pricing?”
Shape a Plan
Once you’ve explored, convert the conversation into a Plan, or open a Plan and ask the Agent to expand each section. Working in two steps — explore, then draft — produces sharper documents than asking for a full PRD in one shot.Take action
- “Create a feature for the CSV export request from Acme.”
- “Link this feedback to the Onboarding objective.”
Edit your writing
Highlight any text in a Plan and ask the Agent to polish, summarize, or expand. Use slash commands in the editor to ask follow-up questions while you write.Where you’ll find the Agent
Agent (top menu) — your open-ended workspace for asking broad questions, exploring Signals, and starting conversations that turn into Plans. Inside a Plan — a chat scoped to the document you’re writing. Best for drafting, refining sections, and pulling in linked data. In the editor — highlight text or type/ to ask the Agent anything about the surrounding content.
Getting better results
A few things that consistently lead to better answers:- Be specific. “Top feedback Signals from enterprise customers in the last quarter” gets you somewhere; “what’s important?” doesn’t. The more context you give — segment, timeframe, product area — the better the answer.
- Keep one ask per turn. The Agent is sharpest when each turn has a clear goal. Bundling several actions into a single prompt (“find X, summarize Y, then create Z”) tends to dilute the result. Break it into steps.
- Explore before you draft. For anything that becomes a Plan or a feature brief, start by asking questions. Use the Agent’s answers to refine direction, then ask it to draft. One-shot “write the full PRD” prompts produce thin documents because there was no thinking step.
- Tell it what’s wrong. If the answer isn’t quite right, reply in the same chat with what to fix — narrower scope, different angle, more detail on a specific point. Don’t start over.
When the Agent suggests changes
The Agent never silently changes your data. If it wants to update a Feature, edit a Plan section, or link records together, you’ll see a proposal card you can accept or reject. You’re always in control.Bring in outside context
The Agent can pull answers from sources outside Lane, so the same question — “what does X actually do?” or “what did the customer say?” — can be answered without leaving the chat. Connect a source once in Settings → Integrations and it’s available to the Agent.GitHub
Connect a repository to ask the Agent questions about your code. This is most useful for non-engineering team members who need answers that live in the codebase — “Is this feature already built?”, “How does our pricing logic actually work today?”, “What does the enterprise tier include in the code?”. Instead of pinging engineering every time, ask the Agent. It’s equally useful for engineering-facing Plans, where the brief needs to be grounded in what’s already in the repo rather than what someone remembers being there. The Agent reads code; it never writes back.Google Drive
Share up to 3 files with the Agent and ask questions across them — sales transcripts, customer research, strategy docs, whatever isn’t yet in Lane. Common uses:- “Summarize the three calls from last week and pull out feature requests.”
- “What are the recurring objections in these transcripts?”
- “Compare the two research reports — where do they disagree?”
Good to know
- The Agent only sees what’s in your workspace — never other customers or anything outside Lane.
- It may ask you a clarifying question if a request could go several ways — that’s intentional.
- It won’t delete records or unlink things.
Credits
Agent activity uses AI credits. See Billing for your plan’s allowance and how to top up.AI privacy and data usage
- Lane does not use your data to train its own AI models.
- Data processed for AI features is shared only with trusted partners to deliver those features.
- Your data is not used to train external models without permission.
FAQs
Q: Can I undo something the Agent did?A: Most changes come through as proposals you accept or reject. For one-shot creations (like a new Feature), edit or delete as usual. Q: Why did it ask me a clarifying question?
A: When a request is broad enough that the Agent could go in several directions, it asks first to avoid wasted work. Q: Why didn’t it find something I know is there?
A: Try rephrasing — the Agent searches by both keywords and meaning, and a small change in how you describe the thing can surface it. Q: Does the Agent work on data I haven’t imported yet?
A: No — it only sees what’s in Lane, with the exception of connected sources like GitHub and Google Drive.