We think the future of customer support is autonomous. AI can now reliably answer questions that have been answered before and instantly verify bug reports, without sacrificing the user experience. This means that humans can finally move on to help customers who really need them.

To support you through this transition, we have shipped hundreds of improvements to the Productlane Support Agent over the last few months, and today we’re launching the largest one yet.

Deploy the agent to all your channels

You can now preview responses and deploy the agent per channel, and it will start picking up conversations like a new team member. Across our customers, we currently have a 65.3% resolution rate, and it’s increasing every week. The best part is that it’s completely free up to 25 resolutions a month on Starter, 50 on Pro, and 100 on Scale. Everything after that is $0.79 per resolution. Pick your channels here and let the agent start saving you time.

Let the agent close conversations

When the agent is deployed on your support channels and confident it solved the problem for the customer, it closes the thread itself. Your daily triage is then only the customers that still need you. The notification in your Slack channel now contains the agents reply so you will still see what‘s happening.

Intelligent handoff

The agent works from two entry points. The first is your support channels from above, the second is the dedicated AI Chat on your Help Center and Widget. Threads from there are marked done by default, and now we added the possibility for the agent to escalate those to a human if they weren’t resolved.

Image annotations

The agent can now mark up the screenshots in your help center, so an answer shows the customer exactly which button to click. Getting those screenshots there is a job for your coding agent. Point it at the Productlane MCP and build a skill that captures your help center screenshots with good demo data. Make it a routine and they stay current. Ideally make the articles Agent only so it doesn’t flood your customer facing docs.

User-facing timeline events

Linear issue status now shows up in the live chat and AI chat, so a customer can watch their requests move. They see todo, in progress, and done. The status never moves backwards, and cancelled work stays hidden. A merged PR shows up too, and none of it sends a notification. Set up a workflow in Linear where your coding agent picks up small bugs, and your customers can watch it live.

How to get started

We’re really excited about helping to make your support autonomous and saving you hours of work, and we’ll continue improving the resolution rate of the agent every week. To get started, you will just need to:

1. Import your help center
   Move your help center to Productlane to make it self-updating from GitHub, Linear, and your conversations.

2. Deploy the agent
   Select the channels you want to deploy the agent to here.

After that, you can improve the agent continually by adding custom connectors or a complete map of your product with screenshots as an internal doc for your agent.

Productlane now works with Microsoft Teams. Teams conversations can flow into Productlane alongside your other support channels, making it possible to manage customer communication in one place. Attachment support is included as well, so files can be sent and received as part of those conversations. This is available on the Scale plan.

Overview

The Microsoft Teams integration brings Teams conversations into Productlane. Once the Productlane bot is added to a team, channel messages and chats are automatically converted into Productlane threads. You can reply from Productlane, and your reply appears back in Teams as a threaded message, with your name and avatar. Files and images sync both ways.

Like the Slack integration, you can link Teams threads to Linear issues, snooze and close them from your Productlane inbox, and collaborate internally with side notes - without ever leaving Productlane.

Your support agent can now access data from any source. Customers can ask:

• Your product database

• Stripe, Shopify, or any billing system

• Internal analytics and usage data

• CRM records, logs, anything

Getting closer to helping you make more of your customer support autonomous.

Let the Productlane Agent access any data

When a customer asks about their recent invoice, the agent looks up their subscription and answers right there, so your team gets the number without switching tabs. It can run a refund or update a CRM record the same way, with a confirmation step before anything actually changes.

Access depends on which agent is asking. Copilot and thread drafts keep full access, since a person reviews the output before anything goes out. Customer-facing agents like email auto-reply and live chat reach the connectors and actions you've turned on for them, and you enable each connector for agent access individually. Actions that make real changes can be marked to always wait for a human to confirm, no matter which agent triggered them.

External links and internal docs for the agent

When a question needs more context, the agent can now also pull from your help center articles or a custom URL you've pointed it to, so it answers from your docs even for topics no connector covers.

Try it: head to Settings → Connectors to add actions to a connector, or turn on agent access for one you already have.

You can now handle multichannel customer support from inside Linear. Productlane creates a synced comment on an issue for every conversation from Live chat, Slack Connect, email, or Teams. You can reply directly from Linear, and responses sync back to the customer in their original channels.

Your support team can work in Productlane with AI drafts and resolutions for high volume, while engineers can stay in Linear and still reply to customers. Also, with coding agents, it's really powerful to have the customer conversation directly on the ticket where a PR from your agent is already waiting.

Turn it on in Settings → Linear under Reply to customers in Linear. The feature is available on all plans. Everyone who wants to reply from Linear needs a seat in Productlane.

AI Assistant for Help Center and Changelog

We have more great news: The Assistant is now included in all plans and works inside the Help Center and Changelog editor. Open any article and toggle it from the header icon or Cmd + J to draft content. It now also has access to your connected GitHub repository, so it has context for your entire codebase, helping you write Help Center articles and Changelogs faster.

Every plan includes a monthly pool of AI credits covering Copilot, docs generation, changelog generation, and translations. Usage past that pool bills automatically at token cost plus a small margin, capped at the overage limit you set. Track spend and adjust the limit in Settings → Features → AI → Usage.

Inbox assistant and UI improvements

The assistant now works across your whole inbox, not just a single thread. You can ask it things like who to answer first or who's at risk of churning. The messages have a new UI for better readability, and the snoozed and done tabs moved out of the sidebar into a shared Open / Snoozed / Done dropdown in the inbox header.

Advanced Reporting early access

Reporting now has dedicated tabs for Overview, Conversations, Team, Responsiveness, and Feedback, covering first response and resolution times, queue depth, AI resolution rate, and reply activity by teammate. Try it alongside other early access features that we're currently working on: Attio, PostgreSQL integration, and Multiple Portals.

Productlane now updates your Help Center for you. The AI agent learns from your past support conversations, code changes in GitHub, and completed Linear issues and writes the articles your docs are missing, ready for you to review. Connect your repository in Settings > Integrations > GitHub to start.

Sourced from GitHub, Linear, and support threads

Three signals feed the agent. A merged commit drafts a new article or updates the one it affects. A closed Linear issue arrives with a draft ready to review. A resolved support thread becomes a new entry, or a suggested edit when the conversation exposes a gap in what you already published.

Review and approve in a lightning fast UI

Proposals collect in a Suggestions panel at the top of your docs sidebar. Each row shows the line diff at a glance: +173 is lines added, -39 is lines removed.

Open one and the full review view takes over. The article sits on the left with every proposed change highlighted in place. A rail on the right lists each edit on its own, and hovering a card lights up exactly where it lands in the text.

Accept, and the article publishes to your portal. Each group in your navigation carries a counter of the drafts still waiting inside it.

Generate your Help Center with Claude Code or Codex

Connect Productlane over MCP and point Claude Code or Codex at your repo: "Review the codebase and every past conversation in Productlane, then write our Help Center." The agent drafts the full set of articles, and you review them through the same Suggestions panel.

Bring an existing Help Center with you: import from Markdown or CSV under Settings > Import/Export, and the agent maintains it from there.

With the new MCP server, you can connect Productlane to Claude, Codex, Cursor, or any MCP-compatible client. Set it up here to enable capabilities like:

• Draft the reply. "Go through all new threads today, check the help center and codebase, then draft a reply for me to approve."

• Bug arrives, fix starts. "When a bug report lands in the inbox, reproduce it and draft the fix as a PR."

• Update the help center. "Review the codebase and all past conversations in Productlane to find gaps in our docs."

• Triage the inbox. "Sort today's new threads by area, tag them, set importance, and assign each one."

• Weigh the roadmap. "Rank open feature requests by the revenue of the companies asking."

• Translate the docs. "Translate every help article into German and French."

• Clean up the CRM. "Find duplicate companies by domain and merge them."

Two ways to connect

Add the Productlane server URL in your client and sign in with OAuth. You pick a workspace, choose what the agent can access, and approve, with no key to copy or paste. The client handles the rest.

You can also connect with an existing v2 API key. Create one under Settings > Integrations > API and pass it as a bearer token. OAuth suits a quick, interactive connect, while API keys fit headless and programmatic setups.

Scopes decide what the agent can do

Whichever method you use, scopes control which tools the agent sees. With OAuth, the consent screen lets you grant full access or hand-pick permissions per resource. With an API key, the key's own scopes do the same. A read-only grant exposes only read tools; anything without admin access hides member management. Revoke access, and every connected client stops instantly.

Coverage across the whole workspace

The agent works with threads, contacts, companies, projects, issues, tags, changelogs, help center articles, the customer portal, and members. Every action runs through the same v2 API as your other integrations, so rate limits, error handling, and audit logging apply exactly as they do over HTTP. Add Productlane to any compatible client to get started. The full setup steps live in our docs.

Improvements

• Redesigned the email settings page and improved the snippets settings experience

• Added the ability to review verified DNS settings from the dashboard

• Added a logout action in CMDK

• Made ticket tag removal clearer

• Refined attachment interactions

• Added a stop button while AI is generating

• Added Reply All as a default reply option

• Improved docs subgroups and nested folder organization

• Added better navigation hierarchy across key areas

• Improved importing content via Markdown

Fixes

• Fixed public roadmap submissions after login

• Resolved duplicate ticket creation and auto issue generation issues

• Fixed cases where the sidebar AI agent failed to generate text

• Addressed email thread performance issues

• Improved link styling for German doc articles

• Fixed screenshots and images not rendering correctly in conversations

• Fixed attachment display issues across submitted requests and threads

• Fixed duplicate threads and wrong customer assignment in the Linear sync

• Fixed Linear customer requests not populating correctly

• Fixed thread sorting not applying properly in the All view

• Fixed dark mode getting stuck

We're excited to introduce API v2 and webhooks today.

API v2 is a complete rebuild of our public API. Featuring scoped keys, cursor pagination, a consistent error envelope, and an OpenAPI-driven reference. Use it to sync threads, contacts, companies, changelogs, help center articles, projects, members, and tags between Productlane and any system you run, with support for the latest features and more.

Every key is now scoped (threads:read, companies:write, admin, etc.), so a CI script that only lists threads can't accidentally invite a member. Keys are stored as hashes, shown to you once at creation, and listed with a last-used timestamp.

Errors land in a single envelope with a stable code, a human message, and a request_id you can quote in support tickets. Pagination is cursor-based across every collection, and dates are plain ISO 8601 strings.

V1 keeps working until November 20, 2026. Every v1 response now carries Deprecation and Sunset headers, so your tooling picks up the timeline automatically. Migrate by minting a new key under Settings > Integrations > API.

Webhooks

Webhooks are now available alongside the API. Subscribe to events like thread.created, message.received, contact.updated, changelog.published, and dozens more, and receive a signed payload at your endpoint within seconds of the event firing.

Every delivery is signed with HMAC-SHA256 using your webhook's secret, so you can verify the payload really came from Productlane. Failed deliveries retry on a backoff schedule, and a delivery log in the dashboard shows the last 200 attempts per webhook with response codes and latencies.

We email you when a webhook starts failing repeatedly, and again if it's auto-disabled. New webhooks added to your workspace trigger a notification too, so unauthorized changes don't slip past.

Set up your first webhook in Settings > Integrations > API > Webhooks.

Usage analytics

The API settings page now includes a built-in analytics block showing how your keys and webhooks are performing. Track total requests, error rate, error distribution by endpoint, webhook delivery outcomes, and response time, all in one place, with windows from the last 24 hours up to 14 days.

It's the same data your customers' integrations are generating, so debugging "why did that fail?" no longer means waiting for a support ticket. Open Settings > Integrations > API to see it.

Improvements

• Improved widget design

• Live chat email fallback delay reduced to 2 minutes.

• Custom views refinements; fixed active state when switching with number keys.

• Docs/changelog editor gains bullet lists.

• Replaced Lucide with Nucleo icons in the whole app.

• Customizable user avatars.

• Show changed lines count instead of new symbol count.

• Subpage counters in navigation.

• AI sidebar drafting and writing‑style controls.

• Set the last author as the contact person on threads.

• Subjects no longer include email addresses in the list UI.

• Responses can be moved into docs more easily.

Fixes

• Dark-mode customer messages are now readable.

• Mobile chat and AI loading animation glitches resolved.

• Inbox navigation and thread selection stability; shortcut reliability improved.

• Fixed replies looping back to the support inbox.

• Changelog menu link no longer returns a "Page not found" error.

• Custom domain TXT record handling fixed.

• Various UI polish: spacing, rounded corners, dropdown and redirect issues, and cut‑off content.

We’re excited to introduce Support Copilot today.

Move through conversations faster and draft replies directly in a thread. Support Copilot gives you a starting point for common responses while keeping the final send in your own hands.

It suggests replies based on your help center, past conversations, Changelogs, and Linear issues. It’s available on the Scale plan, and we launch it alongside two more essential support features today: SLAs and Custom Views.

Custom views

Custom Views are now available to help you organize the inbox around the work that matters most. You can now save filtered views in the sidebar for quicker access. This release also adds filters like revenue and company size, making it easier to segment queues, prioritize important customers, and build workflows around different types of conversations.

SLAs

SLAs are now available in Productlane. Teams can track time to first response, time to resolution, and service trends across teams and timeframes.

This makes it easier to measure support performance, spot bottlenecks, and stay aligned on response expectations. SLAs are available on the Scale plan and can be set up in Settings > SLAs.