What it is
Productboard is a product management platform that collects customer feedback, links it to features, and turns it into prioritized roadmaps and specs. It has held the “voice of the customer feeds the roadmap” position for a decade, next to Aha! on strategy-heavy roadmapping and Pendo on in-app analytics.
The 2026 product is a different shape from the one most buyers remember. Productboard now calls itself an “agentic product management system”, and the AI layer is Productboard Spark, publicly launched on 27 January 2026. Spark reads your feedback, documents (Confluence, Google Drive, Notion) and codebase. It produces cited findings and opportunities, drafts delivery-ready specs, and runs scheduled tasks and reusable skills. Pulse and the old Productboard AI add-on are being folded into Spark: Productboard’s billing FAQ says most of their capabilities move into Spark, and the separate line items disappear when an account moves to the current (“v13”) pricing.
Productboard also now ships a first-party MCP server at mcp.productboard.com, with OAuth and remote HTTP, enabled per workspace by an admin. Claude, Claude Code, ChatGPT, Codex and Cursor can read and update features, initiatives, releases, objectives, companies and feedback through it, acting with the connecting user’s permissions. One limit matters for EU buyers: the docs require the workspace to be hosted in the US region.
Why it shows up in Customer Success stacks
It is a product tool, but CS teams end up working in it because they own the loudest feedback channel in the company.
- Feedback capture without a CRM round-trip. A CSM pushes a Gong call snippet, a Zendesk ticket, a Slack thread or an email into Productboard. It lands as feedback, gets linked to a feature, and is searchable by company and segment. Contributors, the seat type most CSMs need, don’t count toward the maker total that drives price. Business caps them at 25, and Enterprise is unlimited.
- Churn reasons become prioritization input. Spark’s AI chat and feedback reports can answer “what are at-risk accounts asking for” with citations back to the source notes. That turns an escalation anecdote into something a PM will act on.
- Revenue weighting depends on the plan. Weighting requests by account ARR needs CRM data. On the current Business plan the native Salesforce integration is listed as not available, so you are left with the CRM CSV importer or the API. Native Salesforce sync is Enterprise only, and the pricing page does not list HubSpot at all.
- Agents can close the loop. With the MCP server, a CS ops agent in Claude can look up the feature a churned account asked for, read its status, and post a comment. CSMs no longer need to ask PMs for updates.
It is not a CS platform. It does not replace Gainsight, Vitally or ChurnZero. It sits next to them as the place product feedback goes to be triaged.
Pricing
The public pricing page now lists two plans, Business and Enterprise, with no dollar figure on either. Business is “Request demo” and Enterprise is “Contact sales”, and both carry a 5-maker minimum. The Free and Plus tiers that were listed until at least September 2026 are gone. So is the standalone $15 Spark plan from spring 2026.
- Business includes 500 Spark credits per maker per month, pooled across the workspace, plus 25 contributors and 2 Product Portals. It also caps the Jira and Azure DevOps integrations at 5 each. Its last published list price was $59/maker/month on annual billing ($75 monthly), per Productboard’s own billing FAQ dated September 2026. At the 5-maker minimum that is $3,540/year annual or $4,500/year monthly. Treat it as a reference point, not a quote.
- Enterprise includes 1,500 base credits a month plus 800 per additional maker. It adds unlimited contributors, SAML SSO, SCIM, audit logs, custom roles and the Salesforce integration. No public rate exists, and every third-party price you find is a guess.
- Credits and top-ups. Productboard’s own examples put analyzing 100 feedback items at roughly 30-50 credits and a full spec at 50-200. A 5-maker Business workspace (2,500 credits) therefore covers roughly 5,000-8,300 analyzed feedback items a month before specs. Top-ups cost $5 per 50 credits on monthly plans or $60/year per 600 credits on annual plans. Credits expire at each monthly renewal.
- Overage behavior differs by contract. On self-serve plans Spark simply stops when credits run out. On contract plans it keeps running and the overage is billed as extra bundles at month end, unless you ask support for a hard cap.
Legacy customers move to v13 pricing at renewal (September 2026 onward for contract accounts, October 2026 for self-serve). Legacy plans with unlimited contributor or viewer seats keep them.
Best for
A product ops or CS ops lead at a B2B SaaS company with at least 5 people who edit the roadmap. Feedback volume should run to hundreds of items a month across support, sales calls and CS, so that cited AI synthesis saves real triage time. The fit is best if you are already on Enterprise or will pay for it, because that is where Salesforce sync and ARR weighting live. A team with 1 or 2 PMs no longer has an entry plan here.
Versus the alternatives
- Canny — pick it when you want public feedback boards and a published price, and you have fewer than 5 makers. It is the cheaper, narrower tool. See Canny vs Productboard.
- Aha! — pick it when roadmapping is driven by strategy and portfolio planning more than by feedback volume.
- Jira Product Discovery — pick it when engineering already lives in Jira and the main job is idea triage next to delivery. It runs on the Atlassian contract and permissions you already have.
- Pendo — pick it when in-app usage analytics, not written feedback, is the main signal.
- Enterpret — an AI-native entrant on the feedback side. Pick it when you only need AI feedback analytics across channels and already have a roadmap tool.
Watch-outs
- You cannot budget from the website any more. With no published price and a 5-maker floor, the $59 figure is history, not an offer. Guard: ask for a written quote that states the per-maker price, the credit allocation and the overage terms, and compare it with the $3,540/year 5-maker reference before you sign.
- Credits meter the AI you are buying it for. Heavy Spark use (bulk analysis, specs, scheduled tasks) eats the pooled allowance, and on a contract plan the overage is billed automatically. Guard: run a pilot on your real feedback volume (the current page lists no self-serve trial, so ask sales for one), log credits used per week, and write a hard cap into the contract if you want spend to stop rather than grow.
- CRM weighting is gated. The CS case for Productboard rests on ARR-weighted feedback, and native Salesforce sync is Enterprise only, with no HubSpot listing at all. Guard: if you are buying Business, prototype the CSV import or an API sync of account ARR during the trial. If it is too brittle to maintain, price Enterprise or pick a cheaper tool.
- The MCP server is US-region only. EU-hosted workspaces cannot use it today. Guard: confirm your workspace region before you plan agent workflows around it.
- It rewards discipline you may not have. Spark’s findings are only as good as the feature hierarchy and feedback linking underneath them. Guard: name one owner (often CS ops or product ops) to maintain the taxonomy, and review untriaged feedback weekly. Otherwise it degrades into a write-only inbox within two quarters.