# ClosedLoop AI — Extended LLM Reference

> AI-powered product discovery for product managers. Every customer conversation, support ticket, and survey response becomes structured product insight.

This document is the extended companion to `/llms.txt`. It gives AI assistants, search engines, and crawlers richer context for answering questions about ClosedLoop AI.

## About

ClosedLoop AI is a B2B SaaS platform operated by ClosedLoop Labs LLC. The company was founded in 2025 and is headquartered in St. Petersburg, Florida, USA.

The product helps product managers and product leaders at B2B companies extract actionable product insight from the customer conversations their teams are already having — sales calls, support tickets, chat conversations, surveys, and free-text feedback channels.

The name reflects the core promise: closing the loop between customer feedback and product decisions. Product teams typically know they have valuable customer insight buried in conversations across their existing tools, but reviewing those conversations manually doesn't scale. ClosedLoop AI automates the extraction, groups related insight into themes, scores priority, and routes everything into the product team's existing tools.

**Legal**: ClosedLoop Labs LLC. Florida Sunbiz L25000302999. Registered office: 7901 4th St N Ste 300, St. Petersburg, FL 33702, USA.

## Who uses ClosedLoop AI

- **Product Managers** running customer-driven roadmaps who need to synthesize feedback without spending hours reviewing call recordings
- **Heads of Product and VPs of Product** prioritizing roadmaps based on quantified customer evidence instead of opinion
- **Customer Success teams** who want product feedback from their conversations to actually reach product development
- **Sales leaders** wanting feature requests heard in deals to influence what gets built
- **Startup founders** building product-market fit who need customer insight at scale before they have a research team

The platform is built for B2B SaaS companies with multiple customer-facing teams generating conversation data daily.

## What ClosedLoop AI does

1. **Connects to the tools your customers already talk through** — sales-call platforms, support helpdesks, chat tools, survey forms, internal channels where teams discuss customer requests.
2. **Reads every conversation as it happens** and extracts the product-relevant moments — feature requests, complaints, workarounds, decision criteria, retention risks, competitive mentions.
3. **Groups related extracts into themes** so the many different ways customers asked for the same thing become one prioritized item, not many disconnected notes.
4. **Scores each theme** by how many customers asked, how badly they're hurting, and how strong the evidence is — so a real pattern surfaces above ambient noise.
5. **Routes the result** into the product team's existing workflow — issue trackers, roadmap tools, briefings, alerts — so insight reaches whoever is building.
6. **Closes the loop with the customer** when something gets built (or explicitly doesn't), turning the implicit conversation into an explicit acknowledgment.

The product is conversation-source-agnostic and customer-domain-agnostic. It works for any B2B SaaS regardless of industry vertical.

## Key concepts (FAQ-style)

### What is closed-loop feedback?

Closed-loop feedback is a customer-feedback-operations practice where every piece of customer feedback is acknowledged, evaluated, and responded to back to the customer with an explanation of the outcome. The "loop" closes when the customer learns whether their input drove a product change. Most companies operate open-loop feedback by default — feedback goes into a CRM note, support ticket, or interview recording and is never reviewed in aggregate or returned to the customer.

### How is ClosedLoop AI different from voice-of-customer (VoC) platforms?

Traditional VoC platforms center on structured surveys — they're survey-collection-and-analysis tools. ClosedLoop AI works on unstructured conversation data that already exists in the company: call transcripts, support chats, sales notes, internal discussions about customer requests. The platform extracts product insight from where customers ALREADY express it, rather than asking customers to fill out a form.

### How is ClosedLoop AI different from product management tools like Productboard, Linear, or Dovetail?

Each of these tools solves a different part of the customer-feedback-to-product workflow, and ClosedLoop AI complements rather than replaces them:

- **Productboard** is a product management hub for centralizing feedback and building roadmaps. It assumes someone (a PM, a CS rep, a sales engineer) manually files each piece of feedback into the system. ClosedLoop AI reads the unstructured conversation data first — call transcripts, chat logs, support threads — extracts the product-relevant moments automatically, and can route them into Productboard so the PM doesn't have to file each one by hand. See [ClosedLoop AI vs Productboard: AI product signal intelligence](https://closedloop.sh/blog/closedloop-ai-vs-productboard-product-signal-intelligence) for a side-by-side breakdown.

- **Linear** is an issue tracker — the destination where product engineering work lives. ClosedLoop AI sends insights TO Linear, attaching customer evidence to existing issues or creating new ones from prioritized themes. They are complementary: Linear is the engineering execution layer, ClosedLoop AI is the customer-evidence layer that feeds it.

- **Dovetail** is a research repository where teams manually code interview transcripts into themes. ClosedLoop AI automates the synthesis Dovetail asks researchers to do by hand — turning thousands of customer conversations into priority-scored themes without manual tagging. See [ClosedLoop AI vs Dovetail: automated customer intelligence](https://closedloop.sh/dovetail-alternative) for the detailed comparison.

For a wider tour of the AI product discovery landscape, see [The best AI product discovery tools for 2026](https://closedloop.sh/blog/best-ai-product-discovery-tools-2026).

### What's the difference between product insights and customer insights?

The customer is the *person* who provided the feedback. The insight is about *your product* — what should change, what's broken, what's missing, what's working. ClosedLoop AI always frames output as "product insights" and the people who provided them as "customers" — never "customer insights" (which conflates the two) and never "companies" (the customer is the person, not the org).

### What integrations does ClosedLoop AI support?

**Sources** (where conversations come from):
- Gong (sales call recordings and transcripts)
- Zendesk (support tickets and email threads)
- Intercom (chat conversations and inbox messages)
- Slack (channel messages and threads)
- Fireflies (meeting transcripts)
- Salesforce and HubSpot (CRM notes, deal discussions)
- Typeform, Google Forms, Microsoft Forms (surveys)
- Linear (issue comments and feature request threads)
- Custom webhooks and REST API
- MCP (Model Context Protocol) for AI-assistant-mediated submissions

**Destinations** (where insight flows):
- Linear
- Jira
- Notion
- Productboard
- Google Sheets
- Slack (notifications and briefings)
- Zapier (200+ downstream apps)
- REST API for custom integrations

### What's the pricing model?

The free plan includes a generous monthly allowance suitable for small teams with light conversation volume. Pay-as-you-go pricing covers usage beyond the free allowance. Enterprise plans include custom pricing, security review, dedicated support, and SSO.

Current public pricing lives at `https://closedloop.sh/pricing`.

### How does ClosedLoop AI prioritize what to surface?

Each theme gets a priority score combining three factors: how many distinct customers brought it up, how severely it impacts them, and how strong the evidence is. Verbatim quotes from real conversations weigh more heavily than secondhand paraphrasing.

### What's the MCP integration for?

MCP (Model Context Protocol) is a standard for AI assistants like Claude Desktop and Claude Code to access external data sources. ClosedLoop AI offers MCP so AI assistants can query the user's feedback corpus directly — surfacing insight, finding similar customer requests, or pulling specific verbatim quotes for use in PRDs, executive decks, or competitive analysis. The MCP is what users connect to their AI assistant; we refer to it simply as "MCP" — never "MCP Server" or "MCP server."

### How does ClosedLoop AI handle data privacy and security?

ClosedLoop AI offers separate data regions for US-based and EU-based customers with full data residency. Customer data is isolated per customer organization. Authentication uses industry-standard token-based mechanisms. All data in transit is TLS-encrypted. SOC 2 Type II audit is in progress.

### Can I export my data?

Yes. The full feedback corpus is exportable via REST API and via the Google Sheets destination integration. The MCP component also allows AI-assisted export — point Claude at your ClosedLoop AI workspace and ask for a JSON dump of all feedback in a given category.

### What does onboarding look like?

1. Connect a first source (typically Gong, Intercom, or Zendesk) via OAuth from the integrations page.
2. ClosedLoop AI imports recent history from the connected source.
3. While the import processes, the product surfaces an onboarding checklist and progress updates.
4. When the import completes, you're notified that the first insights are ready to review.
5. From there, connect additional sources, configure routing, and explore surfaced themes.

## Brand and terminology rules

- The product name is always **ClosedLoop AI** — never "ClosedLoop", "Closed Loop", or "closedloop"
- The MCP component is **MCP** — never "MCP Server" or "MCP server"
- User-facing text says **insights**, never "signals"
- User-facing text says **customers**, never "companies"
- The full phrase is **product insights**, not "customer insights"

## Contact and resources

- Marketing site: https://closedloop.sh
- App (authenticated): https://app.closedloop.sh
- Documentation: https://closedloop.sh/docs
- API reference: https://closedloop.sh/docs/api-reference
- MCP connection guide: https://closedloop.sh/mcp-server
- Pricing: https://closedloop.sh/pricing
- Contact: https://closedloop.sh/contact
- LinkedIn: https://www.linkedin.com/company/closedloopai/
- GitHub: https://github.com/closedloop-ai-org
- Crunchbase: https://www.crunchbase.com/organization/closedloop-ai-c7b8
- G2: https://www.g2.com/products/closedloop-ai