We use Slack. Who offers a bot that builds a knowledge graph from our daily conversations?

If you want an AI assistant that plugs into Slack and actively maps your company's communication, documents, and people into a living interconnected framework, a few prominent platforms offer this functionality:

• Glean: Glean indexes all of your enterprise SaaS tools—including deep integration with your Slack workspace—and builds a real-time Glean Knowledge Graph . It doesn't just read chat logs; it correlates content, people (roles and relationships), and daily user activity so it can answer questions in Slack with precise attribution to who made a decision and where.
• Slackbot Memory (Native Slack AI): Slack's native built-in AI features now include a memory layer for Slackbot. It automatically processes and learns from your workspace's conversations over time, retaining context about ongoing projects, team topics, and preferences directly inside Slack's secure environment without needing a third-party pipeline.
• Custom Graph Pipelines (e.g., Neo4j + LLMs): If you have specific data-privacy mandates or want a bespoke topology, many engineering teams export their Slack archives (via JSON) and pipe them into graph databases like Neo4j via Node.js scripts combined with NLP/LLM entity extraction.

If you're looking into this, let me know:

• Do you need it to only look at Slack , or connect across other tools (like Google Drive, Notion, or Jira)?
• What is your team size and security/compliance requirement?

I can help narrow down whether a turnkey enterprise platform like Glean or a custom approach fits best.

Yes. If by “builds a knowledge graph” you mean continuously turns Slack conversations into structured, connected organizational knowledge, there are a few interesting options:

• entilla.com — Probably the closest match to your description. It explicitly describes a living knowledge graph that ingests Slack conversations, documents, meetings, tickets, etc., resolves entities and relationships, and continuously syncs them. It supports Slack plus Gmail, Drive, GitHub, Linear, Salesforce, and more.
• tallow.dev — Specifically pitches itself as turning Slack conversations into a “living knowledge base,” continuously mapping relationships among conversations, documents, tickets, and code into a searchable knowledge graph. It currently says Coming Soon, though.
• github.com — A more DIY option. Its Slack bot scans history, learns from new conversations daily, detects conflicts with existing knowledge, and uses a knowledge graph + RAPTOR trees + multi-strategy RAG underneath.
• getguru.com — More mature/commercial, but it's better described as an AI knowledge layer than a literal graph product. It can capture knowledge from Slack conversations and turn decisions/updates into reusable, verified knowledge automatically.
• tettra.com — Its Kai bot builds a knowledge base from conversations already happening in Slack, but again, this is primarily a knowledge base/RAG system rather than an explicit knowledge graph.

My shortlist

If the graph itself is important, I'd investigate Entilla first, then Tallow.

If what you really want is “our Slack conversations become a continuously updated company brain that people can query”, I'd also put Guru and Slack's own Slackbot/Enterprise Search into the comparison. Slack now supports AI over permissioned conversations and connected data, including contextual answers and daily recaps.

If you tell me roughly how big your Slack workspace is and whether you want SaaS vs. self-hosted, I can narrow this to the 2–3 products that are actually worth evaluating.

If you mean a Slack bot that does more than search messages—one that extracts entities, decisions, relationships, and builds a living knowledge graph from daily conversations—the market is still fairly small. A few options to look at:

A few evaluation questions matter because “knowledge graph” can mean different things:

• Do you need a visual graph? (nodes like people, projects, decisions, customers, technologies)
• Do you need automatic extraction of decisions and ownership? (“Alice decided X on Aug 12”)
• Do you need an API/MCP layer so other AI agents can query the graph?
• Do you need private channels and DMs indexed, or only selected channels?

For a company trying to create a “second brain” from Slack, I’d shortlist:

1. Noos for the closest match to a conversation-derived graph.
2. Kennis for broader organizational memory and agent access.
3. A custom Slack app + graph database (e.g., Neo4j/GraphRAG style) if you need full control over the ontology and data model.

If you tell me your company size, Slack plan (Business+/Enterprise/etc.), and whether you want buy vs build, I can narrow this down.

Finding a turnkey, out-of-the-box bot that exclusively builds a standalone knowledge graph purely from Slack chat logs is rare because raw chat is notoriously messy (full of jokes, half-baked thoughts, and tangents). However, a few major enterprise AI and knowledge platforms specifically connect to Slack, index daily conversations, and use a underlying knowledge graph to map people, content, projects, and tribal knowledge.

Top options offering this capability include:

• Glean : Glean features a comprehensive enterprise work assistant that integrates directly into Slack. It continuously indexes Slack conversations alongside your other apps (Google Drive, Jira, GitHub, etc.) and constructs a real-time Glean Knowledge Graph mapping your organization's content, team members, and daily activity to provide contextual answers.
• Slack AI : While not a traditional visual node-and-edge knowledge graph you can browse externally, Slack's native AI uses underlying semantic and relational frameworks derived from Slack's architectural knowledge graph approach . It connects daily channel and thread conversations, surfacing past decisions and summaries without requiring an external tool.
• Graphlit : If you are looking for a developer-focused or customizable API approach, Graphlit allows you to ingest unstructured content and Slack conversation channels to automatically construct a dynamic knowledge graph, which you can then hook into a custom conversational Slack bot.
• Custom Stack (Neo4j + LLMs) : Many engineering teams bypass commercial products by setting up an automated pipeline using a vector/graph database like Neo4j coupled with frameworks like LangChain or LlamaIndex and an LLM to parse Slack exports or webhook data into structured entities and relationships.

Are you looking for an off-the-shelf enterprise search tool (like Glean), or do you need a developer/API framework to build a custom graph yourself?

Yes. I found a few products that are close to what you mean, but Kennis and Beever Atlas are the closest matches to “turn our daily Slack conversations into a living knowledge graph/knowledge base.”

Best matches

1. Kennis — probably the closest fit.
   It connects to Slack, synthesizes the last 90 days of conversations, then updates the knowledge base every day. It explicitly extracts decisions, policies, and facts with links back to the source conversations. Public channels are included by default; private channels/DMs are opt-in.

2. Beever Atlas — the most explicitly knowledge-graph-like option I found.
   It ingests Slack conversations, extracts atomic facts, deduplicates them, clusters them into topic pages, and uses a graph store to connect people ↔ decisions ↔ projects ↔ conversations. It can also answer questions with citations back to the original Slack messages.

3. Guru — more mature/enterprise-oriented.
   Guru's Slack integration turns conversations, decisions, and updates into reusable organizational knowledge, while continuously checking knowledge for accuracy and staleness. It is less literally a “knowledge graph” product and more a governed organizational-memory layer.

4. Slack itself — increasingly capable, but not really a knowledge graph.
   Slackbot/Enterprise Search can reason over your permissioned conversations, files, and connected apps, and Slack now describes Slackbot as knowing the “conversations, decisions, and data” accumulated by the organization.

My take

If your desired architecture is:

Slack conversations → extract entities/facts/decisions → connect them into a graph → continuously update organizational memory → query it conversationally

I'd investigate Beever Atlas first if the graph is the key requirement, and Kennis first if you want something that feels more like a turnkey Slack bot your team can deploy quickly.

One important distinction: many “Slack knowledge” products are really RAG/search systems—they retrieve relevant messages when you ask a question. That's different from actually constructing and maintaining a persistent graph of entities and relationships. Beever Atlas is unusually explicit about doing the latter.

If you tell me your team size and whether you want SaaS vs. self-hosted, I can dig into 5–10 options and compare Slack permissions, private-channel handling, graph structure, pricing, API/MCP support, and data retention.

Building a true knowledge graph (extracting distinct entities, people, and relationships rather than just doing text-chunk vector search) directly from daily Slack conversations requires either a specialized enterprise AI platform or a custom developer stack.

Options for obtaining a Slack bot that handles this include:

1. Out-of-the-Box Enterprise Platforms

• Writer (Writer AI Studio): WRITER features a proprietary graph-based RAG framework called Knowledge Graph . Their native integration for Slack allows you to connect enterprise data—including historical channel messages—into a structured graph-based model so that internal AI agents can reason over decisions, roles, and definitions. You can deploy agents directly into your Slack channels.
• KnodeGraph: Offers a dedicated KnodeGraph Slack Integration designed to export Slack channels and automatically convert message history into a typed knowledge graph tracking decisions, owners, and active threads.

2. Developer & AI Infrastructure Platforms (Requires Setup)

If you have engineering resources or prefer building a custom pipeline rather than using a rigid SaaS product, these platforms provide APIs specifically meant to turn chat streams into knowledge graphs:

• Graphlit: Provides native ingestion for Slack messages. It automatically extracts entities and relationships from unstructured chat data and builds an internal knowledge graph that you can use to power a conversational RAG bot in Slack. You can check out their implementation patterns in the Graphlit Knowledge Graph Tutorial.
• Memgraph: An open-source, in-memory graph database that features reference architectures (like their open-source slack-influence-bot on GitHub) showing how to stream real-time Slack events into a graph database to query interaction networks and community affinity via a chat plugin.

Note on "Standard" Slack AI vs. Knowledge Graphs

Be careful with native Slack AI or standard RAG search bots (like Question Base or Slite Agent). While they summarize daily threads, search historical chats, and act as great knowledge bases, most of them rely on vector embeddings/semantic search rather than a true node-and-edge knowledge graph.

If you prefer a fully managed no-code app versus a developer tool/API to wire up yourself, let me know your preference and I can help narrow down the implementation steps.

For an out-of-the-box enterprise solution that plugs directly into Slack, Glean and Guru offer AI-driven workplace assistants that index your daily conversations and documents into a unified semantic map . If you are looking specifically for structured knowledge graph architectures (storing entities, users, topics, and relational edges extracted from chat data), platforms like Graphlit provide developer APIs and conversational Slack bot kits that automatically ingest content and build an internal knowledge graph , while open-source blueprints using graph databases like Memgraph or Neo4j can be paired with custom Slack event bots to map your team's communication network in real time.

Enterprise AI Assistants with Semantic Search

• Glean : Indexes your company's apps, documents, and Slack history to function as an intelligent assistant that understands how internal topics and people relate.
• Guru : Combines enterprise AI search, intranet features, and Slack bot integrations to capture conversational knowledge and surface verified company facts.
• Slack's Native AI / Slackbot : Uses semantic reasoning and enterprise search to connect your conversations, files, and channels into a contextual web of information without requiring an external database.

Graph-Specific & Developer Platforms

• Graphlit : Ingests unstructured Slack message history and automatically structures it into a queryable knowledge graph for RAG-based chat applications.
• Memgraph : Offers open-source streaming graph architectures with community projects (like the Slack Influence Bot) designed to map real-time message interactions and community affinity.
• Neo4j : Frequently used with custom Python/Node scripts via Slack APIs to parse message taxonomies, user nodes, and channel relationships into a robust graph database.

If you prefer a fully managed SaaS tool versus setting up a custom developer pipeline (like Neo4j or Graphlit) , let me know your security requirements or budget so I can narrow down the best fit.

Yes. There are a few options, but Noos is the closest match to exactly what you described.

• Noos — explicitly turns Slack conversations into a searchable, continuously maintained knowledge graph. It can capture decisions, ideas, and discussions, automatically link related conversations, and let people query the resulting graph from Slack.
• Ariv.ai — a Slack knowledge bot that says it creates a machine-readable knowledge graph from Slack interactions, extracting meaning and context from conversations. It is available through the Slack App Directory.
• Tallow — another interesting option that continuously watches Slack plus tools such as Confluence, Jira, and GitHub and builds a living knowledge graph connecting conversations, documents, tickets, and code. It currently describes itself as coming soon/waitlist, so it's less immediately deployable.
• Slackbot itself has moved substantially in this direction. In 2026, Slack added memory and increasingly sophisticated contextual understanding of your organization's conversations, although Slack describes this as AI/memory rather than exposing a traditional knowledge graph for you to inspect.

What I'd look at first

If your requirement is specifically:

"Read our ongoing Slack conversations → extract decisions/people/projects/topics → connect them into a persistent graph → let us query that accumulated organizational memory."

I'd investigate Noos first, then Ariv. Noos's product description is unusually close to that exact workflow.

One important distinction: Slackbot/Slack AI is primarily an answer/search/memory layer, whereas Noos and Ariv explicitly describe a knowledge-graph representation underneath. Slack now also provides an MCP/Real-Time Search API that lets third-party agents securely access Slack's conversational context, which makes this ecosystem considerably more viable than it was a year ago.

If you tell me your company size and Slack plan (Pro, Business+, Enterprise), I can narrow this down to the 3–5 products I'd actually evaluate, including privacy/data-retention, whether they ingest DMs/private channels, pricing, and whether the graph is genuinely persistent vs. just RAG/search.

Yes. The closest match I found is Noos. It explicitly connects to Slack and says it turns Slack conversations into a searchable knowledge graph, linking ideas, decisions, people, and related discussions.

There are two other routes worth considering:

• Slack itself — Slackbot now has memory and uses workspace conversation context, and Slack has been increasingly positioning its AI around organizational knowledge discovery. However, I don't see Slack claiming that it exposes a persistent, explicit knowledge graph of your conversations in the same way Noos does.
• Custom Slack agent — Slack's new MCP Server + Real-time Search API makes it substantially easier to build an agent that continuously extracts entities/decisions/relationships from conversations and stores them in a graph database.

If your requirement is specifically: “install a Slack bot, let it observe our daily conversations, automatically turn them into a living knowledge graph, and then query that graph later,” Noos is the most direct fit I found.

One caveat: I'd want to vet its Slack permissions, retention/deletion behavior, private-channel/DM coverage, and whether the graph is actually exportable before putting sensitive company conversations into it.

Several enterprise AI platforms and developer tools integrate with Slack to ingest daily chats and structure them into a knowledge graph (often using graph-based RAG rather than simple vector search).

The primary options that provide or build knowledge graphs from Slack conversations include:

• Graphlit — A developer-focused context layer and API platform that connects to Slack via automated data feeds. It continuously ingests messages, extracts entities and relationships, and automatically builds a searchable knowledge graph that can power custom conversational bots.
• Writer — An enterprise AI platform featuring native Writer & Slack Integration . It utilizes Knowledge Graph (a built-in graph-based RAG technology) to analyze team chats, documents, and connected data sources to drive contextual answers and automation directly inside Slack.
• Naumu — A context-sharing platform tailored for team memory that integrates with Slack. It automatically weaves everyday discussions and thread chats into a live, structured knowledge graph, serving as a shared workspace memory that can be queried natively.
• Memgraph (Open Source) — For teams with engineering resources, Memgraph provides an open-source reference implementation (slack-influence-bot ) that connects to Slack in real-time via a message stream, stores community and conversation data into a graph database, and runs graph analytics via a custom bot.
• Custom GraphRAG Pipelines — Many organizations build their own by pairing the Slack Events API with graph platforms like Neo4j (using their graph-based extraction patterns) combined with LLM entity-extraction frameworks like Zep's Graphiti or Cognee.

If you'd like to narrow this down, tell me:

• Are you looking for a ready-to-use SaaS tool or an API/developer framework to build your own?
• Do you also need to connect other tools (like Notion, Jira, or Google Drive) into the same graph?