How to connect internal data to ChatGPT for business

TL;DR

Connect private knowledge sources to ChatGPT and use record mode to capture meeting content while respecting user permissions. Deep Research can search public web and private apps like HubSpot, Outlook, SharePoint, and Teams, enabling targeted queries and cross app references for faster insights.

Transcript

hey everybody thanks for joining today i'm Nate and I lead the chat for chatbt for business product team now today's live stream is all about how we help bring your organization's unique work context directly into chatbt and so way that we do that is we enable you to connect your internal knowledge sources to chatpt uh so that you can find the most... Read More

Key Insights

• Private enterprise data can be connected to ChatGPT using Deep Research and connectors to enable extensive search across both public and private sources.
• Deep Research can search and analyze data from apps like HubSpot, Outlook, SharePoint, and Teams in addition to the web, expanding context for queries.
• There is a focus on permission aware and enforced data access so users only see documents they have rights to.
• Record mode captures and leverages spoken information from meetings to improve information retention and usefulness.
• Users can toggle data sources manually in the composer to direct where ChatGPT should search for information.
• The system provides source visibility so users can verify information and deep dive into specific documents as needed.
• Synthesis features help convert raw data and interviews into actionable plans and documents suitable for planning cycles.
• The workflow demonstrates quick turning of research into a Q3 plan by synthesizing interviews and data into strategy documents.
• More videos from this channel:
  • How to balance accuracy latency and cost at scale
  • How Do OpenAI Dots and ChatGPT Spaces Work?
  • How to use AI as a coding collaborator effectively
• More from this channel: OpenAI

Questions & Answers

Q: How can I connect internal data to ChatGPT for business use

You can connect internal data to ChatGPT using two main approaches shown in the video: Deep Research, which extends search across web and private apps, and connectors that link internal sources like HubSpot, Outlook, SharePoint, and Teams. This enables targeted queries, comprehensive research, and faster access to relevant information while respecting existing user permissions.

Q: What is Deep Research and how does it work

Deep Research is an agentic capability that allows extensive search and analysis of the web and internal sources. It can connect to private apps to pull in contextual data, then run complex queries to surface relevant deals, discussions, and documents. It highlights sources and maintains permission awareness so results are verifiable and compliant with access rights.

Q: How does record mode help in meetings

Record mode captures information spoken in meetings and helps turn it into actionable insights. It works with internal context so you can extract key points, decisions, and data references without requiring manual transcription. This feature accelerates knowledge capture and ensures information is readily usable for planning and reporting.

Q: Can I control which data sources ChatGPT searches

Yes, you can toggle sources in the composer to specify which data sources ChatGPT should search. This gives you direct control over the context and helps tailor results to your current task. It also supports source prioritization for more relevant outputs.

Q: What kind of sources are demonstrated in the video

The video demonstrates connecting to private sources such as HubSpot, Outlook, SharePoint, and Teams, alongside the public web. These sources are used for deep research and rapid synthesis to inform Q3 planning and partnerships for products like Bark to Text and Animal GPT.

Q: How is data privacy handled in these workflows

Data privacy is addressed through permission aware and enforced access controls within enterprise workspaces. Only documents a user has rights to are visible, which ensures confidential information remains restricted to authorized users. This is a core part of the demonstration in the video.

Q: What is the practical outcome of using these features for planning

The practical outcome is faster, data-driven planning. By synthesizing internal documents, interviews, and analytics into a concise plan or document, teams can generate a clear Q3 strategy. The workflow reduces manual data gathering time and improves decision making with accessible, verified sources.

Q: How does the interface support verifying information

The interface shows sources alongside model responses, enabling users to verify the information directly. You can click on sources to deep dive into documents, cross-reference data, and ensure the provided insights are backed by specific sources within your allowed access.

Summary & Key Takeaways

• This video explains two main features: connectors to internal data and record mode. It demonstrates how these features help synthesize internal knowledge with chat GPT for business planning and daily work tasks. The demo uses a fictitious AGI Corp to illustrate practical workflow improvements.

• The presenters show how to search across private sources, verify sources, and generate actionable summaries. They emphasize permission awareness and data privacy within enterprise workspaces. The flow includes deep research, source toggles, and synthesis of user interviews.

• The content highlights using data from various apps to drive Q3 planning and product strategy, including time series, user interviews, and board materials. It emphasizes speed and accuracy when pulling data from multiple sources into ChatGPT for decision making.