# Streamline AI Email Intake Streamline AI’s Email Intake feature transforms how legal and business teams handle incoming email requests. ## AI-Powered Email Parsing The AI Intake system automatically processes incoming emails, identifying key information to create a new request and populate form fields, saving time and reducing errors. ## Intelligent Field Auto-Population AI suggests field values such as request type, counterparty information, and dates, making it easier to complete requests. ## User Review & Confirmation Users can review AI-populated fields and either confirm or edit the suggested values, ensuring accuracy before proceeding. ## Seamless Workflow Execution Once confirmed, requests move seamlessly into Streamline’s workflows and behave like other requests, though your activity history will indicate if the request was triaged by AI Intake, with clear indicators showing that AI processed the request. ## How it works By leveraging OpenAI’s ChatGPT-4o mini, the AI Email Intake automatically parses incoming emails, creates requests, and predicts the value of relevant fields so users only need to confirm the AI suggestions, saving significant time when processing incoming requests. This dramatically reduces the manual effort required for triaging email requests and enables users to focus on higher-value tasks, like decision-making and approvals. ### Step 1 **AI-Powered Email Parsing** The system reads incoming emails and guesses the appropriate request type and populates field values. ### Step 2 **Request Review** The user reviews AI-populated data, with clear UI elements highlighting which fields were auto-populated. ### Step 3 **Edit & Confirm** Users can edit AI predictions or revert to original AI values, then confirm the request to save it. ### Step 4 **Workflow Activation** After confirmation, the request triggers workflows like approvals, rejections, or task assignments. ## Core Use Cases **AI Successfully Guesses Request Type** AI processes the email, assigns the request type, and populates fields. The user is notified and can review and confirm the information. **AI Fails to Guess Request Type** The email is listed in the tracker for manual triaging by the user. The user can assign the correct request type and proceed. **AI Guesses the Wrong Request Type** The user can manually update the request to the correct type. Coming Soon: users will be able to select the correct type and re-run AI Intake. **Hashtag Use** Users can add hashtags to the email body (e.g., #NDA) to categorize requests under the correct matter type. The AI processes the email based on the hashtag. ## Technologies - **Natural Language Processing (NLP)** Powered by ChatGPT-4o mini, the AI processes unstructured email content and extracts relevant data. - **AI Suggestions** Field values are auto-populated based on the email content. - **Real-Time Parsing** AI parsing updates occur in real-time, with visual feedback for any user actions. ## Data Privacy and Use Policy ### Data Storage Email data is stored securely on OpenAI servers for 30 days before deletion as part of general operations and abuse detection. ### No AI Training on Customer Data OpenAI does not use customer data to train its models. All data processed remains isolated and is not commingled between customers. ### Security Practices Communications are secured, and customer data is fully protected, complying with industry standards. > "Streamline AI enabled us to deliver consistent 24-hour SLAs across multiple teams, including Legal, Finance, People Ops, InfoSec, and IT, as we scaled our vendor needs by 20% each quarter." > Ronak Ray, General Counsel at Pantheon ## Frequently Asked Questions **How does AI Email Intake work?** The AI reads incoming emails, extracts key information, and creates a new request with auto-populated fields. Users can review and confirm the AI-predicted data before saving the request. **Can I edit AI-generated field values?** Yes, users can modify AI-predicted fields and revert to the original AI predictions if needed. **How secure is the email data?** Email data is processed securely, with encryption for both storage and transmission. OpenAI retains the data for 30 days for operational purposes before it is deleted. **Will AI always predict the correct request type?** While AI is designed to make accurate predictions, there may be instances where the request type is incorrect. In this case, users can manually triage the request into the correct type. **Does Streamline use customer data to train the AI model?** No, Streamline does not use customer data to train OpenAI’s models. Data is only used to generate the AI suggestions within the AI Email Intake feature. **Will requests generated by AI behave differently from manually created/traditional requests?** No, once the user confirms the entries, an AI-generated request will behave like any other request in the system. **Who should send emails into Streamline?** When thinking about how emails enter Streamline, users typically consider if only the Legal team will be allowed to send emails or if other users across the business will also submit email requests. Most successful customers leverage email intake by: 1. Having Legal individually forward emails and setting the expectation with the requestor that further collaboration happens on the request inside Streamline. 2. Using a dedicated forwarding alias that auto-creates requests in Streamline, while setting the expectation with the requestor that further collaboration happens inside Streamline. **Are requestors required to respond via email?** - **Internal requestors:** It is best practice to set up internal requestors via SSO and user auto-provisioning in Streamline to ensure the business gets notifications about status changes and can collaborate on the request. - **External requestors:** Emails from external requestors are received in Streamline and responded to in email threads. However, they will not have access to the Streamline request in the platform; they will see communications sent to them via the “Email” tab in the collaboration hub. **Should email intake be a primary way requests get created?** If you choose email as a primary intake method, keep required fields in forms minimal to facilitate smooth request flow. ### Easy best practices to follow - **Use email for one-off use cases**: Forwarding select, one-off emails into Streamline is an easy way to bring requests into the system for users who refuse to learn a new workflow. Example: Forward a direct email from the CFO into Streamline and continue the conversation from there. - **Auto-forward from systems and aliases your team already uses:** This allows Streamline to categorize and route incoming requests automatically, saving time and reducing manual review. - **Work out of Streamline - not your email inbox:** Ensure that communication is tracked in Streamline to maintain a clean audit trail. **What types of use cases should I discuss with a CSM?** If you have multiple teams using Streamline or groups that shouldn’t see each other’s requests, email intake may require extra configuration. Work with your Customer Success Manager to set it up correctly. **Avoid these when thinking about email intake:** - Relying solely on email intake without bringing in other organization members as business users. - Forwarding an entire inbox to Streamline, which could introduce irrelevant requests. - Manually forwarding every individual message, leading to duplicate work and clutter. ## Work smarter Scale your legal team's efficiency and effectiveness with modern workflow automation tools designed for in-house legal.