You are here:
Using Unstructured Data in Data 360
Use unstructured data in Data 360 to ground Agentforce agents, generative AI, analytics, and automation use cases with business-specific data that delivers deeper insights for your users and customers.
Unstructured data comes in many forms. Consider these common examples.
| Type of Data | Examples |
|---|---|
| Text documents | Case notes, reports, internal memos, handwritten notes, invoices, business cards, and emails that contain detailed descriptions or conversations |
| Multimedia files | Audio files from customer service calls, podcasts, video recordings from training sessions, tutorials, meetings, images from product reviews, infographics, scanned documents, and medical imaging |
| Social media content | Comments, posts, or messages containing customer feedback, opinions, or sentiments |
| Web content | Blogs, articles, and forum posts that discuss products, services, or customer experiences |
| Chat logs | Conversations from live chat support, often containing nuanced questions, complaints, and preferences |
| Sensor and IoT data | Data from wearable devices or machinery that doesn't follow a standard format, such as raw accelerometer data |
| Customer feedback | Survey responses, feedback forms, or reviews written in free text |
Data 360 supports several connectors and file formats for unstructured data. See Unstructured Data Connectors and Data 360 Limits and Guidelines.
Using unstructured data in your AI, automation, and analytics workflows can help you make informed decisions and improve customer experiences.
Agentforce Agents
Data 360 provides customer data to make agents more contextually aware and knowledgeable. For example, when a customer contacts an Agentforce Service Agent, Data 360 can add unstructured data, such as information from past emails, support tickets, product photos, or voicemails, to help the agent to better understand the customer's perspective. Data 360 hybrid search can then find the appropriate knowledge articles to resolve the issue. When a service case is resolved, Data 360 can guide the agent with next steps, such as automating follow-up emails or preparing detailed chat summaries.
Here's an example of an agent in action.
Generative AI
Retrieval augmented generation (RAG) improves AI responses by grounding them in unstructured data, making them more relevant and accurate. To learn more, see Using Retrieval Augmented Generation and Create a Vector Search Index with Advanced Setup.
In Service Cloud, agents can use hidden data in knowledge articles to answer customer questions more accurately. By working with Data 360, agents can look at unstructured data, like emails, support tickets, and chat logs, to get more information and write better responses, including suggestions for fixing problems before they happen.
In Sales Cloud, agents can leverage unstructured data, including emails, call transcripts, and notes, to generate detailed and contextually relevant meeting briefings and summaries. These insights improve new emails with personalized content and customer-specific recommendations, improving the customer experience and sales team effectiveness.
Automation
Agents can automate the identification of similar records, such as Cases, based on their descriptions to flag duplicates and speed up resolution. Additionally, agents can classify records by finding and using semantically similar ones, improving efficiency and accuracy.
Analytics
Agents can use Tableau to gain insights from unstructured data, such as topic classification. Check out the example in Enhancing Data Analysis with Vector Search in Tableau.
Team-based Examples
Consider these use cases for how to leverage unstructured data across the various functions.
| Team | AI | Automation | Analytics | Search |
|---|---|---|---|---|
| Sales | Draft responses to RFPs using previous proposals, contracts, and documentation. | Analyze email conversations to automate lead scoring. | Analyze recorded sales calls to spot patterns in competitive insights or feature interests. | Prep sales calls with insights from thousands of PDFs and emails. |
| Service | Estimate repairs through photo analysis. | Proactively send repair recommendations by identifying patterns in agent assets. | Analyze chat and product logs to identify common recurring service problems. | Get context-based knowledge article recommendations. |
| Marketing | Find subject lines that have worked on similar email templates. | Automate campaign follow-up based on specific product interests found in support chats. | Spot trends based on customer feedback on websites. | Search feedback from surveys to create more responsive campaigns. |
