Smorgas Chat | Smorgas.ai

Smorgas Chat changes all of this by giving their time back and applies the content to address their tasks and problems. The offering connects with the company’s files and assets (share point etc.), allowing users to leverage intelligent search, giving them the ability to get the documents they are looking for and answer their contextual questions.

It leverages, LLMs and Retrieval-Augmented Generation (RAG) to execute inquiries. Users can securely access documents across a company's disparate data silos and get answers fast.

Resolving the “Data Silo” issue

Most companies have data scattered across Slack, Jira, Google Drive, MS 365, and various SQL databases. Instead of moving all that data into a single vector database (which is expensive and creates security risks), Smorgas uses Federated Search. It understands permissions out of the box — if you don't have access to a specific folder in SharePoint, Smorgas Search won't show you results from it.

Federated Search: It queries the data where it lives in real-time.
Security & Permissions: It respects the existing permissions of the source systems (e.g., an agent can only "see" the files the user has permission to see). users only see search results for documents they already have permission to access
RAG-Ready: It provides a clean, unified stream of data that can be fed into Large Language Models (LLMs) to provide accurate, grounded answers.

Smorgas Chat compliments other Smorgas Solutions

Smorgas Chat is for Supply Chain or Accounting; Smorgas Chat serves as the "Search Tool" the agent uses to perform its job:

In Purchasing: If an agent needs to "find the best price we've paid for this part in the last three years," it can use Smorgas Chat to simultaneously search historical invoices in the ERP, contract PDFs in SharePoint, and recent quotes in Gmail.
In Accounting: If an agent is investigating a reconciliation error, it can use Swirl to pull the related Slack conversation between the buyer and vendor and match it against the transaction log in a SQL database.
In Supply Chain: An agent could use Swirl to monitor external news feeds or weather data alongside internal logistics reports to predict a delay.

‍