Overview

Moveworks offers the ability for users to grant access to search across and execute workflows against their data using API calls to the source system in real-time. Connections to the source system of record that will leverage this kind of execution are called Live API Integrations.

How does Live Search work in Moveworks?

The diagram below explores how we use Live API Integrations within our Enterprise Search experience to pull resources that the user has access to at query time.

1. Once the User makes a query, the Query Planner will send the rewritten query to the Resource Retriever, which orchestrates the search across Indexed Content and all of the Live API Integrations.
2. Resource Retriever will check against the User Token store to understand for which source systems the User has granted Moveworks access for Live API Integrations.
3. Based on those systems, the Query Planner will decide the optimal queries to send to the Live API Integrations and Indexed Search.
   1. Indexed Search will use the query to retrieve results from the Customer’s Ingested Content, with permissions enforced, and further apply thresholding and filtering for relevancy.
   2. Live API Integrations will reach out to the relevant source system with a direct search request to retrieve and map the results.
4. Resource Retriever will bundle these results and pass them up to the Ranking and Policy Enforcement steps, to further enforce only the most relevant results.
5. The final results are returned to the User, and summarized where appropriate.

The overall process is similar to how we retrieve results from Ingested Content for Indexed Search — however, there are a few key additions:

• User Consent: Within our Enterprise Search product, users can provide consent to connect to the source system via oAuth2 and grant access to the content that they have access to. This could include content that is not present within the Ingested Content that has been configured by a Moveworks Admin. This consent is stored as a User Access Token.
• Query Planning: In the case of Live API Search, we see that different source systems, due to the nature of the search APIs that they provide, will benefit from having the query re-written in different ways.
  • Where Query Planning may manifest in one set of optimal query re-writes for one system (e.g. Microsoft Sharepoint), it may decide to use different re-writes in another (e.g. Slack). These re-writes are generally different from the queries written for Indexed Search as well.
  • Query Planning for Live API Search can also include slot detection for API parameters like filters. If the user’s original query is something like “find me the meeting notes from last week” the Planner may decide to break the query into a search parameter like “meetings notes” and a filter parameter such as “> 1 week”.