Making Enterprise Recorded Meetings Easy to Discover and Share

1. Introduction

In global enterprise, more and more knowledge workers attend remote meetings. This trend will not only continue, but also become more prevalent due to globalization. Every day, thousands of online meetings are held. During these meetings, critical decisions are made, important information is shared and future action items are defined. Since these meetings contain rich enterprise expertise, the ability for knowledge workers to discover and share online meetings is highly valuable in an enterprise.

Although online meetings can be recorded easily using existing web conferencing tools (e.g., IBM Sametime Meetings and Cisco WebEx), recorded meeting archives are rarely revisited and shared. There are three main reasons:

• 1.

  Poor Content Indexing: Online meetings are often recorded as audios and videos. Without proper indexing, the content of these meeting recordings is not searchable. Currently, it is difficult if not impossible for users to find meetings that satisfy their information needs.

• 2.

  Scattered Meeting Artifacts: In addition to videos, there are other meeting artifacts (e.g., presentation slides) and metadata (e.g. meeting time, participant names) associated with a meeting. Typically meeting information is scattered on different servers: high-level meeting metadata is stored only on the meeting hosting server; meeting videos are stored and shared on a video server, and slides are stored and shared on a slide server. Since these meeting artifacts are not directly linked to each other, it is difficult to locate all the information related to a meeting.

• 3.

  Lengthy Meeting Recordings: Unlike videos shared in the public domain, which are often short (e.g., average ~4 minutes on YouTube), enterprise meetings are long (normally 30 minutes to 1 hour). Frequently only a portion of a recorded meeting may be relevant to a user’s information needs. Since currently most recorded meeting repositories do not support fine-grained meeting sharing (e.g. sharing the entire video instead of a video segment), users would have to sift through a long video from the beginning to the end to look for specific information. Fine-grained meeting sharing is often more effective and thus would be more preferable.

To address the above problem areas, we developed an end-to-end enterprise meeting service Agora. Our main contributions are:

• 1.

  Deploying an End-to-End Enterprise Meeting Service: We have built and deployed an end-to-end enterprise meeting hosting and sharing service that allows users to host, store, search and share meetings.

• 2.

  Seamlessly Integrating Existing Enterprise Services to Facilitate Fast Development: Developing a comprehensive meeting service is highly complex. In Agora, we focused on developing new technologies that can seamlessly integrate distributed content and functions together to create a new enterprise meeting service.

• 3.

  Enhancing Meeting Discoverability Via Collaborative Content Creation And Indexing: To enhance meeting discoverability, Agora meetings are indexed with a rich set of metadata. They are created either automatically by the system using content analytics (e.g., captions are automatically created for meeting videos) or manually added by humans (e.g., user added tags and meeting description). Moreover, since machine-created metadata are often imperfect, Agora allows users to correct mistakes in the metadata, whether created automatically or by another person.

• 4.

  Supporting Fine-Grained Meeting Search and Sharing: Agora enables fine-grained search by anchoring the meeting index on a timeline. Thus, the search results may point to not only the full meeting recording but also to the period(s) of time that the search result belongs. Moreover, to facilitate easy sharing, Agora allow users to share any portions of a meeting by specifying a start and end time.