Slack’s New Bot To Monitor Employees Shows Promise of Enterprise Intelligent Assistants

slackbot_12Enterprise collaboration software maker Slack is reportedly working on artificial intelligence targeted at automating managerial tasks. Though use cases may vary, the role intelligent assistants play within the enterprise could have profound business impacts, especially for process automation.

Certain job roles — for example, the role of the systems administrator — have long been supported in key tasks by software agents. System admins don’t stay up all night monitoring the health of critical infrastructure. Software products perform this job for them and wake the system admin with an alert in the event of a critical failure.

These system monitoring tools must interpret many variables and possible failure modes. But such systems are far less complex than ones designed to automate processes involving humans. The most difficult tasks call for communicating directly with humans using natural language.

Recently, companies such as x.ai, have built algorithms that enable software agents to automate higher-level tasks involving the coordination of information gathered from human beings. As we previously described, x.ai is focused on automating the complex task of appointment scheduling. There so many permutations of setting up a meeting involving multiple participants that the universe of possibilities a “simple” scheduling algorithm needs to cover is staggering. And x.ai manages the entire process with a software agent that “converses” with people.

Slack’s goal is apparently to construct a bot that can automate the process of collecting team status updates. As a recently retired branch chief at a federal government agency, I can attest to how onerous, time-consuming, and unavoidable it is to feed the weekly status update culture. Slack is already home to many Slack-built and third-party bots that automate everything from scheduling meetings to ordering lunch. Adding a bot that can contact team members and get pertinent information to add to a summarized weekly status update seems like a worthy goal.

Though status reporting may sound on the surface like a simple process to automate, I know from experience that it can be fraught with complexities. Some team members won’t respond to the request for updates. Others will provide updates that are way too verbose. Some team members will forget to mention projects the boss really cares about, but delve into details of all the others. How do you design algorithms to deal with these cases?

It will be interesting to watch the evolution of Slack’s status reporting bot. Many processes within the enterprise could benefit from automation. As artificial intelligence technologies mature, the opportunities for bots in the enterprise are boundless.



Categories: Conversational Intelligence, Intelligent Assistants, Articles