Thought experiment, only a small step from today’s capabilities: Microsoft offers an AI-based query interface to a combination of the whole internet plus all the content in your organization’s Office 365 apps (mail, calendar, documents, meeting content, instant messages). You can swiftly get answers to:
- Which people in our company know the most about ______ ?
- Summarize the key product information about X and compare with Y
- Who does ____ and ____ talk with most often about ____?
- What are employees saying about this leader?
- Please find a slide from before <date> which showed our inventory trend line
- Please find relevant spreadsheets from the last 3 months with data on ____
- Compare the marketing plans for NA and EMEA regions
- How many employees are making resumes?
- Please summarize evidence that anyone on this distribution list are being recruited by a headhunter or are thinking of leaving.
- What is the protocol or procedure for ______?
- What are the three most common items in employee development plans this year compare to last year?
- Summarize the comments posted in the last town hall
- Who are the most likely people to bridge between these two functions in the company based on their email correspondence and meeting attendance?
The benefits would be palpable. This would be a very attractive as a tool to extract more value from the phenomenally large amount of unstructured and messy data. It would also be a boon to reduce the cost of employee turnover. Re-use of information would reduce time wasted re-constructing something already available.
What are the consequences?
- HR and IT groups will need to update rules and constraints about personnel information to ensure compliance with laws (which vary by legal entity). Legal will want to review the rules about retaining information.
- Some people will insist that their information be excluded from this level of query, based on their roles. Attorneys, for example, might have a good case for this.
- Employees might intellectually know that all the content belongs to the company, not them personally, but today few behave this way. Employees will be increasingly reluctant to record anything about their opinions or insights which might surface in unexpected ways.
- Employees will need coaching and help to ask better questions, or intentionally ‘tag’ or label some content (e.g., presentations and meeting notes) to make them more accessible to this search.
- There might be a recognition/reputation opportunity for people who generated most-used content. (Similar to how scientists build a reputation by the number of citations for their published papers.)
- It will be difficult at times to assess the accuracy or relevance of what such a query returns. What is signal, and what was signal then but noise now? “Trust and verify” will be important.
- There will be significant temptations to use this additional capability in “Big Brother” ways to monitor employees.
What have I missed?
