All.ai
Improving Business Communication and Empowering Women in Meetings Through NLP Language Analysis Tools
Last week, I was fortunate to have an opportunity to contribute my data science and engineering skills to a small team led by two inspiring women in tech who were just named to the list of BBC 100 Women for 2017. As part of that nomination they were challenged to a one week hackathon: apply technology to tackle the problem of the glass ceiling.
Our case study was Erin, a confident and capable data science manager at Asana who is nonetheless noticing new difficulty participating in higher-level meetings with mostly male counterparts. Women like Erin often find themselves less supported in meetings with masculine decision-makers. The consequence affects everyone whether they realize it or not: a diversity of representation is essential to reach higher level of performance in group endeavors.
- Hong, Page et al. studied how groups of diverse problem solvers outperform groups of high-ability problem solvers
- McKinsey confirmed that companies with diverse executive boards enjoy significantly higher earnings and returns on equity.
- Forbes identified workforce diversity and inclusion as a key driver of internal innovation and business growth
- Scientific American has elaborated how diversity makes us smarter.
Most smart business leaders know this by now. Change requires going beyond recognizing the problem, to recognizing effective measures for improvement. The issues that suppress diversity in group power dynamics stem from deeply-ingrained, sometimes automatic behavior. Even with efforts to educate, we cannot expect ourselves to be constantly vigilant for every lurking bias, subtle negligence, or subversive social convention. We are only human.
Enter our machine allies.
The app we created, All.ai, listens to how people speak in a meeting and offers personalized feedback to each participant. It works by transcribing speech using IBM Watson, and then performing semantic analysis on the resulting text transcript.
All.ai can easily be extended to:
- tell you much of time men vs. women had the floor
- catch how often you hesitate when trying to be assertive
- highlight the adjectives used to describe others in the meeting
- list the words you use and what they convey to others
- summarize your spoken contribution in a given meeting
- identify sexist language
- do anything else you can do by analyzing text
All.ai is intended to be descriptive and helpful, like a stenographer you can call on at any time. Maybe you’d like to demonstrate to the boss how people aren’t taking your ideas seriously in meetings. Or you’re the manager analyzing your team to ensure the best environment for clear communication is in place. Perhaps you’re a man who’s received scattered warnings of sexist talk, but you have no idea what it is you’re saying that’s giving people the idea, and you want to see evidence.
All.ai mines the rich signals in human language and brings them to the surface for the empowerment of both individuals and the groups of which they are members. The more we understand what we say, the better we communicate, and the stronger we become.
