Abstract:

Microsoft is at the epicenter of the AI evolution and has made a commitment to ensure it is used ethically and responsibly. The Microsoft Research and Insights team shares this priority as we add AI to our research toolkit. Like in the Greek myths, AI is a multifaceted entity that can blur the line between hero and villain. Created with good intentions, if misused, AI can lead to conflict or even peril. The use of AI comes with ethical considerations that researchers must keep in mind as it pertains to data bias, privacy, transparency and security. However, AI is also helping us to glean deeper insights in a fraction of the time, allowing for quicker and better-informed decision-making. With a company-wide and organisational push to experiment with AI, we set out on our journey to demystify its use and see if AI could improve our data and insights.

Our story rests in a classic tale of organisational titans?marketing versus engineering?locked in battle over ?who knows the customer best?. The stakes are amplified by the central character, software developers?a critically important audience to Microsoft and one that is notoriously difficult to research. This is due to developers being a niche audience with specialised skills and interests, making it hard to find a large enough sample size. Other factors include survey fatigue among the developers who are active in online communities or forums, their highly busy schedules, privacy concerns around what they may share to a potential third-party survey and technical knowledge requirements among researchers to pose the right questions. We sought to find a better way to engage developers and improve the quality of their responses while also striving to establish a better way to do pricing research?slaying two beasts as it were.