Using AI with Agile Coaching, Scheduling, and Status
Agile isn’t merely about delivering software faster; it’s about delivering quality software faster. There is no doubt that artificial intelligence (AI) can significantly enhance team performance regarding the collection, analysis, and presentation of data to teams for record keeping or trend analysis. This theme of indirect assistance continues with the use of AI by Agile coaches to assist with the continuous performance improvement of Agile teams. However, the people side of Agile team activities is not well suited to AI support. Teams complaining that it took longer to feed the tool with data than to undertake the scrum activities themselves supports the Agile Manifesto value of “individuals and interactions over processes and tools.”
Scheduling and Status
Robotics is a great way to automate repetitive and predictable tasks, such as the creation of status reports or the calculation of key performance indicators. Typically, these features are found in project management software used for waterfall activities. However, robotic support of this type could also be applied to Agile activities and include automation of the sprint closure report and the production of metrics, such as velocity, cadence, burndown, defect status, and so on. The use of this type of AI support reduces the burden on the team. However, it also means that a team needs to have a certain level of data accuracy to permit AI support and free its time for more productive activities.
AI support can be beneficial in sprint and release planning. In this instance, virtual agents and analytics can assist in evaluating the probability of a team’s ability to complete all work items planned for the release or sprint. This use of AI helps the team guard against being overly optimistic regarding potential performance based on recent performance. The use of AI would probably be limited to making recommendations rather than producing the final plan, which, for team commitment reasons and lack of data for accurate predictions, should be left to human intelligence.
Most organizations use Agile coaches to support their Agile teams, along with AI tools to synthesize large volumes of benchmark data. AI tools, such as Rally (formerly CA Agile Central), can make predictions and recommendations, providing insights into an Agile team’s performance when compared to benchmarks. The Agile coach may then use the analysis in the context of the team’s real situation to recommend intervention actions. The insights provided use the dimensions of responsiveness, quality, productivity, and predictability to suggest optimum sprint length, team size, and so forth. AI tools allow for comparisons between an Agile team’s performance and benchmark data so that Agile coaches can select appropriate actions. This is a bit like the real-time analysis the coaches of sports teams gain using GPS data analysis. Applications can work out metrics that help to ascertain an athlete’s exertion level or performance in any given second. When integrating this information with heart rate, coaches can assess the athlete’s fitness, stamina, and injury status. AI systems provide similar capabilities for Agile teams, and, while not directly encouraging efficiency and faster results, they equip Agile coaches with information that enables them to steer Agile teams and enhance Agile practices to achieve greater productivity.
[For more from the author on this topic, see “Can AI Improve Agile Team Performance?”]