While the pace at which AI solutions are getting built has improved significantly over the last few years, the success of adoption of AI solutions is still low, for example, the self-service automation adoption in many mature client instances is still less than 15-20%, based on our experiences and also market observations. 

The critical question therefore is: How do we ensure that we are applying AI at the RIGHT place in the RIGHT process at the RIGHT time?

The challenge starts with finding the right business scenario where AI capability can be applied, which can also yield measurable outcomes. Once you cross this first big challenge, then it is about getting the buy-in from the business owners and building their trust in AI to let it take decisions. It’s important to be grounded on what AI can do, as there is a lot of hype around it and to set expectations with business leaders on timeline it takes for AI solutions to mature and start giving return on investment. There are unique challenges in AI, like getting “clean data” to train machine learning models. These are typically not given due importance in the plan, as project managers are more used to planning RPA solutions. With the AI skillsets in high demand, lack of right talent can bring down the shutters on the project. Equally important are the domain or subject matter experts. Many AI projects have failed as the domain of the end users has not been understood and contextualized well. In this document, we will look at the various challenges faced in AI Automation projects and discuss ideas on how we can prevent some of these proactively, to help enterprises adopt AI in a much bigger way.

Read more here: