) almost exactly 100 years ago, robots finally became real and started to help weld rivets in automotive engineering plants somewhere around the 1970s. They then further evolved into robot-shaped humanoids with Japanese firms innovating proof of concept walking robots, but these have mostly failed to become part of our homes as yet.
Today we might think that AI is quite ubiquitous in the sense that it could be potentially applied to any aspect of business. Kelleher says that AI should not be viewed as a single application deployment point in this sense. Instead, it should be regarded as a fabric that can be applied to specifically defined solutions that span an end-to-end operating model across a modern enterprise.
Being able to bring this element of intelligence into RPA is only possible if we have an AI model that understands exceptions as they happen so that it can learn and navigate real world ambiguities more competently. This is the point where we can add “agentic skills" to RPA and start to channel that intelligence from what UiPath would call self-healing robots into dynamic planning and dynamic learning in business.