While processing cases, Hyper Automation Bot Workers have access to case knowledge, and this can be accumulated and used in the Hyper Automation process by the same bot worker or other bot workers.
Automating to bring operating efficiency is a short-term goal and knowledge accumulation is a medium / long-term goal.
Transaction knowledge gets stored in enterprise systems and process knowledge gets mostly lost. Operating staff gain this knowledge when they process manually, and the knowledge goes with them when they leave their role.
????????? ??????????
Knowledge derived from processing knowledge is distinct from transaction level knowledge already stored in enterprise systems. This knowledge can be classified into:
1. Digital Knowledge
2. Patterns of Behavior
3. Summarized Patterns of Behavior
??????? ????????? includes special/peculiar handwriting quirks, special characters or separators used, signature stroke information (significant strokes), words generally used in written or email or speech communication, voice prints, fingerprints, facial features, etc.
???????? ?? ?????????: Cases / requests / services processed in by Bot Workers have attributes that are part and not part of the Enterprise System Database. Many of these attributes may be document scrutiny elements. Attributes which are not part of Enterprise System DB must be stored as part of the Knowledge DB. Patterns of Behavior have to be extracted periodically using unsupervised learning to extract commonly and rarely occurring patterns. Essentially, these patterns are a set of attribute values that occur together in cases being processed. Each pattern has a set of these attribute values, with allowances for fluctuations. Each Pattern also a thickness attribute associated with it. The thickness denotes the percentage of cases that have this pattern. Hence thicker patterns occur more frequently, and thinner patterns occur less frequently.
?????????? ???????? are patterns extracted from summarized / collapsed data. Data can be collapsed on multiple dimensions: like time (temporal), geographical region, by customer business segment, by currency, etc.
Patterns of behavior and Summarized Patterns have to be updated continuously as more & more cases are processed and variations in these patterns, representing evolution of behavior can be tracked.
????????? ????? Knowledge acquired can be used by Bot Workers in routine case processing. This usage can be to detect out of pattern behavior, errors in the requests, to predict customer behavior, risk bucketing of cases, etc. Knowledge can play a big role in deriving much higher value from hyper automation.
