Hi Anthony you've got the two most important feasibility factors - application response time and workload capacity.
Applications can't be magically turbo charged, and all we can hope for that a digital worker will be faster and more efficient than a human. But if it takes 10 seconds for an application to find a customer, then using a DW is unlikely to improve improve that time. Maybe the DW will execute the search faster than a human, but the search will still take 10 seconds. In reality I'd have thought that unless you have an application with a highly responsive front end, then UI integration may not be fast enough for IVR/chat requirements, and probably only a 'headless' integration (eg API or DB query) will have the necessary speed. But let's see what examples the community has to share.
To some extent you might be able to design a solution that does not follow the basic 'call and response' pattern, where the user sends a command and waits for confirmation that the work is complete. Perhaps instead of responding to the user with 'your account has now been closed' the digital worker could say 'your request has been received and your account will be closed shortly'. Such an approach will depend on business requirements but it may be necessary to rethink the user-robot flow, possibly even engaging more than one robot per user, in order to achieve the best solution.
Obviously you'll need a digital workforce that can maintain SLA during the peak period, which begs the questions of how many workers are enough and what does that workforce do during quiet periods, sit waiting or do something more productive? Perhaps the 'economy of scale' means that any idle time is offset by the benefits brought when the DWs are productive. It's something to weigh up as part of a project feasibility study.
------------------------------
John Carter
Professional Services
Blue Prism
------------------------------
