1. When Everything Seems too Easy
My first impression when I first hands on a RPA Tool, everything seems so easy because of the drag and drop features, pre-built functions, puzzle like building blocks that allow you to construct a sequence of actions just at your fingertips. The creator of these tools made it so easy to implement and deploy automation, this causes a lot of headaches and technical debts down the road if we do not do our homework to have a full understanding of what we are trying to automate and what control are needed in-place.
2. Rush into RPA
Due to the hype related to RPA technologies. When we first started to adopt RPA in our company oftentimes we made a huge mistake calculating ROI by just using 'Current Time Spent' minus 'Future Time Spent' without really think and consider about infrastructure and cost of maintenance. This often due to our team rushing to build up momentum, wanted to prove the capabilities of the tool and most importantly to get buy in from upper management. But realistically most of the current processes is not ready, ended up the team is spending more time maintaining the bot than the manual process.
3. No everyone can do it well
Too naive to think that everyone can use the tool that have been offered. Business user often been asked to attend RPA course and started to automate some of their task. Unfortunately, most of these automation developed were not properly tested, and assume everything is been taken care automatically. We normally started off working on the "Happy Path" without considering the scenario that are usually more complicated than that. These could be variant of an input or the input can be in a differnet format that you expect. RPA can complete your job 10x faster, but also can create mistake 10x faster.
4. Just focus on building and Robots are ready, set and forget
From day one of deploying our first bot, not a single bot can be set and forget. Bot needed to be manage operationally and maintained. We have made mistakes for not properly documenting and communicating automated process with business SMEs. When a bot breaks, no-one will know how to recover the process and leads to unnecessary over-time to figure what had happen and what needed to be done.
On the other hand, due to rapidly changes in business requirements, system upgrade and process changes have causes us lots of bot down-time. Getting involvement of IT and Business SMEs and knowing these changes before hand could reduce unneccesary over-time trying to fire fight to get your bot up and running again.
5. Failing to Scale, not building for scalability
After 3 years of RPA experience. Started to feel the impact of not building for scalability. Most of our solutions were built in silo mode rather than enterprise-wide. We are much now spending alot of time re-visiting our solution, re-designing our bots.
Another mistake which was choosing the right candidate processes for RPA, we had wrong thinking about using automation as a quick fix which now end up deploying multiple bots which are doing similar job while ideally they could be designed as a single bot.