Agile RPA

During my RPA journey, I have encountered different organizations and projects which over promise and under delivers to their clients, causing deception on the RPA technology, and it is ok to feel this way due to inefficient projects. But why can these projects fail?

Well the answer is simple, wrong project management skills combined with a waterfall vision of the solution. But let`s make it more explicit.

Water fall projects refer mainly to get all the requirements at the start, and work on them until a project is delivered.

But what is wrong on this approach?

Well, let`s start by the end, the product delivered. We might deliver the best product ever created, but if it takes a lot of time to be produced, for sure, something simpler will have already taken the position on the market, and if it is an improvement, probably the process would have changed.

And what about automation?

Well, if it is for automation, delivering a robot for an entire process, not only will take months, resources and time, but as well, the risk of doing something useless increases, and I mean “useless”, because the process might change, and the bot might not be required any more.

Does it really happen?

Of course it does, I usually encounter this on the companies, I have seen people believing on their automation, even if it takes 2 years to complete (real case). But at the end, the process was changed due to the new technology and the RPA solution became “useless”.

What do I propose?

Well, lets start by how we work on the scope of the project. We should do it some sort of waterfall, lets construct what we want, I mean, let’s stablish the base of the result, it might change, of course it might, but at least we know where we go. It is like a soccer game, you know you score at the opponent’s side, how you arrive, changes every time.

Once we have a base, we do a process map of the system, and split the processes in stories as if we were doing “Scrum”, so we program the bots from the most valuable one, and each small automation is delivered to the operation

But how to select the most important? Well, now we play lean manufacturing. Why? Because we need to identify the bottle neck of the process and start with that small portion of the automation. 

So, at the end you mean, to do RCA, but splitting as Scrum, then ordering through lean, while planning as Waterfall?

Yes, of course that’s the approach, at the end if the bot takes years of developing, every month we will deliver a small part that can be used, and this small part belongs to the bottle neck the usually stops the process velocity. 

About the author. 

Andrey Mendez + 7 years experience in quality and Project management. Certified as: Quality Engineer, 6 sigma Black Belt, PMP, PRINCE2, ITIL & ScrumMaster/Product Owner.