The Automation Imperative Part Three: What will you automate?

Having established that you will automate and that you want to automate, the next question is where to start? Every company I have worked with has begun their automation journey with a set of 3-5 pilot projects to gain experience and confidence with automation technology. Here are some guidelines for selecting the right pilot processes.

• First, what measurable impact will be achieved? In order to build support for automation in your organization you will need a case study that illustrates to stakeholders the clear benefits that can be achieved through automation. So there are really two questions embedded in this first step -- how will you measure the impact in a way that people in your organization will understand and agree with the measurement methodology. And second, what level of impact would constitute a meaningful contribution that would give your stakeholders a reason to believe that automation could also help them.
• Second, avoid implementation risks! In evaluating possible first pilots, watch out for process complexity, system access restrictions, challenges in getting employee adoption, and anything else (regulatory, organizational) that may add substantial time, cost, or reviews to the implementation schedule.
• Finally, in selecting processes for your pilots, choose ones that can illustrate multiple values (not just cost avoidance). Stakeholders in your business may be more impressed with the opportunity to mitigate risk, reduce cycle time, provide new insights, or improve customer satisfaction -- these can often be more compelling than simply reducing human labor

Some questions that you should ask as you evaluate different candidates for your pilots:

What is... time-consuming, tedious, high-volume, resource-heavy? And out of those, what is mundane, does not require judgment, repetitive in steps, consistent in outcomes?

Some examples:

• Finding the right data? An auto loan company needed to use public records (available through the web) to verify that the auto dealer specified in a given loan request was a legitimate registered business. Each US state has a slightly different system and information format so the validation of this by human workers was time-consuming and tedious but very repetitive and consistent. Because this automation also did not require any integration with internal systems it was a great candidate for a first pilot.
• Following up on items? A manufacturer needed to maintain up-to-date certifications on a variety of topics for all of their suppliers. Tracking when certificates were due to expire, sending emails (and follow ups) to request updated certificates, validating the information and then updating the tracking schedule was a time-consuming and repetitive process - one that took more resources than the company was able to allocate, resulting in frequently missing and out of date certifications. Almost every step in this process was automated resulting in a substantial improvement in compliance (and measurable risk reduction).
• Creating reports? A marketing organization needed to consolidate information on online advertising activities in order to inform decisions on where to increase or decrease spending. The information had to be gathered from multiple different suppliers online systems, standardized from varying formats into a consistent set of data for comparisons to be made, and then sent to decision makers for review. Due to the level of effort required, reports were being generated only once a month with a resulting delay in changing advertising spending decisions. Automating this process released a marketing employee to do high value work but also resulted in more frequent reports and updates to spending decisions - resulting in a much higher return on invested advertising dollars.

What will you automate first?

Previously in this series:

• The Automation Imperative Part One: You will automate
• The Automation Imperative Part Two: You want to automate