The trend toward automation is not new. The Industrial Revolution started it in the 19th century, but there has never been such rapid automation progress as today. All forms have accelerated, often without understanding their effect. Humans have become choke points in operations, points of disease and legal liability, and friction to smooth digital pathways, and the pandemic acted as the tipping point for remote work, which will continue, with automation the key enabler.

The pandemic, the surge in demand, and ongoing worker shortages all accelerated investments in tech and automation. In fact, 43% of businesses said they expect to reduce their workforces through new uses of technology, according to a global survey of nearly 300 companies by the World Economic Forum. The automation imperative is here to stay. The technology itself is neither good nor bad, but all comes down to how it is applied, in which contexts, and if there is a human-centered approach to design and implementation.

What do we mean by random acts of automation?

“Lacking a definite plan, purpose, or pattern” is how Merriam-Webster defines random. What does this look like when it comes to automation?

A random act of automation might be:

  • Automating a bad process (e.g., generating bad loans faster).
  • Faulty, biased algorithms.
  • Physical harm or safety risk posed to humans.
  • Creepy experiences using robotics.
  • Awkward and frustrating self-service (e.g., bad chatbots).
  • A negative impact on an employee’s mental state (e.g., fear of job loss).

Take, for example, the case of the security robot that accidentally fell into a fountain in Washington, DC in 2017. Steve, as the robot came to be known, patrolled the office complex area with the mission of picking up any misbehavior or parking violations. When it came to stairs, however, he fared less well and was found in a shallow fountain. One passerby notably tweeted, “We were promised flying cars, instead we got suicidal robots.” In the aftermath, a memorial was set up for Steve, the drowned security robot, on his charging pad. And while instances like these may be comical, there are other examples of random acts of automation that have much more serious consequences and can inflict real damage (physical, reputational, monetary, psychological). Take, for instance, the $9 billion dollar mistake made by Zillow’s AI predictions or the 47 false arrests that were made as a result of a system failure at car rental company Hertz.

In our report, we explore the 10 most common automation pitfalls and how to avoid them to reduce your risk of random acts. We highlight 10 factors across three categories — process characteristics, employee experience, and customer experience — that, when viewed collectively, help an enterprise avoid problematic and sometimes dangerous automation outcomes.

In addition to reading this report, use Forrester’s Automation Decision Tool as a way to gauge your readiness for automation and evaluate the “random act” factors before moving ahead with automation.

To read more, check out our recent report, Random Acts Of Automation: 10 Pitfalls To Avoid. If you want to tell us your story, share feedback, or have questions, please reach out to VP, Principal Analyst Craig Le Clair ( and Researcher Renee Taylor-Huot ( Forrester clients can easily schedule time with us to discuss.