Implementing automation? Plan for the long-term
It is now widely understood that automation can offer huge benefits to the business world. And yet, for every story of a company successfully leveraging automation to streamline its operations and boost productivity, there seems to be another example where it has failed to deliver on initial hopes.
In 2022, I had the privilege of chairing an OpenOcean roundtable on automation with a robust line-up of leading thinkers and business leaders. One of the central takeaways from the event was that stories of automation and AI’s failure are very often the result of the same mistakes in implementation being made repeatedly. However, there are a few key areas that organizations need to focus on to ensure they can avoid these common pitfalls and introduce automation successfully.
Make a plan aimed at specific capabilities
Like any other business strategy or tool, automation will not lead to any significant benefits if it is not applied deliberately and precisely.
A company should only even think about automating once it has identified a specific area that needs improvement and metrics by which this improvement can be measured. By doing so, businesses will be able to track how well their approach is working throughout its implementation and enable them to adjust the solution as needed based on concrete feedback. It is worth noting as well that automation and AI rely on a nuanced understanding of ROI that emphasizes experimentation and embracing the benefits of change.
Fundamentally, an effective strategy for automation rests on first building a detailed, end-to-end understanding of the process, known as 'process mining'. A Forrester report found that 61 percent of executive decision-makers are using or considering process mining to simplify their operations. These insights provide the necessary foundations to move from siloed, specific task automation to more holistic process automation, and make a more tangible impact.
Update infrastructure and processes
Another widespread mistake is leaving pre-automation infrastructure largely unchanged. Some companies expect to see success by attempting to add AI or other automated elements on top of what they had previously been doing, but this reflects a fundamental misunderstanding of how these elements work and what they can offer. Instead, companies implementing AI and other automation solutions should be willing to reassess their processes at every level.
Structures that may have functioned acceptably with less advanced systems are often not capable of supporting effective AI. Often, for example, different divisions of a company might use different systems and services that do not interact with one another, creating a constant drain on productivity -- the opposite of what AI and automation aim to deliver.
To realize the AI-driven orchestration of work, we need to break down silos between applications. The focus should be on an integrated enterprise and truly holistic orchestration of tasks, processes and resources. To achieve this, it will require artificial integration to ensure optimal orchestration. The result could be integrating these capabilities into a single end-to-end platform or through what Gartner calls a composable platform architecture approach.
Have a long-term vision
An idea that came up several times in the roundtable was that long-term thinking is absolutely essential to successful automation. There are two distinct senses in which this is true. The first is that organizations trying to implement AI cannot be afraid of some degree of short-term failure -- there is no way of avoiding the fact that some parts of the plan will not lead to the exact desired result.
This is an inevitable part of the process, and the companies that implement AI most effectively will be the ones that embrace it in the spirit of experimentation. For example, one intelligent approach is to use lots of small-scale implementations that will very likely fail, and then use the lessons from these to inform a larger-scale implementation.
Long-term thinking is also essential because the technologies used for automation evolve at such a rapid pace. They can quickly go from cutting-edge to outdated, and unless this is planned for, even a company which has some initial success will struggle as time goes on.
This means that immediate profitability cannot be the only concern. The companies with the best chance of having sustained success with automation will be those that adopt a longer-term perspective by investing heavily in innovation and are willing to constantly update and question the platforms that they use.
Tom Henriksson is General Partner at OpenOcean.