The key driver of automation and digitization today is accessibility and low cost. Front-end users can drive the development, design and build automated operations that previously were out of reach for all but the complicated big IT-projects.

As bottom-up innovation tends to be driven by front-end users, there is a possibility of miscoordination and individualized solutions. Overcome this and improve the sustainability and impact of such approaches.

Here are three ideas to help manage and orchestrate the bottom-up innovation and digital automation:

1. Prioritization

This needs management input. The basic rationale for the business case is one of impact and feasibility. Impactis mostly about identifying candidates, those processes which are repeated recurrently, and consume much time and energy of the staff. Feasibilityis about understanding which ones of those candidates are actually feasible for automation.

The low hanging fruits would be those processes that have highly structured data input, in digitized form, where processing does not require cognitive input, and where outputs are also in structured form.

The difficulties here are much about disentangling various workflows and ad-hoc processes that may exist in the organization.

And slightly more complicated, the workflows may need redesign in order to be automatable. Perhaps a new workflow concept is needed. Separate out processing that is exclusively using structured data, from those that requires dealing with unstructured data and also requires cognitive input.

Don’t leave this to the coders alone.

2. Collective solutions

Coding is a very individualized activity. Managers will need to overcome this tendency by introducing some concepts and infrastructure to help ensure that solutions are collective, are accessible to those who need it, and does not disappear when the individual disappears. It shouldn’t be necessary to fully emulate the practices of tech design and development houses.

Your focus is on front-line usability and maintenance. Practical solutions for this are often about shared hard drives, and shared operational concepts for how to manage file versions, data and document management. And, if possible, ensure that there is some residual capacity in the team to understand the code so that adjustments can be made by more than one person. Document the code by commenting in then scripts so that it is easier for multiple users to engage and to maintain over time.

3. Be conversant and have some fluency with code

The world is still in a bit of a transition stage whereby too few people understand computer code. A degree of fluency is immensely helpful, and to an extent necessary.

It is just too important to be a black box. Significant decisions are made by the coders. Critical decisions are made about controls, dependencies, definitions of variables, and configuration of reports and metrices. It is not a given that the coders have the same understanding of core workflow issues, priorities and requirements. Yet, they make critical decisions that will determine performance, and insight, for years.

Coding is very hard. And being an efficient coder is even harder. Yet, all managers do not need to become fully efficient coders. But a level of fluency with core principles is necessary, and that’s goes beyond just being conversant.

Good news is that understanding the core functionalities of the workflows and information processing is the basics. A good business process manager may already be on top of that. The step-up to looking the coder over the shoulder and understand how this is turned into an automated procedure is not that far-fetched.