This article continues a series that explores some of the novel tension RPA can introduce between the business and IT. If you’re unfamiliar with such tension, that’s a lucky thing. You may want to read on regardless. If these issues haven’t yet manifested themselves, there’s a good chance they will.  

The business sees RPA as a low-risk, high-value new disruptive technology. It empowers them to speed up repetitive tasks and set free creative labor. This technology encourages them to design and push forward their own automation initiatives, perhaps without the “permission” of IT. 

IT might agree that a UI-based technology is a relatively low risk, but it represents technology nonetheless. Technology must be managed and governed. IT exists to do just that.

If you’ve started your RPA journey, have a few processes in production, and are still contending with these issues on a regular basis, it’s a good time to consider your organization’s automation Center of Excellence (CoE). Your CoE implements the Automation Operating Model unique to your organization. By governing how your organization innovates and operates RPA, a good CoE mitigates much of the aforementioned tension and can actually be a catalyst for collaboration.

To illustrate the value of a mature CoE, consider the tension between business, IT, and Legal in this simple use-case. 

Business customers submit credit applications via a form. For each application, someone must check for completeness and pull a business credit report from one of the bureaus. Based on the rating, the application is then routed to one of three Credit Analysts for a decision. The organization receives 200 such applications per month, and each one currently requires someone to spend about ten minutes processing prior to the Credit Analyst reading the final package to make their decision. The organization assigns a standard blended cost for the Credit department of $100 per hour. The benefit outcome of automating the receipt, credit report request, and routing of this process works out to a savings of 33 hours per month @ $100 = $3,300 plus subjective improvements such as 24*7 processing and employee satisfaction.  

Since the process uses structured data and 31-50 steps, it looks like a medium-low build with a simple payback of less than one year.  

A no-brainer, as automations go. Full speed ahead. What could go wrong?

Imagine the IT group that maintains the organization’s form repository has scheduled migration of all their customer-facing forms to an entirely new platform within five months. Were development to start immediately, it would need to be redone five months hence. 

Had they been consulted, Legal would have informed the RPA team that the organization is switching from one business credit bureau to another effective immediately due to a data breach with the current bureau. Credit will need to redesign their SOPs to understand how to apply the similar but different rating scales of the other bureau.  

Even if the RPA team clears those two hurdles, they’d better check with the Enterprise Application team. They will not do so. Why would they?  At no point in this process does the Credit Analyst or Bot deal with the ERP or CRM systems. Had they a venue to do so, the RPA team would have learned that the CRM system roadmap shows it taking advantage of a new feature within three months – automatic pulls and storage of Credit Report data. There is no need for RPA to make an external call because the Credit Report data will already exist within the prospect’s record the moment their FEIN is entered into the CRM. That’s prior to any Credit Application. If the automation proceeds without regard to this, the C-Suite will not be pleased to discover the automation pulling redundant data from the bureau, having wasted development and testing resources. Worse, the bureau charges per pull, so the organization is now paying the same bureau twice for the same data.

The CoE resolves such tensions by orchestrating what might be disparate voices within the organization (Credit, IT, Legal, Security). What harmonies these voices produce depends largely on how mature the CoE has grown. At their origin, the RPA initiatives may be decentralized as each business unit explores value propositions in something like a silo.  Even at this stage, those in charge of understanding the technology and evaluating the use cases might be in some position to consult other groups to avoid the aforementioned pitfalls. This becomes easier as the CoE itself grows in coverage and becomes a centralized body across the organization.  With this scale tends to come increased participation from those groups whose perspectives might influence an otherwise attractive automation. As this scale increases, a mature CoE operates in a Federated model. In this paradigm, the business units are empowered on their own to implement use cases when following the SOP’s and benefiting from the support and governance of the CoE.   These SOP’s should include collaboration with a diverse steering committee in which all aforementioned voices reach the right ears.

 By providing a voice to all those holding a stake, a good CoE tends to bring such tensions into the daylight where parties can reconcile them.