Bob manages Accounts Payable for his Organization. The group always had an issue with the tedious work of processing Invoices.
The IT group after all the cajoling had come up with an Oracle custom solution to reduce the work. It was an incomplete solution and also failed all the time.
Seeing all this trouble, Lin the tech-savvy intern in Bob’s group downloaded an RPA solution and demoed what can be done. Bob was excited and showed it to his boss who was even more excited.
They reach out to an RPA vendor who is happy to put in a place a quick solution within a month. Bob is a hero. No one knows about Lin though he did everything 🙂
Bob’s peers are envious. Some are scared that this automation is going to take away their jobs. They come up with an idea to use Oracle upgrade to stonewall any rollout for their group.
And the upgrade did come! One screen changed drastically – the whole automation is failing. Lin is on the phone with the vendor who is having a tough time fixing this as he had moved on to a dozen other projects that he now has.
Bob is now a Zero and so is the Organization.
In hindsight, Bob’s boss and the management thinks a CoE would have helped.
What is an RPA CoE?
An Automation/RPA Centre of Excellence (CoE) is generally a unit within the organization that provideshigh-quality Automation services to meet the enterprise’s strategic objectives.
The unit provides a centralized structure for all aspects of automation including resources, messaging, assets and security standards, and best practices. Quality resources and messaging would have helped Bob roll-out and manage his automation more smoothly.
Why do you need an Automation/RPA CoE?
The Centralized structure that CoE provides creates a framework for smooth roll-out of RPA and other automation technologies. It enables:
Robotic Strategy & leadership – ensures maximum value is delivered
Demand management – Process identification and prioritization
Change Management – Involve key stakeholders like IT, Compliance, Security, Procurement etc.
RPA Tool management – Efficient use of automation vendors, Tools, and Licenses
RPA Capability development – A pool of quality resources across all phases
Consistent Implementation – Standards for Design, Implementation & Support
Organizational IP – Reuseofrobotassetsandcode
Benefits Tracking – Establish, measure and report metrics
So an RPA CoE enables your automation journey right from Strategy to reporting.
Most organizations tend to try out RPA and evaluate it’s potential before rolling out a CoE. It is wise to lay the groundwork for your CoE as you do this though. You can then grow it out as you progress on the journey. For eg. You can define your Strategy ahead and have a few team members focussed on the key aspects like Change Management, Tool Management and capturing Best practices during the initial implementation.
That was a quick look at What and Why of RPA CoE. We will look at the structure of typical CoEs, when and how to set it up in a future post.
The best approach to identify the processes for automation is to use a funnel concept.
The objective is to first capture most of the processes in the organization across business lines. We then filter them out till we have processes with maximum automation potential and returns maximum benefits.
I suggest taking a 3-step approach to identify the processes. Once the processes are identified, we can prioritize and build them out.
The suggested three steps to identify the best opportunities for automation include functional analysis followed by process analysis and then finally filtering them for the benefits. Let us Look at each step in detail.
The first step is to uncover all possible automation opportunities within the organization at a very high level. We can go about identifying opportunities across the organization in multiple ways – Lines of business, geography or business function. The approach we have taken in the past is to go at it based on the business functions. So let us take that as an example approach here.
For each area (Business function here), it is advisable to identify a champion to help us coordinate with the group and their sub-groups. I have seen it it’s best to start by improving the digital or bot fluency of the different groups. This can be achieved by setting up sessions with the group through the champion.
The sessions are then followed by questionnaires to the group to identify the processes and provide data on the same. Once the information is received from each of the groups, you can do a first level analysis for automation. At this stage, we are largely looking for processes that are largely manual and have high volumes. Based on this high-level analysis and based on past experience on RPA, flag the automation opportunities within each of the business function.
You can then come up with a heat map like the one below:
As part of functional analysis, we took a first-level look at the automation opportunities throughout the organization. We now have a prioritized list of Business functions that we like to start with.
Now in the next step, we will dive another level deeper into the processes for the business functions that we prioritized. For eg., if we prioritized Finance as an area to automate, we will analyze the Finance and Accounting processes further.
This next level of assessment is usually conducted through workshops with the sub-groups of the identified business functions. During these workshops, the identified processes are analyzed for the following:
Along with the group champion, assess which of the individual processes for the function are ripe for automation. These processes are usually assessed for:
Manual tasks where considerable time is spent on repetitive tasks
Process maturity – Processes that are clear and is usually well documented
Manual hand-offs: Processes where information is entered repeatedly in the same or disparate systems
Processes that involve searching, collating or updating information
Processes that require the matching or comparing of data
This should give you a list of processes that are suitable for automation.
Ease of Implementation
Next, look for the ease of implementing the automation based on the technologies that we are looking to use to automate. All the tools have areas that they are strong at and areas they are not so good at. For eg. If the process has handwritten documents, this would be hard to implement automation using RPA.
For understanding ease of implementation, look for:
Number of Exceptions. Processes with lesser exception paths obviously are easier
Level of Digitization. More the digitization, easier it is to implement
Security needs. More the levels of security, the harder it could be
Compliance. Too many compliance requirements can make it harder to automate
Process stability. It is better to automate when the process is more stable
Change readiness. If the team is amenable to change, it is easier
At the end of the above two steps of process analysis, we have a list of automatable processes for the business functions we like to target. These are generally the processes that we now look forward to automating and moving to production.
Before that though, we should ideally be ensuring that the processes meet our benefit objectives.
In the final stage of Process identification, we want to filter the processes based on the likelihood of meeting your strategic objectives.
Some of the common criteria to be considered for this include:
After this final identification step, generally, about 20 to 30% of the potential processes come through the filters and would be ideal automation opportunities.
You can then prioritize the processes and select the ones most suited for a pilot program and subsequent sprints.
You now take the processes that we identified and maintain a running log along with a priority to implement them. This can be tracked in a spreadsheet or any system that you and team are comfortable with.
It is best to use a template to log the processes and rank them based on a scorecard. This scorecard is used to rank, prioritize and build the processes one by one. A sample process prioritization scoreboard could look like this:
You can have more criteria like Productivity, Risk mitigation, Impact, Compliance, etc.
Process prioritization and management enables effective management of the automation pipeline. This sheet can also be used to report the progress of the automation initiative to identified stakeholders.
We now have a list of prioritized processes or use cases to build our bots from.
We now have enough data to summarize what you should consider before adopting RPA. This is thanks to the early adopters and the trials and tribulations they have gone through. If you are one of the organizations coming onboard, you can take advantage of the learnings so far.
It looks like RPA will reach near universal adoption in the next few years. More than half of the respondents(53%) have already started their RPA journey as per a Deloitte study. About 75% of the Organizations would have some form of RPA implemented in the next 2 years.
Implementing RPA brings significant benefits. RPA continues to meet and exceed expectations across various key areas. This includes improvement in Quality, Speed, Compliance, Productivity, and cost.
While there is a lot of success around RPA, only 3% of the Organizations scale up as per the study! As per one Mc Kinsey study, “it is clear that the first act in the ‘robotics evolution’ has not been a slam dunk for many, especially when companies try to scale”.
How do we scale our future implementations? What are our learnings so far? It turns out most of the failings can be traced back to a few things we could have considered as we started on the journey. Here are 6 top considerations.
1. Strategic priority
It is important to have sponsorship and executive buy-in on the RPA journey and Roadmap. All the stakeholders should jointly own the results. This is easier if this is a strategic direction and everyone is rowing in the same direction.
This is easier said than done though with the competing objectives of different stakeholders. It is a balancing act at the end of the day but is a very tough act if automation is not a priority for the organization.
I like to say that there should be a “Pull” from each of the business groups involved rather than being a “Push”. It’s much easier that way.
2. Business Lead
Intelligent Automation(IA) or RPA is a Business tool and not an IT solution. So, RPA projects should not be considered as another technology implementation. IT has to be involved from the beginning but in a slightly different way. IT has to ensure proper technical governance (eg. Hosting, Security etc.) and let Business lead the automation of processes.
This is a major mindset shift and probably the most important one for successful RPA implementations. Most of the initiatives on RPA either do not take off or ends up in failure when RPA is considered an IT tool and goes through the usual Tool evaluation and implementation process. It is not easy but worth the time to have discussions and align everyone with the notion that business would lead the process automation.
3. Manage change
There is a lot of hysteria around job losses due to Automation. So, it becomes especially important to plan and manage the communications. I would start by defining the roles of the stakeholders clearly. Have a plan that involves supporting teams like IT, Audit/Compliance and Procurement as early as possible.
As with any initiative, there would be resistance from various levels and groups. It is important to understand the concern and address at each level and early. We have seen resistance from mid-level managers mostly as well as people working on the tasks. It is important to educate and communicate to raise awareness of automation and its benefits. This can help improve buy-in on automation.
4. Right Process
It is important to select the right processes especially at the beginning and show a quick return on investment. People have found it best to start with simple processes that are visible but not very critical to the working of the group. It is better to select processes with a few exceptions paths as possible to start with. Studies also suggest that you keep away from processes that have a lot of Audit involvement in the beginning lest we end up with questions you are not ready to answer yet.
Selecting the wrong initial processes could slow down or seven stop the Automation initiative. Choosing the wrong pilot process has been one of the major reasons for failed Automation initiatives. Now that we have a database of processes that are fit for RPA, It is advisable to choose the initial processes with a strong fit.
A planned approach to implementation ensures that you have the right foundations to scale the automation initiative. So, choose the right tool and ecosystem for the specific use case(s). Take care of your security considerations right from the beginning.
In terms of implementations, choose the methodology that works best for you. Agile methodologies are usually better for quick wins. Finally, establish Governance for managing Intelligent Automation throughout the life cycle.
A well thought out launch with detailed IT controls, Using the right tools and Business Processes is a hallmark of successful implementations.
6. Track metrics
It is helpful to identify the relevant metrics and track them on a dashboard. Agree with the business teams on the success criteria by identifying the post-automation performance metric and how it would be measured. It is useful to measure and track the cumulative Savings and ROI of the automation. Successful initiatives have the right level of reporting and leaders that understand the ROI & TCO of the automation projects.
In conclusion, while the above points summarize the common considerations, each organization, team, and implementation is unique. You learn a lot by doing and starting out. It will take a few cycles before you discover your unique challenges but these considerations based on past data is a good start.