How to deliver large scale process automation
Virtual Operations' executive chairman and AIIA advisory board member, Nick Andrews, outlines his organization’s approach to optimizing automation programs and discusses the key enablers for achieving large scale benefits through automation
While there are considerable benefits to be gained from a well-structured robotic process automation (RPA) transformation program, equally we are seeing a growing number of businesses that fail to reap the full benefits because the corporate objectives and the RPA strategy are not aligned.
At the recent, highly-acclaimed HfS Research FORA (Future of Operations in the Robotic Age) Council meeting in Chicago, there was general acceptance that the era of deploying a single RPA tool aimed primarily at reducing process FTEs (full-time employees) is ending.
Service and software providers have realized that deploying RPA alone cannot deliver the benefits, the scale or cost reductions promised by the industry hype. The industry has concluded that the way forward involves the integration of multiple tools and a more holistic approach.
It is certainly our long-held belief that the path to process excellence and sustainable benefits starts with the adoption of a strategic automation plan.
For the vast majority of those seeking automation at scale it is simply not possible to achieve this through the tactical deployment of RPA alone. This is partly due to the fact that the numbers are not there and partly because of the limitations of RPA tools in their current form.
Most organizations’ processes are disparate, involving paper-based processes and procedures and other analogue sources that are followed to varying degrees. RPA alone cannot handle inefficient work practices.
Similarly, even the largest organizations only have a small number of staff whose work could be automated easily. In the majority of cases, the ‘low-hanging fruit’ solutions have already been implemented.
We are delighted with the new focus, not least because we, as a company, have been coaching, selling and delivering the next wave of automation (strategic—transformational) for over five years.
We still strongly encourage tactical automation projects as an introductory catalyst because it provides proof of concept and reassures the client as they see the benefits from small steps and gain confidence in the tools. This paves the way for rapid scaling-up and generates momentum for the more involved strategic automation projects.
Key enablers to derive optimal benefit from automation
Here are some experience-based observations from the work we have done:
1. Clear objectives
It is vital that the business is clear on its corporate objectives as well as the key change drivers, which vary by industry and by business, including: compliance and regulatory confidence, customer satisfaction and retention, competitive advantage, supply chain optimization, resource augmentation, management information and analytics, launching new business lines, re-work elimination, location optimization—repatriation of off-shore work, reducing /eliminating fraud, generating cash—increasing cash-flow.
We estimate that in over 90 per cent of all automations currently underway worldwide, headcount reduction is nearly always the primary and only goal.
In our view, cutting the people cost out of a business offers a quick win for providers and is often the focus of their marketing. However, as we have found, simply targeting head count reduction is limited and there are far greater rewards in moving away from the process-by-process approach and deploying automation at the cause of business problems—not at the symptoms.
We are frequently invited to participate in requests for information (RFIs) or proposals. The purpose of these is typically to identify appropriate RPA tool(s) for the client and this is a bit like choosing a car before you know its purpose, how much you want to spend, how long you want to keep it and whether you will need to upgrade it at a later stage.
It is far better to understand what you are looking to achieve overall through automation in terms of objectives and scale before you choose your technology.
If the business objectives are unclear and the processes that are candidates for automation are vague, then your choice of tools will be less effective and will commit the organization to unnecessary costs for tools that may or may not have the correct functionality that the business requires. (Many organizations are in this situation).
2. Making the business case
Building a business case and developing a successful transformation program requires the organization to build the momentum, control, standardization and deployment of best practice. The elements required are:
- Early successes: Choose your (tactical) pilot processes carefully. Then choose your core RPA product. The next step is to broadcast and deliver the resulting benefits, both direct and indirect. (Note: a surprisingly high percentage of automation benefits are never actually tracked nor delivered).
- Build your automation plan: Select the required team depending on timescales for delivery. It is crucial that the team are trained in their roles and our advice and approach is to help our clients to build a COE (Center of Excellence) largely comprised of their own people. It is also highly beneficial to dedicate people to the automation team and prioritise all participants’ automation activities over business as usual.
- Train the team to use best practice: At Virtual Operations, we deploy our comprehensive ‘VOLT’ Methodology. This ensures that all of the team uses the best practice tools and templates, all following the same procedures, which reduces risk and costs, eliminates failure and facilitates Q&A. Supervisors can far more rapidly resolve issues if they understand how the process automations have been designed and built, and recognize the areas of non-compliance.
3. In-house capability
Once the corporate objectives are clear, the business needs to review its in-house capability to procure, install and support the appropriate RPA tools.
The few organizations that do have the tactical automation scale will almost certainly need to use optical character recognition (OCR) or machine learning (often mislabeled ‘cognitive’) tools to translate the inputs or extract data in a manner that the RPA tool can handle (i.e. structured and digital). This requires both technical and commercial integration and is not as simple as it may first appear.
Our suggestion is to learn from, and work with, one the few organizations that has successfully integrated multiple tools onto a single platform or that can deploy multiple tools from a single platform.
4. Strategic automation
This is usually the point where our clients invite us to help them solve business problems or realize objectives by deploying sophisticated automation techniques within broader initiatives such as process excellence.
For example, we have looked at fraud levels, customer retention, reducing shipping container time at sea, inventory levels, regulatory cash reserve levels, acquisitions/divestments and much more.
It is vital to think of strategic automation as only one part of a transformation toolset rather than stand-alone solution. It is equally important not to focus on the ‘low-hanging fruit’ approach of headcount reduction or see it as an end in itself.
In one project we recently completed, it may have been possible to automate the work of up to 7 of a team of 14 involved in chasing unpaid or partly paid invoices. By applying a strategic automation approach to the business problem, rather than the usual tactical solution at the people involved in it, we were able to eliminate the team doing the invoice chasing AND dramatically reduce the amounts withheld from invoices. This was a genuine win-win solution where our approach helped solve the business problem rather just automate the work of those dealing with it—an example of attacking the cause rather than alleviating the symptoms.
5. Scaling up
The next consideration is how you will build and scale the solution across the business. Although most of the tools that we deal with are scalable, the real challenge with achieving rapid scale is building a competent team.
The automation market is currently seriously under-resourced and likely to remain like this for at least a year or two. Therefore, the only options available to companies at present are:
- Outsource to a ‘scale’ partner: The trouble with this approach is, however many resources the partner may have, they tend to be very inexperienced and the providers are also under-resourced so you can end up paying higher rates for recently accredited ‘green beans’ who are inexperienced with the providers and unable to provide sufficient experienced staff to provide mentoring and supervision. This has resulted in the automation industry failing to deliver in around 50 per cent of cases.
- Hire or recruit internally and provide training: This will deliver in-house ‘green beans,’ but the same problem results in that they will only learn the basics and will not be adequately competent for up to six months or so. It will take many years until they then become expert in the field.
- Build ‘joint agility teams’: This is where your own people are trained and mentored by a specialist company until such time as you become almost self-sufficient. It is really important that both training and mentoring is of high quality (i.e. conducted by highly experienced practitioners).
To illustrate the point, here are a couple of examples of strategic automation where the overall cost benefits realized for these clients far exceeded what could have been expected through tactical automation alone:
In the early days of automation, a leading bank automated the processes involved in dealing with a regulatory issue which stemmed from over-selling of loan insurance. This was an industry-wide issue and most of the bank’s competitors had to hire in large numbers of expensive consulting staff for sustained periods.
The outcome was that the bank with automated processes had far higher compliance levels, the cost was massively reduced and each new regulatory requirement could be met by a simple adjustment to the process maps and reporting algorithms which had been installed at the outset. Improved regulator confidence and buy-in was also an important benefit.
Consider organizations paying for healthcare. The sheer volume of invoices being processed makes any meaningful claim validation practically impossible and unaffordable using conventional means to administer. The best that the payer can do is sample, assess and reject claims that do not pass some fairly simple guidelines.
Historically, this had been conducted on a relatively small scale and was largely ineffective as no real management information was gathered and the manual assessment ‘work’ was done by the payer. Simple rules or parameters were applied to most categories of treatment, meaning that any invoices that fell outside of these rules were treated as exceptions (e.g. where clinical justification was required).
If practically ALL invoices were validated, then it was a fair assumption that the costs of treatment would go down substantially as the providers of primary care would be required to stay within the guidelines and would have to use expensive clinical staff to justify exceptions.
All this could be done, very cost-effectively by robots. If you then applied analytics (AI or specialist analytic tools) and compared primary care providers’ average costs for similar treatments, then an intelligent system could enable a competitive environment which would drive down national primary care costs even further.
Nick Andrews is executive chairman and founder at Virtual Operations, and a member of the AIIA Advisory Board.
Want more? Become a member of the AI & Intelligent Automation Network for free access to industry news, whitepapers and articles. Sign up today.