Adrian Iaiza, Formerly Head of Process Automation and Improvement, TAL Australia
A brief scan of literature suggests that while most organisations have embarked on Intelligent Automation, the majority of us are struggling to scale. Why does this happen?, What is holding us back?
Goldratt’s Theory of Constraints tell us to identify the most important limiting factor standing in the way of achieving a goal, and systematically improve that constraint until it is no longer the limiting factor. Then move onto the next constraint, and so on.
As no two organisations are the same, I’ll suggest a number of common constraints to consider, leaving it up to you as to which to take up, and in what order.
C-Level support is conflicted or waning.
This is usually as a result of a lack of Value and/or Alignment.
Value: We all know to prioritise effort against value, but value is much more than cost alone. To generate and maintain enterprise wide engagement, value must extend to areas such as Customer Experience (time and quality), Risk Reduction and Avoidance, Revenue Generation, and fast tracking future new Product Releases.
Once value has been defined it needs to be benchmarked, and progress widely reported. Too many organisations define value holistically, but then fail to report more than just the number of processes automated or hours of work avoided. Value needs to be reported holistically in order to generate, and sustain enterprise wide engagement. Reporting of benefits needs to be expanded to benchmark and report on the other key holistic measures such as reduced time to market, avoided regulatory risk, and improved revenue generation and NPS.
Alignment: Hard fought budgets are tight and getting tighter. This capability is best used in areas using legacy systems that are expensive, or difficult to change and integrate. It is important that the initiative is positioned as complementing, and not competing, with straight through processing projects, using API or similar approaches.
There are also synergies to be attained, and the question to ask is “Has the opportunity been seized to align your automation initiative with your digital transformation program?” IA has a thirst for clean, structured, digital data. Don’t let this opportunity slide by.
Commonly Machine Learning and Chatbot initiatives run into difficulty when moving from decision to action.
IA is best placed to fill this need cheaply and quickly by using the digital inputs generated from machine learning, and then processing the relevant transactions across all underlying systems cheaply and effectively.
Change Management: Are the benefits aligned against your corporate objective? Is IA seen as a way to deliver these top down, and bottom up throughout the organisation? Is there a residual fear around the intent of this capability? Are all members of the CoE change agents, and have you developed a network of supporters at all levels across the organisation?
Centre of Excellence considerations.
The common themes that surface are:
A shortage of endorsed, high value processes that have been shortlisted or prepared: It can be overwhelming for a small team to spread themselves throughout a complex enterprise in an effort to understand and prepare a pipeline of opportunities. Organisations have found success in the alternative approach of the CoE developing the assessment and process preparation frameworks. These are then used by the users themselves in determining what processes, and in what order, the automation needs to be developed. The framework enables transparent and objective decisions in an unconstrained, distributed way. It also fast tracks process preparation without sacrificing quality.
Business users find managing this capability harder than expected: There should be no distinction for any operational manager in managing virtual and human agents. Management of virtual agents should be integrated into your BPM. KPI’s such as productivity, utilisation, quality, resource planning and associated metrics should be seamlessly integrated into your management operating systems.
The cost of managing a small fleet is more than anticipated: On face value, the total cost of ownership of any virtual agent should equal and ideally exceed that of any BPO offering. Often this might not be the case. Here area few things to assess:
Vendor Management. Studies indicate that 79% of organisations work with 2 or more vendors, and 55% expect this increase to 3 or more. With IA, it is better to select a single provider that can support your attended as well as your unattended use cases, and seamlessly integrate these into your BPM. Less is more.
Maintenance cost. This can present itself as the hangover from the night before. Is reuse baked into your process design as well as your technical design standards? How effectively is this managed? It is not unusual for 20+ processes to share a common screen(s), for example, Customer details search and update. If these screens change do you really want to retrofit, test, and release these changes into 20+ separate baselines?
Infrastructure. High fixed costs against current scale. Try investigating options internally or through providers that can operate and securely manage this as a service for you. You’ll be glad you did.
Lack of environment integration. The need to have integrated environments and synchronised static data is essential in order to deliver continuously with quality. This must be plaguing other development efforts, so combine forces and develop solutions.
SDLC integration. Is the resilience of the IA fleet put at risk by unforeseen releases, or do development teams resist or find it difficult integrating this capability into their projects? Ensure project funding includes provision for IA and ensure your SDLC is updated to include the needs of both IA and development teams harmoniously.
Hopefully, this article has highlighted constraints to focus on how to achieve the scale, and reap the benefits that you anticipated at the outset. IA is a maturing capability that has much to offer, as long as its managed appropriately.