We’ve had robots for ages!
Put simply, Robotic Process Automation (RPA)is a type of software that provides ‘virtual FTE’. So rather than the automation seen in industry, with rows of robotic arms welding car bodies, assembling circuit boards, or moving deliveries around a warehouse, these robots are instantiated purely in software. Like their industrial counterparts however these software robots replicate the work of people who would otherwise perform the tasks being automated.
In the RPA environment, those tasks are the operation of existing business processes and existing application software in order to achieve a business outcome.
Allianz provide an account creation example where a manual ‘as-is’ process (figure 1) is automated using RPA (figure 2).
RPA focuses on automating business processes with software ‘robots’ mimicking exactly what a human employee would do (e.g. by logging into a system, entering data, copying and pasting data between systems, etc.) in the business...
What are the benefits?
In its explicit focus on business process and the drive to analyse, model and automate those processes, RPA links to a focus on Business Process Management and that area’s focus on process as a strategic business asset. There are several clear business benefits specifically provided by RPA:
robots work 24X7
improved responsiveness (robots are faster than humans and can operate at all hours)
highly scalable allowing addition of more robots during peak hours
higher accuracy at lower cost
full audit trails for processes help achieve compliance and reduces business risk.
licenses for robots are a fraction of the price of employing someone
frees staff to work on more value added activity
Moreover, in contrast to business change programmes which require traditional system integration projects, RPA provides a comparatively straightforward route to automating business processes across multiple functions and which use disparate existing systems. RPA:
executes existing business process with existing applications
robots work with existing IT architecture
infrastructure sits with existing IT infrastructure and is governed by IT
robots can however be controlled by suitably trained business teams
So RPA provides a neat way to drive automation in business processes. RPA delivers several accuracy and cost benefits and should cost less than a change programme based on integration of legacy systems.
Bring on the robots!
The use-cases for RPA are unarguably extensive (see table 1).
Finance and Accounting
Sourcing and Procurement
Regulatory and Compliance
Financial Risk Management
Cyber Risk and Resilience
Manage incoming vendor invoices
Process vendor payments
Validate sales orders
Create and distribute customer invoices
Process customer payments
Maintain master data
Audit expense reports
Systematic vetting of contract details Extraction of key data and metadata
Automatic evaluation of orders and routing to appropriate teams for exception resolution
Data entry validation for tax and licencing forms
Periodic attestations renewals
Review of waiver requests
Reconciliation of trade requests with pre-authorisations
Processing of new affiliation and interest disclosures
Registration of gifts and entertainment in central register
Evaluate credit limits
Aggregate and segment data for standard reports
Complete timeliness, accuracy and completeness checks and remediate where needed
Examine system logs to identify suspicious or illegal activity
Check customer documentation and enter in to account opening systems
Compile information into standard reporting formats
Reconcile recorded data across operational and risk systems
With benefits attributable to RPA ranging from the possibility of 24*7 transaction processing capability, to quoted near 99% accuracy, and over 60% reduction in turnaround times, the technology undoubtedly has merit.
The attractiveness of RPA is reinforced by existence of several specialist vendors in the market, and the involvement of the large technology companies - by purchase of build-from-scratch. Popular vendors include: UiPath; BluePrism; Automation Anywhere; NICE; and Work Fusion. As a technology RPA is also comparatively mature with numerous specialist consultancies, or consulting practices within larger firms, supporting implementation of the tools.
RPA’s is attractive in financial services because of the weight of legacy systems in use, all the business processes built around them, and the cost and execution complexity of replacing them.
Hitting the change wall.
With all the benefits RPA does still have limitations, notably with handling of unstructured data, and in situations where judgement is needed regarding what action should be taken in a process. In both situations AI can help. Processing unstructured data is a task for which AI has found inumerable recent benefit, and automation of decisions requiring judgement is a task where AI techniques apply. Unsurprisingly, RPA vendors are increasingly incorporating AI into their products and cognitive process automation is marketed as a next step for the technology.
Without delving into the strengths that cognitive automation can bring it’s worth highlighting that even enhanced with AI, RPA is limited as the means to transform a business operating model. RPA’s limitation is closely tied to its strength, specifically, it focuses on automating existing process. Inherently is that existing process which is the foundation. Existing process is essentially to be maintained with the injection of automation in to the process where the technology allows. The result is a faster, more accurate and cheaper version of the existing process.
It is worth restating that this can undoubtedly be beneficial! However by itself it can also be limiting - because it misses examination of how an existing model or existing process might be completely reenvisioned.
AI provides tools with which an operating model can be completely restructured. For example:
rather than automate the flow of identification, data extraction and verification checks for an identity verification, AI could be used to extract and verify the data
rather than automate the workflow for classifying and judging insurance claim evidence, and calculating compensation, AI could be used to automatically determine the claim
rather than automate the process through which processes flow in a large legal document review, AI could be used to deliver the heart of the process
rather than automate the evidence collection and assessment process for a tax or pension calculation, AI could be used to automatically determine liability or entitlement
So while RPA by itself can tinker at the edges of transformation, understanding the possibilities offered by AI tools opens possibilities for truly restructuring business delivery.
Surfing the AI wave.
Truly taking advantage of AI is not a ‘standard’ business change effort. Amongst all the things to address is simply the understanding of AI, its benefits, strengths and limitations, by senior management, through business change teams, and in to IT. There is misunderstanding of what AI is an what it allows.
In this context we advocate AI focused education from the Board down and this issue’s article (and discussion in previous issues) about Board focus and practical methods for effectively engaging with Boards elaborate on this. Successful delivery depends on senior sponsorship of strategic change. Senior sponsorship should drive buy-in for the strategic re-examination of business models and processes, and a structured examination of the best way to deliver those either through existing vendors, through building internal AI capability, or both.