Charles Wright

Head of Technology Solutions

  • Digital Transformation
  • Hyperautomation
  • Data
  • RPA

15 Apr, 2021

Hyper-Automation: Beyond RPA

Combining emerging technologies to drive innovation

Robotic Process Automation (RPA) is an incredibly valuable tool that can provide a great ROI if used correctly. But when simplicity does not fit the task, Hyper-automation provides a scalable solution.

What is RPA:

Robotic Process Automation features the use of digital “robots” to autonomously complete tasks that require interaction with multiple systems, commonly via their user interface.

RPA use cases:
Robotic Process Automation (“RPA”) is a well-explored, previously emerging Technology, which has been adopted by many businesses to enable their transformational journeys. As a result, use cases leveraging RPA are well understood and provide great potential for cost savings. A few classical process characteristics which will provide a scalable return on investment include:

  • High volume, repetitive
  • Currently completed manually
  • Contain structured data
  • Stable or infrequently changed
  • Simple or logical decision making
  • Integration requirements with legacy technology

This provides the opportunity for RPA tools to replace meticulous, manual processes in all or part of a function, allowing for cost reduction, reduced error rates and all the classic benefits of RPA to be realized.

The development of RPA is simpler and faster than software development because of built-in components which create a friendlier user interface that does not require end-user coding capabilities. These easy to use, minimally technical, platforms allow low-cost pilots to provide generous ROI when the correct activities are chosen.

RPA Success Factors

Many businesses fail to effectively utilise RPA due to a common set of factors that can largely be consolidated into lack of planning. These are some of the typical things to consider when looking for success in RPA:

  • Choose the right operating model – One organisation may suit an internal Centre of Excellence model while another may be better served by outsourcing
  • Choose the right tool – There are a number of RPA tools available, the right one must be carefully considered through its features, feasibility, pricing model and more
  • Get buy-in – RPA implementation can face a range of challenges and cause considerable change within an organisation, leadership buy-in and support across the organisation is essential for success
  • Gather the required data – Depth and breadth are key, look across the organisation for potential use cases and then map and quantitatively analyse them to build an accurate business case for each
  • Pilot and scale – Having identified a range of use cases, prioritising the automation of short duration, low risk and high return processes build momentum

What is Hyper-Automation?

Hyper-Automation is the next level of RPA and automation technology. Hyper-automation combines the capabilities of RPA technology with AI to allow complex decision making processes to be computerized.

In large part this capability has only been feasible over the past 5 years. Recent developments in Computer Vision and Natural Language Processing allow for advanced processing of unstructured data (text and images), allowing computers to perform semantic activities such as information extraction or summarisation.

Capabilities are rapidly increasing as research into machine learning continues, new models and algorithms are regularly released and results constantly improve. More and more commonly Machine Learning exceeds human performance across a range of tasks and applications.

Use cases for Hyper-Automation

The combination of these technologies into the automation of processes allows for numerous new use cases previously unreachable via RPA alone. A small selection of Chaucer’s key hyper-automation projects include:

  • AI augmented information extraction from invoices, emails and other unstructured data types used to populate internal systems
  • An intelligence platform which collects data from the web, applies summarisation and sentiment analysis, followed by conversion to structured data with metadata tagging
  • Translation and recreation of documents in alternative languages across multiple formats

In addition to replacing an entire process, the sections of automated processes that previously required human checkpoints or intervention increasingly require less interaction and error handling. With the strength and progress in modern AI typically we are able to provide high accuracy AI while automating entire processes, dramatically reducing the need for human contribution.

Is there a downside?

While the promise and ever improving capabilities are exciting, there are significant considerations be made for using these technologies.

Hyper-automation can be costly to build out due to the number of moving pieces. If not executed thoroughly, there can be expensive maintenance which will increase the overall development time of the product among other risks.

The largest being the new infrastructure required to support AI technologies. Computer Vision and Natural Language Processing both require specific data pipelines and GPU driven compute power generally provided in the Cloud, which incurs both a financial cost and an increase in development time. Additionally, it requires new skillsets to implement and introduces further maintenance requirements.

Takeaways

There are many methods to automate a process, it is important to perform detailed analysis and data collection before selecting the best approach.

We suggest the following steps to get started with hyper automation:

  1. Build an automation strategy which considers vision, technology, operating model, culture, readiness, and governance.
  2. Gather buy-in from leaders across the organisation and identify a wide range of automation opportunities across multiple functions.
  3. Don’t lift and shift processes and potentially the bad behaviour. Make sure you review how the process should be and have the right quality data to back it up.
  4. Implement your operating model and execute your highest value, lowest risk use cases as pilots first.
  5. Scale and industrialise your automation program, incorporating increasing numbers of use cases alongside maintenance, making sure to realign to the original vision where necessary.

At Chaucer we have developed and iterated on an automation strategy model, including a process analysis and automation framework in collaboration with multiple clients, enabling us to identify and gather all the details required to select the correct method of automation and generate detailed businesses cases in support. Our insight is built from multiple industries, leveraging both horizontal applications of technology and specific industry requirements which can benefit across the board. We support this with deep technical capabilities in AI, Cloud, RPA and software development to implement and scale any use case.

If you would like to scale, redesign or implement the first use case in your hyper-automation journey, please reach out to Chaucer’s Data & AI Centre of Excellence:

  • Elodie De Fontenay – Data & AI Insight Partner: Elodie.DeFontenay@chaucer.com
  • Charles Wright – Head of Technology Solutions: Charles.Wright@chaucer.com

Charles Wright

Head of Technology Solutions

Chaucer's AI specialist delivering data strategies and capabilities for Fortune 500 organisations. He is passionate about driving data led digital transformation to enable organisations to realise the benefits of machine learning and holds both an MBA and MA in Educational Leadership and Management.

Blog 23 Jul, 2021

Data, RPA, AI, Robotics, Machine Learning

Preparing for your first Machine Learning Use Case

23 Jul, 2021

Charles Wright

Head of Technology Solutions

Blog 23 Jul, 2021

Data Strategy, Data Science & Analytics, Data Visualisation, Data, Data Ethics

Five risks of getting ‘return to work’ wrong and how to avoid them

23 Jul, 2021

Elodie De Fontenay

Insight Partner – Data & AI

Blog 06 Jul, 2021

Privacy, Cyber Security, Data, Ransomware, Information Security

Handling a ransomware data extortion attack

06 Jul, 2021