The Intersection of Conversational AI and Robotic Process Automation (RPA)

AI, Bots

The Intersection of Conversational AI and Robotic Process Automation (RPA)

By 6 Minute Read

Who will win the process war? Chatbots or RPA.

Terminology abounds in the area of artificial intelligence (AI) and multiple definitions can lead to confusion and blurring.

In industry sectors like manufacturing, utilities, energy, finance, healthcare, and telecommunications, routine, rules-based, and repeatable business processes have been subjected to increasing degrees of automation throughout the decades. Increased accessibility and advances in the area of AI are creating another wave of opportunities to not only increase the degree of automation but also to apply greater intelligence and cognitive understanding to workflows and processes.

One area of technology evolution in process automation has been in the area of robotic process automation or RPA. But more recent developments in conversational AI may be set to disrupt how businesses  approach process automation. There seems to be confusion regarding these two topics with an increasing blurring of the lines between the two. So how do these two differ and how do they intersect? 

To try to clear up any confusion we decided to explore the characteristics of both, compare and contrast them, and uncover their business use cases.

The Characteristics of Robotic Process Automation

RPA applies software robots or bots, to further automate processes, eliminating inefficiencies, cutting costs, and improving speed and performance. It is the next stage in the evolution of traditional process automation, which has been around for decades.The technology can vary from basic rules-based automation to very complex solutions based on machine learning. At the more basic end of the spectrum, solutions are often an extension of traditional business process and rules management while at the more advanced end, the complexity means that data scientists are required to build highly custom AI solutions that are both expensive and difficult to scale or adapt with shifting demand. 

The key characteristics of RPA are:

  • It is applicable to routine, repeatable, rule-based, or predictable business processes.
  • It is primarily based on automating existing business processes.
  • It focuses on eliminating manual or human involvement in process steps.
  • It is governed by structured data inputs and not human inputs or chat.
  • It is a  process-centric, bottom-up approach
  • It decreases friction across process tasks by limiting human handoff
  • It focuses on back-office processes governed by rules and business logic, for example in the areas of finance, operations, production, HR, and distribution departments
  • It lacks flexibility to adapt quickly to changes or exception-handling
  • Finally, RPA implementations are generally IT-led with input from the business

Defining Chatbots and Smart Bots

Where the confusion between RPA and Chatbots comes in is that both are strongly associated with robots or ‘bots’. In the case of RPA these are automation bots and in the case of Conversational AI these are chatbots (often also referred to as bots). RPA bots don’t have a chat element where conversational- or chat- bots do. But the issue with chatbots is that they are evolving beyond just being virtual assistants to being much smarter conversational bots that can also handle specific business tasks and user journeys i.e. a form of intelligent process automation.

These chatbots are becoming Smart Bots, applying natural language processing (NLP) and AI technologies to simulate human conversations, perform automated business tasks, and become smarter through learning from the environment and their experience.

The key characteristics of a Smart Bot:

  • It is applicable to any user-centric process or journey that is initiated or conducted via voice (phone, physical or voice-activated interfaces) or message (text, chat, email, web, etc).
  • It is primarily based on transformation rather than pure automation
  • It focuses on simulating human conversations and understanding human intent to perform automated tasks.
  • It is governed by free form or guided conversations that are less structured
  • It is a  data- and conversation-centric, top-down approach
  • It decreases friction across process tasks by limiting human intervention
  • Current focus is on front-end customer-facing processes (for example customer service, sales, and marketing) but they are equally powerful in many back-office tasks where they are gaining traction.
  • It is very versatile, adapting quickly to changes and gaining intelligence and capabilities through knowledge gained from data and experience
  • And, chatbot implementations are business-led but can involve IT

Chatbots or smart bots always involve some form of conversational interaction, via either voice-activated or messaging interfaces. RPA, on the other hand, can be applied to a discrete process that does not involve any type of user chat or interactions. For example, RPA could be applied in generating customer invoices by downloading customer order information from back office systems or screen-scraping and extracting the information when it is not available from code. A bot, on the other hand, can be used to read a customer email or text requesting details of an invoice, interpret the customers intent, retrieve the appropriate data required to fulfill the request, and present the information to the customer. 

Recommended Reading: The Fundamentals of a Successful Chatbot Strategy

RPA and Chatbot Approaches Differ

Process automation has been around for decades with robotic process automation representing more of an evolution in approach. In RPA, processes are automated by teaching robots what to do rather than automating them through code or scripts. So it’s a step up the curve towards more intelligent automation.

As RPA evolves it is increasingly intersecting with conversational AI technology. In fact, more and more, recent media articles have added the word bot or chatbot when referencing RPA, making things even more blurred.  However, it’s important to recognize that the RPA approach is much different to the conversational AI approach.

Conversational AI represents a revolutionary approach, where conversations and user intent dictate the tasks and processes that are executed in order to fulfill the intent. By it’s nature it is more free-flowing and unstructured requiring a high degree of flexibility and versatility in orchestrating the right tasks at the right time. This contrasts with RPA where the processes are less fluid, and more rigid. Layering chatbots on top of RPA creates the potential to force a more guided and process-driven, rather than a conversation-driven flow of activities.

Bot orchestration is probably one thing that makes the conversational approach much easier, faster and more versatile than a RPA approach.  At ServisBOT, by using a clever bot orchestration approach, a virtual assistant can hand off a high-level customer intent to an appropriate task-oriented bot who can either execute on the  intents or invoke the help of another bot, if needed. In this way an Army of Bots can be assembled to handle multiple tasks in a customer journey and individual bots can be deployed or removed according to how the conversation guides the execution of tasks. This is a much more flexible and modular approach than traditional RPA.

But will Conversational AI replace RPA? 

Whether RPA will be fully replaced by smart bots is doubtful as there will always be use cases that lend themselves more to RPA, but there definitely are strong arguments that may lead to Conversational AI stealing some of the thunder from RPA. Or maybe there’s merit in the two coming together to unite the best of both worlds? 

Conversational AI is a greenfield technology and approach to smart conversations and these are increasingly not just focused on simple Q&A type interactions. Rather, a chatbot can now understand a high-level customer intent and pass that to a task-oriented bot to gather any additional data needed and execute the necessary tasks. As this evolves and matures, smart conversational bots are set to eat away at the RPA space.

The emergence of conversational AI platforms, unlike RPA, represent the latest technologies, tools, and approaches and are not lumbered with legacy systems and methodologies. And that can be a double-edged sword. The conversational AI platforms have to prove their approach and technology while the RPA players are generally incumbents that are deeply embedded with middleware solutions in enterprise. 

The Use Cases for Chatbots

The broad differences between RPA and Chatbots have been mentioned above as have some of the high-level use cases. Wherever there is a chat element to a process think in terms of conversational AI. And if the process is highly administrative with no chat element then maybe RPA is the right choice.

Don’t think too narrowly about conversational AI. Many of the current implementations are in the areas of more basic customer service use cases. However, businesses are increasingly extending their horizons beyond customer service, recognizing the potential for bots to go broad as well as deep across and within their organization and to be used in outbound marketing campaigns, in field sales or service operations, or in employee-facing engagements. Here are some use case categories that may help guide you in choosing conversational AI 

  • Service Bots:  These bots help automate simple tasks, service requests or FAQs. They can be customer facing or work behind the scenes, making service experiences more convenient and available 24/7, across multiple channels.
  • Journey Bots: These bots are more involved as they manage multi-step and sometimes complex customer journeys such as making an insurance claim, onboarding a new customer, or resolving a complaint. They guide customers through specific workflows, fulfilling particular tasks in order to achieve desired outcomes.
  • Automation Bots: These bots automate recurring tasks such as billing, renewals and appointment settings. They are responsible for proactively automating routine tasks to improve overall process efficiency, freeing humans to focus on more complex tasks.
  • Campaign Bots: These bots are typically deployed for outbound tasks, executing marketing campaigns such as proactive loyalty outreach, promotions, retention, winback, collections, and upsell or cross-sell opportunities.
  • Employee Bots: Don’t forget employee engagement when it comes to bots. Workplace bots can handle internal employee-related tasks, improving efficiency of field service workflows, HR-related tasks, and/or IT helpdesk operations.

For more information about prioritizing the use cases for conversational AI and building and deploying smart bots that can intelligently engage with your customers and handle complete customer journeys, you can visit our resource library or download a free copy of our eBook: Transform Customer Engagement with An Army of Bots

Close this Window