What is Conversational AI?
Conversational Artificial Intelligence (AI) is the technology that enables automated assistants to understand, process and respond to human language in a human way. Unlike traditional chatbots, conversational AI has the ability to generate complex conversational dialogue based on the specific policies, vocabulary, and processes of an organization.
Conversational AI uses natural language processing (NLP) engines to interpret unstructured customer queries and convert the language into a structured format that can enable an automated assistant to provide a response.
Automated assistants can be designed to collect, collate and assess information alongside presenting data back to end users. They can also be embedded with multi-media capabilities, such as videos, forms, photo upload, and signature capture to enrich the customer experience.
What are the Benefits of Conversational AI?
The key benefits of deploying conversational AI technologies include:
- Acting as triage for queries that can be automated, allowing customer service agents to deal with more bespoke requests
- Cost savings as an increased number of customer interactions can be automated
- Increased customer satisfaction as requests can be resolved faster without long wait times or lengthy queues, with the option to provide out of hours support
This technology can be implemented in short timescales and with a small number of use cases to demonstrate value quickly.
What are some of the Use Cases for Conversational AI in Insurance and Retail Investment Firms?
Conversational AI Chatbot
The following abilities can be supported by conversational AI:
- Answering customer FAQs and more specific queries, including questions relating to:
- specific firm policies
- product information
- policy renewal dates
- insurance, and investment terminology
- Responding to internal support queries by providing information on internal processes and compliance policies
- Performing processes for a customer such as
- death notifications for life insurance
- changing a customer address on a policy
- submitting an insurance claim and uploading the necessary evidence
- Onboarding customers including document upload and ID verification
- Producing outbound campaigns and marketing materials for new insurance and investment products
ChatGPT is at the Forefront of Minds, so how is this Impacting Conversational AI?
ChatGPT (and its underlying engine – GPT4) analyzes information from data sources across the whole internet to formulate responses and the knowledge cutoff point is September 2021 – meaning answers to queries may not always be reflective of the truth. It would be a high risk strategy to open ChatGPT directly to customers, as firms will want information released to customers to have compliance approval and GPT may ‘make up’ answers to questions if it cannot source one.
However, ChatGPT’s underlying generative AI technology can be utilized within a specific data set (e.g. a firm’s web page) and subsequently be used to formulate responses to queries in a conversational manner. Additionally, when firms are designing responses to queries, ChatGPT can be used to provide training data for employees building responses – improving the employee UX, whilst also enabling firms to build automated assistants at a greater scale.
Overall generative AI is certainly improving conversational AI but needs to be used in a controlled way. In the future, as models increase in maturity and become better trained, the impact generative AI will have on conversational AI will only increase as it becomes more proven and the risks reduce.
For more information on how generative AI can impact Insurance and Retail investment firms, please take a look at our recent white paper [Generative AI for Insurance & Retail Investments].
What are the Key ‘Don’t Get Wrongs’ in Deployment?
From our experience of implementing conversational AI technologies, there are a number of key ‘don’t get wrongs’:
- Be prepared to iterate and pivot the strategy to drive value for the customer – a negative customer experience may be worse than keeping the status quo
- Have early engagement with IT to have a clear understanding of both how the technology fits into the firm’s wider architecture and the number of integration points
- Ensure InfoSec requirements are clear at the outset to avoid the need for unplanned security testing
We recommend starting simply by identifying 1-2 key use cases, then piloting a proof of concept to unlock benefits as quickly as possible and drive learnings for future use cases.
How can Alpha Help?
Alpha has experience in implementing conversational AI technologies and has helped clients understand from creating a business case to process design and optimization.
To learn more about Alpha’s conversational AI proposition and how we can help please click here and we will be more than happy to continue the conversation.
About the Authors
Keith Aylwin
Executive Director
Keith is an Executive Director at Alpha FMC who brings extensive global experience with 25 years in the Insurance industry. He has successfully led engagements covering strategy development, core technology change, target operating model design and implementation, process improvement and re-engineering, business process management, and distribution.
Henry Dalton
Manager
Henry is a Manager within the technology practice of Alpha’s Life, Pensions & Retail Investments division. Henry’s main area of focus is supporting clients deliver value through technology enablement, and he currently leads Alpha’s Conversational AI offering, aimed at delivering enhanced customer experiences and operational efficiencies using AI. He has experience managing the business case creation, design and implementation of conversational AI chatbots within the Life & Pensions industry.