 Powerful tools for personalisation: Using large language model-based agents, knowledge graphs and customer signals to connect with usersSeth Earley and
Sanjay Mehta
Applied Marketing Analytics: The Peer-Reviewed Journal, 2024, vol. 10, issue 3, 271-288 Abstract: This paper discusses how large language models’ (LLMs) agentic workflows powering ChatGPT types of applications can use a combination of enterprise data sources to hyper-personalise information at scale for customers or employees. Typical use cases include marketing communications, customer support, content creation and digital assistants. The approaches described are at one level established in theory; however, practical adoption has been challenging and the combination of templated prompts with LLMs and agent call outs to external application programming interfaces and knowledge sources are new. The data sources using these approaches include knowledge, content and transactional data with near real time and real time customer signals. Customer signal data can include first, second or third party data that describes the characteristics of a customer or employee, as well as real time ‘digital body language’ — click paths, searches, responses to campaigns and chatbot dialogues. Two use cases in two industries — automotive and industrial manufacturing — will be detailed to illustrate how the same principles and approaches can be applied in situations that are very different, and how a knowledge architecture combined with retrieval augmented generation (RAG) should be developed and applied. Analytics to monitor outcomes and enable manual and automated course corrections will be discussed. The outcomes are unified and contextualised experiences realising the sometimes ambitious designs of user experience developers. It is easier to storyboard a design than it is to make it a reality. Marketing organisations are more and more responsible for the end-to-end customer journey and experience. However, the customer journey is a knowledge journey. At each step of the process, they are asking questions about the company, product or service. What product and solutions do you offer? Which ones are right for me? How do I choose a particular offering? How do I purchase or procure the product or service? How can I maintain it, and get service or support? How do I get the most from my purchase? What are the options for upgrading or enhancing my solution? These are marketing communications that consist of educating the prospect rather than selling to them. Today's prospects are empowered with greater information and understanding of offerings and the competition than ever before. Marketing is therefore responsible for helping them make the decision based on information and references that are presented at each stage of the journey. Keywords: marketing personalisation; customer journey; RAG; retrieval augmented generation; hyper-personalisation; customer experience; generative AI; artificial intelligence; ChatGPT; LLMs; large language models; knowledge management; LLM challenges; LLM solutions; knowledge models; metadata models; knowledge architecture (search for similar items in EconPapers) Downloads: (external link) Related works: Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text Persistent link: https://EconPapers.repec.org/RePEc:aza:ama000:y:2024:v:10:i:3:p:271-288 Access Statistics for this article More articles in Applied Marketing Analytics: The Peer-Reviewed Journal from Henry Stewart Publications |
Is your work missing from RePEc? Here is how to contribute.
Questions or problems? Check the EconPapers FAQ or send mail to .
EconPapers is hosted by the
School of Business at Örebro University.