In the last part of this three-part series I want to explore where Generative AI may lead us and imagine the potential it offers marketers.
What comes next does feel a little bit like science fiction as we need to consider Artificial General Intelligence (AGI). AGI is still, for most of us, a theory and nothing more. While organizations are making inroads toward AGI none have yet passed the Turing test and therefore cannot be labelled as artificial intelligence machines. Most leading experts do not believe we will achieve true AGI before the year 2300 which puts it out of scope for any immediate marketing application.
If we accept that AGI is likely to remain just a theory for the foreseeable future what is next for GenAI after multimodal GenAI ? In my opinion the next logical step is neuro-symbolic AI (NSAI).
NSAI is an approach that combines both neural AI techniques and symbolic AI architectures to address the known weaknesses in each of the two approaches. In essence it is a step on the pathway to AGI giving us AI solutions that are capable of reasoning, learning and cognitive modelling.
The key is in the two approaches. We have known for some time that sub-symbolic solutions such as GenAI are good at solving certain types of problems but the way they achieve it can often be hard to audit and validate. Sometimes it’s even hard to understand the proffered solution as the networks are constrained by their data sets which inhibit strong generalization and can also introduce significant bias. Symbolic AI on the other hand, based on rules, logic and reasoning, may not be as efficient as sub-symbolic but has a high degree of transparency and is fully explainable giving high levels of confidence and trustworthiness to any solution. So, an approach that combines the two techniques of low-level data intensive perception and high-level logical reasoning provides us the potential for transparent, reliable, and trustworthy solutions.
How are organizations already employing these techniques? Well one example is neural network analysis being used on data collected from CCTV cameras to provide emergency services with the ability to route ambulances and fire crews to incidents avoiding road congestion and preventing those same vehicles causing road congestion. Each CCTV camera has its own neural network model, and the neural networks are then assembled in a symbolic layer to make sense of the whole street network in an area.
But how does this translate to marketing? Well, every marketing organization sits on a wealth of data that can be used to support the symbolic layer in the NSAI. This comes from transactional data, demographic and behavioural data, and digital journey data used to support attribution and other modelling techniques. Most successful marketing teams are putting this data to good use. Today GenAI, the neural element of the NSAI, is breaking ground in customer service, employee support and content creation.
If we bring these things together in an NSAI based solution the applications are exciting. Imagine a bank that combines customer transactional history and customer service interactions to derive a deeper level of understanding automatically. Perhaps the customer has received a job promotion and salary increase, maybe they are getting divorced, maybe they have a child on the way or one leaving for college – imagine being able to automate the creation of journey options available to this customer and individualize the steps in the journey with personalized content generated by using multimodal capabilities.
Recommendations for which journey options are best and the likelihood of success for each journey option will be managed by the symbolic capabilities. All aspects of the content and communication will be managed by the neural capabilities (which as we have established has already fed data into the symbolic side of the equation from the customer service interactions in the first place giving insight to sentiment, context and eventually emotion of the customer interactions).
As marketers we are always keen to embrace change but the whole topic of Generative AI and the potential it offers is one of the most exciting things to happen in my generation. I believe that the opportunity given to us through approaches such as generative, multimodal and neuro-symbolic AI will allow us as marketers to unlock the true value of the wealth of customer data available to us today and complement this with a deeper and clearer understanding of individual customer needs. And at the same time to scale the creation and measurement of content across multimedia and multi channels.
We sit on the cusp of the greatest opportunity to lift customer experience to new heights. Let’s make sure we all make the best of this for the organizations we work for but most importantly the customers we service and support.
