by Sam Carter
LONDON - Artificial intelligence (AI) is often an abstract term in marketing, over-inflated by clickbait headlines about robots taking over the world and self-driving cars.
While AI and its machine learning (ML) subset are already a very real part of the marketing and advertising ecosystem, businesses will struggle to apply these technologies in a meaningful way if this perspective remains. The industry should be considering the application of AI and ML on a more granular level to solve business problems, looking at real-world use cases and finding tangible solutions to their associated issues, whether practical or ethical.
Real applications of AI today
AI and ML are already empowering marketers to do things that were previously unheard of. This is largely due to the ability of these technologies to make sense of large and growing volumes of unstructured data from numerous sources and formats across the digital media marketplace.
By stitching together the complex consumer journey, which typically involves multiple channels and devices, AI and ML enable a single, precise view of the consumer as well as determining patterns in how they interact with a brand and what encourages them to convert. Insights from this process allow unified measurement of marketing performance through algorithmic models, which was impossible before AI entered the mainstream, to create accurate channel mix predictions.
The potential to piece together the consumer journey also enables AI-powered personalisation. By leveraging data from a diverse range of touchpoints, AI can provide a complete picture of consumer activity and can intelligently segment audiences using highly specific interests and behaviours. This approach allows marketers to tailor the consumer experience, delivering relevant personalised messaging that drives results.
So what is holding marketers back in embracing this new technology?
The practical concerns
While the marketing industry is still at the very beginning of what AI and ML can ultimately deliver, it is important to address issues surrounding the technology that will only be amplified as its use accelerates.
One of the biggest concerns is a practical one Ė the use of incomplete data and siloed data. The results achieved by AI and ML will only ever be as good as the information used to feed them, so organisations need to unify their data before jumping into the use of these technologies to generate the best outcomes.
While the General Data Protection Regulation (GDPR) and other impending regulations such as the California Consumer Privacy Act (CCPA) appear to restrict the use of personal information, they are actually having a positive impact by encouraging businesses to focus on organising and connecting the first-party data they already have. Because compliance requires all personal information to be linked to an individual, data held by a business is increasingly being joined up rather than stored in silos.
Connected data underpins effective use of AI so, by compelling organisations to think in this way, data regulation provides a real opportunity for marketers to gain a robust data foundation and make better use of the technology.
The ethical threat
The second major concern relates to ethical use of AI, and as an issue with deep roots it will undoubtedly be more complex to address. Removing all human-led decision-making can be risky, so AI must always be used in combination with human consciousness rather than replacing people altogether.
At the same time, it could be argued the people behind AI are a bigger concern than the machines themselves. A recent survey by New York University research centre reveals a lack of diversity in the AI field is perpetuating gender and racial bias in the systems it creates. From defining the problem that must be solved to collecting and preparing the training data, conscious and unconscious human bias can creep into AI at multiple stages in the process, and standard practices in computer science are not necessarily capable of detecting it.
The long-term solution to this issue is to build more diverse and representative teams, but this will take time. In the meantime, regulators need to get ahead of the curve and require evaluation of algorithms where the output has a material impact on peopleís lives to ensure the absence of bias.
AI isnít an abstract concept for the future, itís part of the here and now. Marketers should consider the technology in a meaningful and helpful way, looking at practical use cases such as measurement and personalisation as well as tackling real issues around incomplete data and ethics. Only by understanding the technologyís application at a granular level today can the marketing industry build a foundation for future AI-powered success.
Sam Carter is CEO of Fospha, a data-driven marketing firm