Marketing, digital, and artificial intelligence: what relationship can we highlight?
There is a very close relationship between these areas: it is no coincidence that the main industrial players in AI come from environments that we can qualify as digital and extremely close to marketing and content production.
The myriad of micro-decisions that must be made in real time starting from data, with an approach that Ilya Katsov (author of “Algorithmic Marketing”) would classify as programmatic marketing, have driven the development of a huge variety of decision support systems, which in turn have pushed the industry to accelerate developments and self-finance. The personalization of a content feed on a social network, the analysis of a query on a search engine, the management of the auction mechanisms behind online advertising, personalized pricing, and recommendation systems that suggest which product to add to our cart – these are all examples of contexts where it has been possible to directly link algorithms with value generation.
What is happening today is that a technology that until yesterday operated in a limited context is starting to spread pervasively, moving out of the niche in which it developed and preparing one of the greatest revolutions that mankind has ever seen.
In an era where AI models are increasingly accessible, how crucial has it become for companies to own, organize, and value their data?
Just recently, Gartner cut its growth estimates for global IT spending for 2024 from 8% to 6.8%; although generative AI is generating a lot of hype and interest from companies (Gartner also estimates that the CAGR for the AI software market between now and 2027 is 19.1% and that the generative AI component will rise from the current 8% to 35%), it is still not yet a driving force.
We are still in a preparatory phase for the growth of the AI market; the technology allows for fantastic things, but companies must prepare adequately to handle it. While we see companies hesitating in adopting AI, major digital players are investing heavily, further widening the gap and laying the groundwork to become mandatory entry points for working in this sector: the investments of Microsoft, Google, Amazon, Meta, and OpenAI are massive. Just consider that Meta aims to have over 600,000 GPUs by 2024, and this is only part of the investment planned by one of the Big players, who will be able to enter markets where they are not yet present even faster, which will be completely reshaped by the use of this technology.
Companies that do not want to be left behind must do enormous work from a cultural perspective and in building specific skills, and fortunately, I see a lot of attention on this. It will be increasingly important to be able to identify opportunities for efficiency and redesign of notable business processes in a distributed manner where artificial intelligence can effectively serve as a solution to known problems or as an accelerator for growth, and this can only be done through widespread awareness regarding the potential and methods for handling these technologies responsibly. We must quickly move out of the bubble of pure temporary experimentation and artificial intelligence reduced to training on how to use ChatGPT, to immediately enter the mindset of developing transformative processes for business. If we talk about marketing, it’s not about designing a few graphics or writing a few articles with AI, but about rethinking how customer interaction is managed or how functional information for strategic decisions is collected.
In this sense, the major enabling challenge, which is certainly not new but today becomes unavoidable, is that relating to the structuring of digital processes for data collection and organization – fundamental to be able to think about working on projects where artificial intelligence does not have a simple aesthetic purpose but can effectively be transformative.
If we talk about customer communication, for example, one cannot think of reaching an acceptable level of personalized interaction in line with current market standards if digital interaction data is not collected granularly and appropriately linked with CRM data; the prediction of customer behavior, support for cross-sell activities, and many other use cases enabled by artificial intelligence are reduced to simple toys if they do not rest on solid data sets.
Fortunately, artificial intelligence today also helps in this field, facilitating the collection and analysis of data that is so important.
How can companies prepare for the introduction of artificial intelligence and achieve a real impact on their business processes?
As happens in many other contexts, the game is played by reversing the perspective typically used for these problems; initiatives that do not start from technology, but that resolve known business problems (process optimization problems, customer problems, etc.) thanks to technology, are successful.
This is why artificial intelligence work teams are rarely composed only of software developers, requiring instead a strong understanding of the context in which they operate; the highly acclaimed “breaking down of silos” that is so often discussed becomes an imperative for competitiveness with the advent of artificial intelligence.
Consequently, technology in the enterprise sphere is following this trend, increasingly trying to use composable software approaches. Instead of working on monolithic systems that are the same for everyone, systems consisting of independent components that can work well with each other or with other systems are developed; thus, it is not the company that has to adapt to the software, but vice versa. Generative AI in this sense facilitates the journey as a whole, breaking down the barriers that make interaction between human and machine difficult thanks to some extremely interesting artifices.
Systems based on autonomous agents, for example, allow working on the execution of complex tasks that involve multiple actors and the use of company resources through interaction in natural language.
Many integration and automation activities today can be developed, after appropriate setup, thanks to interaction with natural language: if the company has customer data accessible and well-cataloged, it becomes easy, for example, to develop complex marketing automation flows, involving diverse touchpoints and tools, with personalized messages, without having to write a line of code, focusing on defining objectives and guidelines – an area in which we must increasingly be able to develop skills and ways of working.
How can artificial intelligence help companies identify market spaces and opportunities?
Data is the greatest treasure that companies can value at this historical moment; it is so both because it helps to direct and personalize how artificial intelligence can be put at the service of the company itself and the ecosystem in which it operates, and because it continues to perform its informative and decision-support function assigned by tradition.
Over the last twenty years, we have been buried under increasing volumes of data, which at times have seemed unmanageable or unusable, increasing attention toward topics such as big data analytics and leading to the explosion of strategic professions such as that of the data analyst and data scientist. As already mentioned, artificial intelligence facilitates the extraction of value from data, allowing for better organization, speeding up the identification of actionable insights and anomalies that can be the basis for important strategic decisions.
On one hand, in marketing, we have all the value that can be extracted from first-party data that companies collect on their customers: data coming from CRM, digital analytics systems, customer care interactions, apps, and all touchpoints through which the customer (or user) interacts with the company. This data must be appropriately collected, aggregated, and reconciled with unique identifiers that allow linking all traced interactions, so that through AI, interests can be inferred, and predictions and recommendations can be developed that are useful both at an implementation level, but also and above especially in terms of analysis to support strategy. Analyzing trend interests of the target audience and identifying behavioral anomalies can be the first and important signal to be able to change course before competitors, leveraging precisely one’s own information assets.
On the other hand, we have data accessible to everyone in a digital context (social posts, reviews on local listings systems, searches on search engines and e-commerces, geo-data coming from apps): a true hidden treasure not manageable with classic analysis methods and which instead artificial intelligence allows to handle quickly, identifying interest trends, hidden opportunities, and potential threats for reputational crises that companies must absolutely keep under control. In this, generative AI has helped tremendously as a tool for text analysis and structuring unstructured data, leading to a leap forward in the quality of outputs from sentiment analysis and phrase classification systems.
Transparency and responsibility in using AI and marketing data: what are the main points of attention for companies?
The topic of privacy and data usage transparency is now at the center of the discussion when talking about marketing and must be even more so with the ever-increasing adoption of artificial intelligence to support the processing of this data; in this phase, I believe that common sense must prevail and that one can be inspired by the principles of GDPR, on which Europe arrived first and served as a lighthouse for many regulatory processes in the West.
It makes one smile, for example, to see how part of the Ad Tech industry, with the gradual disappearance of third-party cookies, is finding refuge in even more invasive technologies (and above all, less communicated) such as so-called “alternative IDs,” most often based on device identification through techniques collected under the umbrella of fingerprinting.
I don’t think there are bad technologies and good technologies; instead, I am convinced that one must be extremely transparent with users and customers if they want to work on stable solutions, even having the courage to leave some points of marginality on the table in the short term. I am talking about mechanisms for sharing value with customers, who can and must be educated regarding the value of their data: these are now the norm in many contexts, and I don’t see why they shouldn’t be at the basis of any data collection initiative.
Transparent data management, the use of algorithms and technology at the service of users, and attention to these elements are and will increasingly be extremely important elements in the relationship that brands will have with their audience, and it is good to approach the topic correctly right away, avoiding having to resort to yet another brand washing when it is already too late.
