Within a year of ChatGPT’s launch, more than half of businesses (54%) were taking steps to implement GenAI in various parts of their organisation. In the marketing world, this included tools that could help generate imagery and text and create smart personalisation when communicating with customers.
Disappointment with GenAI, Who’s to blame?
But, by the start of 2024, businesses had realised that something was missing. Despite the enormous potential, GenAI wasn’t proving to be the route to quick success many thought it would be. There are numerous examples of companies that have struggled to effectively apply this technology due to incomplete, out of date or false information. As such, it has started to fall from favour.
The problem isn’t GenAI though. The generative capabilities of this technology remain truly impressive. Where businesses are falling short is with the quality of the data they are using to feed GenAI applications. This is no small issue. According to Gartner, just 4% of CMOs say their data is AI-ready.
To be fair, almost half of business leaders (44%) have already recognised this problem and are actively driving forward data modernisation programmes to support their AI ambitions. From a marketing perspective, this modernisation process will heavily focus on organising customer data.
This means, however, that instead of just investing in GenAI, companies are also having to invest in other forms of AI. This is necessary to build a trusted foundation of data that marketers can use to create effective campaigns and engage with customers in a more personalised way.
Messy data
The challenge for marketers is that customer data tends to be very messy. Firstly, it comes from multiple sources such as in-store sales, ecommerce platforms, email campaigns, customer surveys, CRM systems, etc. Secondly, it is rarely in a uniform or consistent format. For example, a customer could be named Cathy Smith, Catherine Jane Smith or C. Smith depending on the channel they interact with. How do you know that they are the same person?
For marketers who are looking to personalise customer communications, it’s an issue if you don’t really know who you are talking to. Effective personalisation requires marketers to know what their customers like, what they are likely to buy next and what messages are likely to resonate, and on what channel.
Resolving customer identities is key. It’s what will help marketers gather insights, make predictions and turn customer information into actionable data. And this is where those other forms of AI come into play.
Getting to grips with data
A significant trend in the data management space is a shift towards ‘lakehouse’ architectures - a combination of data lakes, a centralised repository for raw data, and data warehouse, which stores more structured data.
This architecture is helping marketers see data wherever it is, regardless of where it resides within the marketing stack. No longer do they need to extract, transfer and load it anywhere else to build a unified view of a customer. This ‘zero-copy’ approach is saving a huge amount of time – while improving data security and privacy.
It’s also helping to generate customer insights. Deploying AI tools within a lakehouse architecture stitches various sources of data together and finds the connections between different data points. This enables marketers to understand if Cathy, Catherine Jane, and C. Smith are, in fact, all the same person.
Generating customer insight
By applying AI within this data infrastructure, marketers can also unify every touchpoint in the customer journey, from first interaction to last purchase. This is providing a comprehensive customer profile - which is what is enabling far more effective personalisation across all marketing communications with customers.
It is also helping marketers to extract insight and activate campaigns faster. Whereas developers were required to write SQL code to enable data queries, GenAI empowers any member of the team to ask the lakehouse a question in natural language. An answer that would have taken days can now be retrieved in minutes.
Last mile AI
When GenAI is used for content creation, this is really the last step in the process. To steal a phrase more commonly used in logistics, it is simply delivering the message to the customer over the last mile.
So, instead of thinking about AI as a tool for content creation, marketers should be looking at how AI can also add that critical layer of intelligence that will help them to generate customer insight.
This will give brands a better understanding of how customers truly behave, determine the best way to engage with them, produce the most relevant content and, ultimately, help them to activate more effective campaigns, quickly.
Editor’s Note:
Matthew Biboud-Lubeck is general manager of EMEA where he is responsible for the commercial expansion of Amperity, a leading customer data platform trusted by brands like Reckitt, Under Armour and Wyndham Hotels & Resorts.
Lubeck joined Amperity in 2017 to help launch the company and has served in a number of key roles building sales, customer success, and marketing functions. He established Amperity’s LGBTQ employee resource group (ERG) and is a trusted advisor and customer-centricity change agent to the C-suite across leading consumer brands.
Prior to Amperity, Lubeck spent 10 years with global beauty conglomerates Estee Lauder Group and L’Oréal as Group Head of Customer Data Strategy and Analytics, leading 30 brands across luxury, mass and salon professional divisions to better use data & unlock incredible beauty experiences, establishing L’Oreal as an industry leader. He resides in London with his husband and young daughter.
