With the hype around GenAI reaching a fever pitch, many organisations are rushing to implement without a clear strategy – often leading to inefficiencies or missed opportunities. In this article, Nick Yang, Head of Martech at data company fifty-five London cuts through the noise to shed light on some of the mistakes being made with GenAI implementation, and explores how marketers can take a smarter, more systematic approach that prioritises commercial value, long-term scalability, and above all, customer impact.
Generative AI has been the flavour of the decade so far when it comes to emerging technologies, with businesses scrambling to identify how they might benefit from its potential. The rush to stay ahead of competitors raises the question of whether marketers are taking the best approach when it comes to implementing GenAI? The truth is, most of the time, the answer is no.
The pressure for marketers to utilise AI is overriding opportunities to take a more strategic approach. There are many ways in which GenAI could be implemented as part of marketing and advertising plans, from content creation to data interpretation or automation, but one size does not fit all. Understanding how individual use cases can address issues or improve effectiveness within your own organisation is critical, focusing on AI-optimisation rather than AI for the sake of AI.
With so many possibilities and multiple stakeholders involved, it can be challenging to avoid the potential pitfalls. The common issues organisations experience can often be avoided with a more strategic or systematic approach, or by following best practice approaches, highlighted below.
1. Setting up for failure with a lack of planning
The first step of building any AI strategy should be to work with internal stakeholders and/or external consultants to establish where AI could add commercial value and drive success, whether that is in resolving issues faced by internal teams or customers or in improving efficiency.
Auditing the status quo and understanding what is possible, before building a detailed strategy for development and implementation, affords the opportunity to prioritise projects based on budget availability, personnel requirements or the scale
of impact.
For example, when working with a global fashion retailer recently, we reviewed their existing MarTech architecture and worked closely with key stakeholders to define their business priorities before developing a strategy. This strategy included a prioritised list of use cases, from foundational quick wins to more sophisticated AI-based applications, alongside a view of the necessary technical enablers to facilitate them.
2. Running before you walk
Larger companies or groups often have legacy systems in place that make it harder to implement company-wide initiatives. While one area of the business may have highly responsive systems that are able to process the scale of data required across many GenAI projects, others may be lagging.
Over the past 18 months, we have been working with a global FMCG holding company to deploy a consistent, robust analytics setup across all of their brand properties, which is a prerequisite before more advanced applications can be enabled. If the end goal is a company-wide roll out, then it is important to either work within the restrictions of your current stack or invest in updating them before you get too far along in the process.
From performing an initial audit of brand websites to understand the state of play to defining a standardised analytics tagging plan that could be implemented across an initial pilot brand before being rolled out across all properties, this work is essential
before taking any next steps.
3. Following the crowd
While it is helpful to understand how competitors are approaching AI, it is important to stick to your own strategy while it still serves your organisation. Every organisation is different, facing its own pressures, priorities and challenges in a tough economic climate, so it’s important to think about the areas where there is the most potential for AI to make a difference for your company. And it may be that its best application is within something more foundational.
For one global retailer we worked with recently, fifty-five created a GenAI-driven solution to automate the manual and time-consuming task of updating product stock keeping units (SKUs). The average product feed can take four to five days per month to properly maintain and update to ensure best practice adherence and make sure it covers all existing and new product SKUs. By using GenAI to analyse historical product feeds and train it with best practice guidelines, the solution can automatically generate updates or new descriptions by extracting information from pre-defined sources, eradicating the vast majority of manual work previously required.
Ensuring that each SKU had the correct information around variables such as price, description and images helps marketing campaigns, such as Shopping ads on Google, to perform more effectively. While the solution in this case resulted in a 34 percent improvement in their Shopping campaigns Collaborative Performance Advertising Solutions (CPAs), it might not be beneficial to all retailers who have different processes or marketing strategies – and blindly following course could result in wasted resources without the desired result.
4. Forgetting customer needs
Another fatal flaw often made by those embarking on their AI journey is to lose sight of the customer journey in favour of AI-driven efficiencies. There is a balance to strike between initiatives that could streamline existing processes and reduce overheads and the impact these might have on the customer journey. When poorly implemented, AI can negatively impact the customer or brand experience by creating a lack of human interaction or providing inaccurate or irrelevant responses, causing frustration and additional stress.
Last year, we worked with a global CPG brand to set up an AI chatbot – identified as an improvement to the customer journey. When research revealed that women were uncomfortable having conversations about female health products, the client set out to offer a service that enabled them to have a private conversation about a sensitive topic without directly engaging with a person. The GenAI-powered chatbot ensured that they would still receive the required information in an engaging, personalised manner, whilst remaining within a private, safe environment.
The secret to unlocking GenAI’s potential lies in laying the groundwork for success and focusing on the use cases that will be most beneficial to your organisation and its customers. Of course, you could try to pursue every possibility – but it is an expensive, inefficient and ineffective approach that will not deliver the results. GenAI should be seen as a business support tool rather than a replacement for entire processes or groups of personnel. If your organisation has the capability to introduce AI, then there is no reason it cannot be used to increase efficiency, allowing teams to focus on high-level strategies that can otherwise be deprioritised.
