Eight non-technical factors for successfully implementing AI in marketing teams

Right now there’s a lot of attention being paid to AI tech. Not a day goes by without an announcement about a new LLM version, a new AI company, agentic AI, or the imminent arrival of artificial general intelligence (AGI). Amid this tech frenzy it’s very easy to get caught up in the technology at the exclusion of all else. All that’s needed for success is to select the right model or vendor and start automating business processes.

I’ve worked in technology for decades, and this is always the way tech sells itself. The huge period of hype, followed by rapid adoption, inevitably followed by the realisation that reality is somewhat different. It’s why Gartner invented the Hype Cycle, because nearly all technology follows the same pattern. For those unfamiliar with the Hype Cycle, it identifies five key phases for any new technology:

First comes the “Innovation Trigger” - a breakthrough sparks interest. Early successes and media hype inflate expectations until they reach the “Peak of Inflated Expectations”, people expect that merely applying the technology alone will deliver results. Soon however, project failures and challenges cause expectations to fall, leading to the “Trough of Disillusionment”. As a result, projects become better defined, scope becomes more contained and gradually organisations learn how to not only apply the technology but also make the right non-technical business changes necessary for success. Finally, widespread adoption occurs, and business benefits are realised – called the “Plateau of Productivity”

For marketers, generative AI tools like ChatGPT and DALL-E reached the "Peak of Inflated Expectations" in 2023.? Marketers imagined a future where AI could automate content creation, optimise campaigns, and predict consumer behaviour with unprecedented accuracy. By 2024, many business users encountered significant obstacles, ranging from poor ROI, poor data quality as well as ethical concerns, placing generative AI firmly in the "Trough of Disillusionment."

In 2023 Gartner reported that 80% of AI projects failed to meet business goals. That’s almost double the failure rate of other IT projects. It highlights the huge gap between expectations and reality. Boston Consulting Group reported a 70% failure rate due to poor data quality and misaligned objectives. S&P Global also recently reported that 42% of businesses abandoned most of their AI initiatives this year, up from 17% last year. That’s some serious disillusionment.

They key reasons for failure across multiple pieces of research can be summarised as:

1. Misaligned Objectives - lack of clarity about what they want AI to achieve. Projects are started without well-defined goals or alignment with wider business strategies.
2. Poor Data Quality - AI models require clean, relevant data for training. Many marketing datasets are fragmented or incomplete, leading to unreliable predictions and insights
3. Overhyped Expectations - Generative AI tools are often seen as "magic solutions," leading marketers to adopt them without fully understanding their limitations. This results in disappointment when outcomes fall short.
4. Lack of Expertise - Teams lacked technical expertise in data science or machine learning, leading to poor implementation and management of AI tools.
5. Unexpected consequences – wider business issues elsewhere in the process limit the benefits, or sometimes lead to changes with result in a poorer result e.g. a diminished customer experience
6. Pilot Paralysis - Many organizations struggled to scale successful pilot projects due to inadequate infrastructure or unclear pathways for production deployment.

Overall, the reasons for failure are often skewed towards non-technical factors. While selecting the right LLM or AI model is important, it really isn’t a key determinant of success.

The right factors for AI success

Most success factors for marketing teams tend to be human and organisational in nature. Yes there are some technical factors relating to skills and data, but overwhelmingly success is driven by people and process.

Key success factors for successfully implementing AI in marketing teams are:

1. Having clear marketing objectives e.g. reducing specific campaign costs by x% , or optimising ad spend by Y%, rather than simply building a demonstration of the technology.
2. Work closely across departments and teams. You can't do this alone, working with other teams such as IT, sales or customer services has always been vital for marketing success, AI is no different. You need to ensure the technical capabilities and business goals are aligned.
3. Improve AI skills, to ensure your marketing teams can make use of the technology, and even more importantly fix problems when they arise.
4. Invest in data quality – taking time to improve data and apply governance to the data to ensure consistently high quality. Make sure other teams across the business understand how important it is for them to get data right, even if it doesn't directly affect them.
5. Adopt new ways of working. AI can have a profound impact on business process. If you’re adopting AI but planning on keeping everything else the same, there’s a good chance you won’t be successful. Think about what impact your changes will have on the experiences of end customers, internal customers and your own teams. Don't just think about things that people might have to stop doing, but also new things that are enabled.
6. Start small and gradually making continual improvements over time. Big bang changes are high risk and take a lot of time and money. If you get it wrong it can cause a lot of disillusionment, look to start small and make iterations.
7. Look out for new business models – AI can fundamentally change where value is created in your business, this often requires new approaches to business models to better captialise on new value created, or to prevent value from being eroded. For example applying automation to a standard business services and unintentionally eroding value from a paid premium service.
8. Continually measure success – have the right measurement processes in place so you can measure the performance of AI driven activities, such as marketing campaigns and adjust based on real world results

Applying product thinking

This might sound like a daunting list of success factors. However, many of these can be addressed by applying product thinking to the problem. By productising your marketing operations, you can improve AI adoption and address most of these factors simultaneously.

Productisation can ensure you produce clearer definitions of your marketing processes, improving visibility and understanding of your offerings across business teams. ?It applies clear success metrics and makes each service more measurable. It also changes ways of working, applying structure and modularity to your offerings making your marketing processes more scalable.

Essentially, taking a productised approach gets your marketing operations in a ready state for automation. You can make informed choices about what to prioritise for automation and ensure that the process is contained and measurable. It’s far easier to manage rollout of AI in a predictable way, rather than applying AI to unstructured business processes with unpredictable outcomes.

If you’d like to understand how your marketing organisation is placed to adopt AI and to apply productisation, why not complete our free self-assessment test? ?It takes less than five minutes and will give you an instant, detailed assessment of areas that you could improve with product thinking.