While 2023 was the year of discovering AI, for many 2024 will be the year of action. Acting on the appetite within organisations to deploy AI, on boardroom pressure to see value from AI and on the many solutions pitched to enterprises that could solve their problems and keep them ahead of the curve.
But weak data foundations will continue to be the major blocker. Without the right infrastructure, quality data and necessary controls to ensure risks are understood and mitigated, AI will continue to be an experiment and rolling it out at scale will continue to be a pipedream.
To help organisations begin to make this a pipedream a reality, let’s look at how current businesses are already using AI. Let’s investigate what businesses and their leaders should do in order to capitalise on opportunities. Finally, let’s ask what they should establish to get the right data foundations.
Those businesses using AI are mostly focused on low-level, relatively tactical uses.
According to Mesh-AI’s research report, The State of AI in the Enterprise, 32% of business leaders say they are already using AI to increase operational efficiencies. This is followed by improving customer experience (20%) and improving business decisions (16%). This could take the form of content creation, automated chatbots based on corporate data, or knowledge extraction.
The promise of Generative AI, when paired with strong data infrastructure and the right controls, is already being realised by a small number of enterprises. For example, we’re working with an international bank to use Generative AI and prompt engineering techniques to help them validate ways to scale Generative AI in their business in a safe and ethical way. This will save them time and resource experimenting when they could be rolling out at scale, and leading to multi million pound uses.
Mature data capabilities are holding back enterprises from making the most of AI opportunities, with data quality the leading obstacle for AI innovation according to our research.
While 2023 was the year businesses woke up to the importance of data quality, 2024 will be the year that the enterprise connects the dots from a robust data-driven approach to providing tangible, and reportable business value.
CIOs will have been told that the solution to their data problems is a data mesh, data fabric, data mart, and data lakehouse. But instead, they should consider a more pragmatic approach that looks to combine the best of all of these concepts for one main goal: how do we make high quality data available across the organisation to deliver value to the business?
Extracting value from data on a large scale demands that companies transform how they’ve traditionally managed their data. The emergence of data mesh has been a catalyst for this, setting the foundation for what future data-driven organisations could look like.
Data mesh is not just a technology upgrade to an existing data platform, even through the acquisition of top-end new tools. While the journey to adopt data mesh demands effort, it's a worthy endeavour when aligned with value outcomes as defined by the business. Tangibly, this could be to support compliance and regulatory requirements without compromising on innovation, while ensuring data is trustworthy, high in quality and accessible to all areas of the business.
There are a number of factors to consider when beginning any AI transformation journey, and they fit into five key areas: use case, strategy, data foundations, talent and compliance.
Consider the value you want to see, aligned with the goals of the business of course, and who is going to be the main user group.
Develop a framework based on how risky and how valuable each use case is. Make assumptions based on how you would start each one, and ask yourself if there are multiple use cases associated with each other that the same solution will work for. By bucketing your use cases, you may find there is a more streamlined route forward.
Consider your organisation’s appetite for AI and how well defined the strategy around AI already is, or if one even exists at all. If you’re not comfortable with your strategy, take our accelerator to help you define this further.
Build a matrix of your strategic initiatives and figure out how AI can support them. Some will be simple, low risk and consistent from one organisation to the next - i.e writing proposals, pitch decks, or consuming information. Some will be larger issues that require more long-term planning.
For the organisations Mesh-AI speaks to, we find that many have a backlog of AI use cases, but no way to start them. The answer often lies in the low quality disparate data they possess, and tackling how data is created, collected, processed, stored and accessed is a crucial step in the AI journey.
A more modern approach to data can enhance the quality and accessibility of data, leaving fewer gaps in data and creating fewer hurdles for AI adoption.
Successful AI adoption requires a robust environment containing the right data engineering skills, machine learning capabilities, and expertise in prompt engineering. CIOs should ensure they have the necessary resources and skills for their chosen approach.
Our research found that many lack the necessary in-house skills, with only 33% of respondents saying they have the talent to safely and securely innovate with AI. But not having the right skills now doesn’t mean you can’t get a head start. Hiring a team of AI experts just to get an AI operation off the ground is unrealistic, so partnerships can help to introduce capabilities and make your business a more attractive place to work on such AI projects.
There is a general shift at the moment to consider regulations as guardrails, not hurdles, when it comes to AI adoption, given the considerable pace of change we’re seeing.
CIOs need to consider data privacy, authorisation, and governance, while also being aware of regional and industry specific regulations. Compliance and legal teams must be involved to ensure that data is handled responsibly.
Our experience across highly-regulated enterprises has shown us to achieve success, you need a prescriptive framework focusing on people, process and technology to baseline how ready your organisation is to become data-driven.
In doing so, you can quickly see results from your investments, leading to tangible value as you improve, optimise, enhance and monetise how you deploy your data and innovate with AI.