Emerging Trends for Data & AI Strategy in 2024

As an organisation specialising in Data & AI Consulting pre-Chat-GPT, it has been interesting to follow the market during this period of rapid disruption. It remains clear to me is that no emerging technology, in isolation, can be a "silver bullet" for the myriad challenges organisations have with data and as ever, the devil is in the detail (and the very real-world context!).

Here are some of the key insights into the trends we're seeing in the market and have recently experienced in our partnerships with clients.

#1 Outcome-driven Data & AI Strategies

Data Strategies started being embedded in organisations at the turn of the century for tech-forward industries. Early strategies were heavily focused on technological enablers and data management. These were strategies largely owned by IT and devolved from business strategy.

As time has passed, many of these strategies did not receive the investment they needed or failed to deliver tangible impact to business outcomes. That's not to say the work in these strategies was not valuable, as that's not true, activities around fostering a culture of data-driven decision making and data use has been pivotal in re-shaping how organisations function.

However, the ad-hoc use of data has created its own challenges, siloed practices, lack of single source of the truth and tremendous data quality challenges across the board. This is the backdrop against which there has been a renewed interest in refreshing and creating outcome-driven Data & AI strategies.

Modern data strategies resemble much more a business strategy than earlier strategies and are clear on the golden thread from data related activities to business value. We have found the creation of Strategy Blueprints has been incredibly valuable for our clients to draw this link, showing how prioritises high-value use cases can deliver business value, underpinned by strong data governance and management practices.

As budgets are constrained and we come beyond the inflated peak of expectations from AI, we expect the need for simplified and actionable strategy will bloom. Showing incremental value and designing strategies which are shorter-term and mindful of the pace of the Data & AI market will be most successful.

Designing such strategies requires pragmatism and a strong problem-solving approach, and existing frameworks will not cut it, leading to the need for tailored strategies as the maturity gap in AI and Data between organisations varies so significantly.

Democratisation of AI

It's hard to miss how easily accessible AI tools are to each of us in our daily work and personal lives; these will only become more embedded as time goes on. This is a huge opportunity for productivity gains, but in sensitive areas, without adequate appreciation of the Data & AI Literacy required, can bring significant risks.

This is particularly pivotal now, as the range of tools skyrockets and there are multiple options to choose from for the same use case, e.g. a custom LLM vs. using Microsoft Copilot's built-in capabilities. These each come with trade-offs around cost, risk and bias which should be evaluated for each use case.

As our clients embark on journeys to leverage AI innovation, we have found an increasing need to support with increasing data and AI literacy across the board. This is the key to successful democratisation of AI, and in many cases, there is a need to go back to basics around data literacy to prevent a magnification of impact.

This will be a key area of focus across organisations in the coming months as our client’s pilot new tools.

Data Value & ROI

AI and data initiatives have been around for a long time, in one form or another. So why have they not been widespread? Well, it links back to Democratisation of AI. The investment and expertise required was a barrier to entry, now that barrier is significantly lower, we need to consider if the solutions are value for money and providing value to the business.

We're past the days of creating proof-of-concepts which have no place in business as usual. We've all seen data initiatives which have failed to deliver value at scale, and do not provide ROI, fail. We now have the experience to understand and estimate benefits and costs for different AI tools and cloud platforms.

There is a trade-off with much of the emerging technology in the marketplace, which is why what makes GPT4 the correct tool for one organisation and role correct, may not be the correct solution for a smaller team who may benefit from a custom solution built with open-source AI models. We've been working in AI since before ChatGPT, and will be working in it far beyond, so understand how to determine the best technology, be that AI, Automation or something simpler.

We have seen this recently with a client of ours, where the challenge was establishing a benefits case for an LLM-based summarisation tool for case handling; we provided support in designing a robust experimental approach with metrics to understand whole system time savings and evaluating these along with cloud consumption costs. These underpinned a business case for BAU deployment of the tooling, which was driven by the size of the user-base and security requirements which influenced the choice of model and cost-profile of the solution.

It is becoming increasingly important to provide detail around costs and benefits so ROI can be determined for implemented solutions, even before we start experimenting. ?

Data Governance & Responsible AI

While Data Governance went "out of fashion" in recent years with the hype around AI, it's making a comeback as organisations realise the Foundational Data enablers are essential for emerging tech to be utilised to its full potential.

This includes an increased focus on data quality and remediation, starting small, where there is a clear benefit beyond housekeeping. The evolution of legislation and policy around AI leaves a lot open to interpretation, and there is a need to understand and design effective business wrappers and guardrails around technology.

It is important to note that a safe technology, with a poor business wrapper can be even riskier than a low-quality AI solution embedded in a business process with adequate guardrails and can still drive efficiencies. We are finding that an increasing number of our interactions with clients are around providing this challenge and AI leadership across their innovation teams. There is a safe middle ground with much AI implementation, and this is not only dependent upon technology, but on the interface and usage of that technology.

We can't continue to outsource Responsible AI to engineering teams, and need to instead bring our expertise to challenge their ideas and ensure the understanding and ownership sits with business users, who have levelled up their AI literacy, with our support.

We'd love to talk more

If any of these points resonated with you, please do connect and reach out for a chat with one of our experienced team of Data & AI Consultants.

Our Data & AI Consulting practice focuses on three pillars of Data & AI Strategy, Data Foundations and AI Foundations. This encompasses aspects of Data Governance, Responsible AI alongside Strategy and business case creation.