Key AI Readiness and Data Trends For Technology Leaders

KEY AI READINESS AND DATA TRENDS FOR TECHNOLOGY LEADERS

CRUCIAL FACTORS TO KEEP ON OUR RADAR FOR 2025

‍

At the beginning of every year, I compile some of the trends that are emerging in data and AI. My overall feeling is that the data and AI landscape continues to evolve at a breakneck pace.

With this pace of change and growth, I am sure technology leaders like me are faced with the dual challenge of keeping up with advancements while ensuring our organisations are AI-ready.

In this article, I hope to explore key trends shaping the data and AI ecosystem and provide a starting point for building readiness to grow and adapt in this ever-changing environment.

‍

1. The rise of AI Agents

If you read or watch some of the media that highlights data and AI trends, you will see that most hyperscalers and AI companies are talking about AI Agents or Agentic AI. This is a shift from using AI as an assistant that helps you perform tasks to AI running tasks and actions autonomously. Companies like Microsoft, Google, OpenAI, UIPath, to mention a few, have released the AI studios to create agents

Built on the growth of Large Language Models and in some cases existing workflow platforms, AI agents can perform functions on behalf of a human role player. The agents themselves can be created through Generative AI prompts to reduce the time to build and configure agents.

The role AI agents play in an organisation are limited by the imagination of those configuring them, from answering emails, to coding, to structuring information and many other use cases.

‍

2. Retrieval-Augmented Generation (RAG) and Contextual AI

Related to AI agents and in most cases underpinning agents is retrieval-augmented generation (RAG). The need to improve AI accuracy and relevance is instrumental in this trend. It is emerging not only as a game-changer but also a potential improvement for some of the criticism on LLMs around stale training data. RAG uses curated data to enhance the contextual understanding of AI models, enabling them to deliver more personalised, actionable and more accurate insights. This trend underscores the critical role of data quality and observability in AI readiness. Organisations that intend on using RAG in their deployment of AI applications must ensure that their data pipelines are designed and built on solid foundation ensuring proper data security, observability and quality.

“Ensuring data quality is essential for generative AI success,” notes Núria Emilio on the Bismart blog. Continuous monitoring and proactive resolution of data issues will be vital for organisations aiming to deploy reliable and impactful AI solutions and enabling the effective use of RAG.

‍

3. Large Language Models (LLMs) Drive Data Infrastructure Transformation

Large Language Models (LLMs) have redefined AI capabilities, and their influence on data infrastructure is set to deepen in 2025. In an announcement in January 2025 by Sir Keir Starmer, the British Prime Minister, was all about building infrastructure to “turbocharge” the use of AI.

The infrastructure upgrades revolve mostly around computation and storage and using additional data capabilities such as Vector databases to enhance the efficiency of AI.

Companies such as NVIDIA are leading when it comes to advancement of AI focused computation and processing components.

I believe that transitioning to such architectures will be critical for using LLMs effectively with the need to balance the cost and impact of these architectures. An interesting view from Andrew Rabinovich from Upwork highlights, “Smaller models, specifically custom models, will become a common solution for many businesses, as they are precise, efficient, and tailored to unique needs.”

There will be a need to strike a balance between deploying powerful LLMs and optimising infrastructure costs by adopting more tailored solutions for specific use cases.

‍

4. From Data as an Asset to Data as a Product

When engaging with customers, I have seen that on the data side, over the past few years, there has been a shift towards treating data as a product. This trend emphasises applying product management principles to data assets. Teams are implementing service-level agreements (SLAs), product roadmaps, and user-centric designs for internal and external data products. This approach allows for greater accountability and ensures data products meet organisational and customer needs, with the added benefit of identifying how data can impact an organisation’s revenue.

Maria Korolov of CIO.com observes, “The fields of customer service, marketing, and customer development are going to see massive adoption of AI-driven tools because of their proven ROI.”

Businesses that shift towards data as a product need to emphasise treating data as a first-class citizen in their organisation and provide both their business and technical teams with the adequate human and technological resources to shift the focus on bringing their data to an AI-production ready use.

‍

5. AI Readiness Through Converging Disciplines

I have seen a convergence of software engineering and data science, and this trend is likely to continue in 2025. Multidisciplinary teams that integrate software engineers, data scientists, and business analysts will become the norm. The realisation that AI is a team function, not an individual capability, ensures that AI solutions have a strong technological foundation, scalable, and aligned with business objectives.

The organisation’s use-cases for AI can vary widely which means that solutions will vary. In some instances, a solution may be as simple as connecting an internal service to the API of a major LLM provider. However, as leaders we need to understand that AI solutions go beyond the simple to ones that are not only technology related, but also cover areas such as data privacy, ethical considerations and specific country legislation. Having the right people with the right skillsets in any AI team that are aware of those nuances will certainly help building the right solution.

‍

6. AI Readiness: Preparing for Regulation and Ethics

As AI adoption accelerates, so do regulatory and ethical considerations. Fuelling this trend are acts such as the EU’s AI Act, along with initiatives like California’s transparency law for generative AI, shows the growing need for compliance frameworks. Vivek Mohan of Gibson, Dunn & Crutcher LLP emphasises the importance of proactive measures: “Companies must track the provenance of their training data and ensure accountability in AI decision-making processes.”

Ensuring responsible AI practices, including bias mitigation, explainability, and transparency, will be essential for organisations to maintain trust and avoid legal pitfalls. Businesses will need to invest heavily in tools and processes to reinforce their data pipelines with the right checks and balances in place to ensure reliable data observability, quality and security.

‍

7. Driving Efficiency with Rightsizing and Cost Optimisation

Following from last year’s trend of FinOps as applied to rising cloud infrastructure costs, organisations are prioritising right-sizing—optimising resources to achieve maximum impact at minimal cost is high on the CIO’s agenda. Tools for monitoring metadata, analysing usage patterns, and streamlining deployments are becoming indispensable.

“Balancing the use of more data with cost reduction will be key to maximising AI’s impact in 2025,” writes Núria Emilio. Companies must adopt a proactive approach to infrastructure optimisation to ensure long-term sustainability. This year we can expect to see more LLM models being optimised even further – in some cases this might mean same quality output with models smaller in size or perhaps getting far better output with more curated but smaller training datasets.

‍

Building Readiness for 2025 and Beyond

The trends shaping data and AI in 2025 highlight the need for organisations to balance innovation with readiness. This year we can expect seeing more robust productionised solutions in the making. By investing in the right infrastructure, enabling multidisciplinary collaboration, and addressing ethical and regulatory challenges, companies can unlock the full potential that AI can offer. As Vanessa Cook from Bank of America Institute aptly states, “Gen AI will catalyse an evolution in corporate efficiency and productivity that may transform the global economy.”

For all of us, let 2025 be the year of action and preparation. By focusing on AI readiness, technology leaders can position their organisations for sustained success in an increasingly AI-driven world.

‍