The Three Legs Every CTO and CDO Stands On When Implementing Data and AI

Many organisations today possess much more data than they can make sense of. The potential for artificial intelligence, data science, and machine learning to convert this data into value in operations is considerable. However, the journey from this potential to business outcomes with measurable effect is tricky.

Following the roadmap I discussed in my last article on things I learned in the trenches, in this piece I will touch on something professionals coming from a software development background face when they take over as CDOs or CTOs. Many technology leaders have developed software their whole careers without realising how nuanced and specific data projects are. Now as AI and data are blobbing their way onto the technology field, these leaders often ask: “What should I focus on to support a true data-driven push?”

As many of my colleagues, friends, and readers know, my work experience in data transformations across various sectors has shown that success depends not only on technology. Strategy, people, and culture are huge factors. Today I will discuss the three legs of the stool leaders will likely be standing on when implementing data products and AI for large organisations.

The First Leg: Data

Some organisations fall into the trap of trying to “catalog the universe” before even starting to engage the potential use cases and business value they could obtain from becoming data-oriented.

Collecting, transforming, storing, and retrieving data are complex, costly, and messy activities. Redundant types and often contradictory data, collected by different methodologies and under different definitions, are the norm and a main source of cost, confusion, and inefficiency. In large companies, data integration projects frequently face difficulties from data formats that differ and conflicting standards, especially when interacting with systems from an earlier time.

Therefore, establishing data quality, data consistency, and implementing governance with depth and breadth is a foundation step. Setting benchmarks for data quality, fully sanctioned and cared for by responsible data owners, guarantees the organisation has data that can be trusted by anyone. The technology area must establish and implement governance mechanisms that operate with automation, including data lineage tracking and data dictionaries with detail that is comprehensive, and assist in maintaining quality when operating at scale. These systems aid compliance with regulations such as GDPR or HIPAA.

This is such an important and underappreciated basic part of the job.

The Second Leg: Infrastructure Foundation

As the quest for taming the data deluge goes on, in my experience a continuous assessment and strengthening of the technological infrastructure that supports the whole thing must be carried on.

The scope of this work is wide, encompassing the various IT platforms that support both current business operations as well as the ensuing data applications. From data analysis, ML development, and automation to AI agents integration, those applications inevitably add stress onto the overall infrastructure responsible for running apps, legacy systems, ERPs, and every aspect of a company’s business.

I can tell you that you will be responsible for maintaining and improving what currently exists and works, while expanding it to perform crucial data-driven activities like collecting old and new data, cleaning, organising, development, and testing. Once this is all over, then the infrastructure will consume, deploy, and support all the delicious innovations and AI trinkets we love.

Hence, the infrastructure foundation shall support a multi-structure and full-cycle data usage designed for AI automation, which will make the analysis and suggest the decisions (or even make the decisions autonomously). This is a game changer as the analytical environment has to be fully integrated with the operations environment, to measure, evaluate, and improve the consequences of those decisions.

Most companies are not even there yet. Most still have in mind the analytical part of that data-driven arch: creating dashboards to support “better-informed decisions”. But even in those, the game has moved from ingestion and dashboards to complex integrations, particularly with platforms from earlier eras, embracing new environments, flexible designs, modular components, and innovative architectures.

I favour architectures built from interchangeable components and solutions designed as a service for their interoperability and capacity for growth. This allows component reuse without waste, avoiding rework that incurs high expense. This discussion is far beyond the familiar on-premise versus cloud. It is about how to best use multiple environments to construct a data backbone that shows resilience as businesses expand and demand scaling infrastructure capabilities.

The Third Leg: Implementation, Innovation, Scale, and Data Products

With data foundation and infrastructure progressing, attention moves to balancing innovation pursuit with steadiness in operations. It is expected that Chief Technology Officers guide digital transformation while keeping the lights on and the business operating seamlessly. This is a balance act that requires a keen eye on the future of the industry and the company, slicing technology adoption in stages, and non-stop process optimisations.

However, this third leg is about something completely different from the technical topics covered so far. This is about dealing with the consequences of a successful journey until this point. This third leg is about developing people so they can reap the fruits of organised and comprehensive data and a powerful technology stack capable of making their most innovative business idea come true.

This is where I saw so many companies stumble.

At this point, companies are excited about the success and novelty of the technologies introduced, the power of data, the workshops, the internal stakeholder engagement, the encouraging results of MVPs and POCs. That is when internal demand explodes as every single department in the organisation wants to jump into the bandwagon and get the benefits of this data-driven digital transformation process.

It is time to scale from a central data function whose job is to carefully curate data and skilfully develop data products, to an open and wide full-scale transformation where data assets are tailored for, or adapted by, individual business departments to new requirements, under standard frameworks of common governance.

While working on this leg, CTOs and CDOs are required to provide training, keep a watchful eye for compliance and cybersecurity, support data and software engineers to learn non-technical skills (from agile and customer centricity to empathy and text interpretation), while providing salespeople, lawyers, accountants, medical doctors, surgeons, marketing experts, and cooks the technical training they need to develop their own data products and AI agents, in a responsible and effective manner.

In my experience those three legs must not be built apart from each other, but all at once, one supporting the other. Often there will be different maturity levels for each of them and that is fine. Though they are not the only aspects that guarantee the success of data projects, I hope that if you step up to a leadership role in data and technology, you will find these thoughts useful.

The stage is set for data democratisation, data literacy, people engagement, data security, and strategic alliances to come into focus.