Data Architecture in the Age of AI
2025 is turning out to be a watershed year in many respects. While the political turmoil is perhaps the story getting the most attention, the transformation of the workplace through the adoption of ever-more powerful AI tools may in fact have the more significant long-term impact (on both the economy and society). There is no sector completely immune from these impacts, but today I’m going to address the IT sector specifically and then zoom in on one of the roles that’s being impacted from the perspective of someone who has worked in such a role off and off for more than 25 years.
AI & IT
The early to mid-1990’s represented the beginning of the current IT boom. Previous to that explosion of new roles and jobs, things had been pretty stable for what was then referred to as Management Information Systems (MIS). Nearly all of the work was done on Mainframe computers with relatively little access to systems outside the strict confines of IT departments. The Mainframe server rooms were the early version of what later became Data Centers (and is now the Cloud).
The underlying premise behind the majority of innovations associated with Information Technology (IT) has been focused around various forms of systems-based automation. For the past 50 plus years or so, each new wave of IT innovation has led both to greater automation as well as to the creation of new job opportunities and workplace roles associated with those new technologies. The cycle was never meant to last, though – in other words – if the ultimate goal is total or near-total automation, then eventually, jobs and workplace roles would be begin to disappear. In 2025, after many had been predicting this outcome for decades; it's finally beginning to happen.
This phenomenon is manifesting itself in several tangible ways:
1. Multiple rounds of layoffs in Big Tech (companies). This is occurring even in cases where the companies are making large profits. Typically, the companies in question don’t overtly announce that the layoffs are associated with the replacement of personnel with AI tools, but in some cases it has been. The exact number of layoffs since 2023 is hard to track and subdividing it even more specifically into AI-driven layoffs is even more difficult – but the likely figure associated with AI replacement in IT is now in the 100’s of thousands.
2. Targeted job replacements (across all industries). The areas where AI replacement is most prevalent across industries at the moment are associated with Software Engineering, Customer Service and other low to mid-level management and administration roles. Other targeted areas though include writers and other creative personnel across multiple industries.
3. The job market for new grads is cratering. This may seem counter-intuitive at first, given that the new grads tend to have the most / best AI-related skills; however, most of this demographic initially fills lower-level or junior positions and those are precisely the ones being targeted first for replacement.
Obviously, these trends when combined will radically change the hiring landscape for both IT and potentially all other White-Collar jobs. It is quite likely that within the next 5 years, the IT industry will shed millions of jobs as AI replaces an ever-greater percentage of positions and even roles. How much will be replaced in that timeframe is difficult to predict, but it could be anywhere from 25-50% of all current IT jobs and a like number of roles (that will vanish or come close to vanishing). For years, young people were told to study STEM curriculums, learn to code and focus on IT as a stable career – this is all about to change if it hasn’t already. As recently as last year, I was still advocating to all the young people I knew to consider a career in Information Technology – as of this year, I’m no longer making such recommendations. The IT Boom is over. Besides changing the landscape for young people just starting their careers, this trend will impact the jobs & careers of nearly everyone already in the field. Those impacts will include (or already include):
1. Fewer positions with more competition (across the board).
2. Lower salaries (and / or contract rates).
3. The elimination of some job roles entirely.
4. Few, if any, new replacement roles in IT to shift to.
AI & Data Architecture
Data Architecture is a relatively new practice within the larger field of IT. While there was always an aspect of data design associated with system development and maintenance, there were relatively few people who likely considered themselves to be “Data Architects” until sometime around the mid-1990s. The change that occurred during the 1990s that led to an increased need for such a role was related to a proliferation of data sources, new standards and new tools that hit the market (and workplace) more or less at the same time. In other words, the field became sufficiently complex (over a short timespan) to justify a number of additional roles for specialized data-focused roles - including a “Data Architect” role. Ever since then, Data Architecture and the Data Architect have played a prominent part within most mid-sized and large enterprises. And this was only one of perhaps more than two dozen new roles introduced around the same time in relation to IT innovations being adopted in the workplace.
What is a Data Architect anyway?
We’ll go with a top-level definition here: “A Data Architect is someone who is charged with oversight of the entire data footprint associated with an organization or is charged with some defined portion or data domain within that enterprise (in a very large organizations). The Data Architect is typically involved in Data Strategy, Data Design, Data Governance and sometimes data development and helps to ensure that the organization gains the most value possible from their data resources.” This definition is necessarily a bit broad, but the idea is that this role allows one person to act as the Point of Contact (PoC) across many systems, data issues and groups within an organization. The exact mix of associated data roles depends on the size of the organization; but from an industry perspective, there are probably in the neighborhood of around two dozen additional data-specific roles that are often used in a typical enterprise. This covers everyone from Database Engineers, to ETL development and BI report designers, etc. Several years ago, there was a surge of “Chief Data Officers” or CDOs being hired across industry, but this has since tapered off. A CDO (role) is really an extension of the Data Architect role with more focus on strategy and less on design.
How is AI Impacting Data Architecture?
Notice I didn’t say “how will…” – that’s because it’s already happening. AI is changing the practice Data Architecture in real time. This is happening from several perspectives:
1. Analysis & Data Augmentation
2. Data DevOps Automation
3. Introduction of new AI-driven capability
The Data Architect is involved in each of these areas as both the main advocate and key oversight resource. We’ll look at each of these in turn.
Data Analysis – Historically, this has been an onerous task in large enterprises with highly complex systems and / or lot’s of systems to deal with. In this context, AI empowers existing roles w/o total replacement and makes completion of these tasks more likely and the resulting designs more effective. Data Design both within systems and across systems is becoming increasingly automated, though and this will reduce the need for Data Modelers & Database Engineers over time. While the Data Architect is sometimes involved in directing or participating in data design and definition, less of this will be required now due to AI.
Data DevOps Automation – This spans various areas of development with a large focus on testing and configuration alignment. Much of this can now be automated and less oversight will be required here than has been required historically.
New AI-driven Capability – At first, this represents additional work for the Data Architect as he/she will be responsible for determining which AI capabilities to adopt and how to integrate / rationalize them within the enterprise. Later, some of this work will be done by the AI’s themselves. In this context, Data Architects might also be considered AI Architects given the extraordinary amount of overlap that will occur (but only for a limited time).
Ultimately, AI tools will allow Data Architects to accomplish more than they did before, but over time their duties will be diminished and begin to mirror more of the previous CDO role that has largely vanished. As Architect / CDO, the main focus will likely become more centered around interaction with Business leadership to allow them to do some data work directly and also continuing the enterprise oversight role with fewer human roles supporting – but still remaining answerable to leadership for the full range of activities & capabilities in the enterprise associated with data.
This view covers the next five years or so for Data Architects. Things are likely to change further when looking 5 to 10 years out. It’s entirely possible that all IT Architecture-related roles may disappear somewhere around 10 years out. Why? Because the extreme levels of automation being introduced are eliminating the very complexity that required IT Architects in the first place. This prediction assumes that no serious attempt from an industry perspective to save certain job roles will happen (in order to ensure that humans “stay in the loop”), which certainly seems to be the path that we’re on now.
Copyright 2025, Stephen Lahanas