How AI reduce IT team size & where the impact comes from

 Whenever AI and IT teams come up in conversation, I notice the discussion quickly moves to jobs and headcount. How many people will be replaced? Which roles will disappear first?

Personally, I think that discussion misses the bigger point. From what I’ve seen, the biggest impact of AI isn’t just task automation. What really changes things is the way it cuts down handovers, reduces repetitive coordination, and removes a lot of the delays that teams have learned to live with.

If we look at a typical software delivery organization, work is often split across many specialized roles. Business analysts gather requirements, product owners manage backlogs, developers write code, testers validate functionality, DevOps engineers prepare releases, infrastructure teams manage environments, and support teams handle incidents.

There’s nothing fundamentally wrong with that model. But in practice, every handoff creates overhead. Meetings need to happen, tickets get opened, documents get written, and context has to be passed from one group to another. In my experience, a surprising amount of energy goes into moving information around instead of creating real value.

I don’t think AI will make all of that disappear, but I do think it will reduce a lot of it. And once that happens, responsibilities that are currently split across several roles can start to come together in broader, more outcome-focused positions.

The impact will vary across the SDLC. Roles that are highly repetitive, documentation-heavy, process-driven, or focused on information transfer are likely to see the largest reductions.

Below is a simplified view of the potential FTE impact across a medium-sized product or value stream.

SDLC phase	Before AI	With AI	Estimated reduction	Where the impact comes from
Requirements & Discovery	3–4 FTE	1.5–2.5 FTE	35–50%	BA, PO, Process Analyst and Data Analyst activities merge into Product Discovery Lead. AI helps with user stories, meeting summaries, acceptance criteria and process maps.
Planning & Delivery	3–5 FTE	1–2 FTE	50–65%	Scrum Master, PMO Analyst and Project Coordinator work is reduced. AI can generate reports, RAID logs, sprint summaries and dependency views.
Architecture & Technical Design	2–4 FTE	1.5–2.5 FTE	20–35%	AI supports diagrams, options and documentation, but senior judgment remains important.
Development & Testing	9–15 FTE	6–10 FTE	25–40%	Developers, QA Automation and basic testing start to merge into AI Product Engineer and Quality Engineer roles.
DevOps, Release & Infrastructure	5–8 FTE	2–4 FTE	40–60%	Build, release, environment, cloud and sysadmin work moves into Platform Engineer and Release Reliability Engineer roles.
Operations & Support	7–14 FTE	2–5 FTE	50–70%	L1 support, NOC, monitoring and basic incident work are strongly impacted by AI agents and automation.

When we look at the entire delivery chain, the shift becomes even more visible.

View	Before AI	With AI	Overall impact
Medium product stream	29–50 FTE	14–26 FTE	40–50% potential reduction
Enterprise or regulated environment	29–50 FTE	20–35 FTE	20–35% realistic reduction
Mature platform organisation	25–40 FTE	12–22 FTE	40–55% realistic reduction

One thing I think may surprise people is that developers might not be the area where the biggest reduction happens. Yes, coding productivity will improve. But to me, the larger shift may come from everything around the code: reporting, coordination, manual testing, release management, environment setup, support operations, and monitoring.

Testing is a good example. AI can already generate test cases, create test data, run regression scenarios, and spot coverage gaps. I don’t see that as making quality less important. If anything, I see it making quality more embedded in engineering instead of treating it as a separate manual step.

I see the same thing in delivery management. A lot of PMO and coordination work is really about collecting updates, building reports, tracking actions, and stitching information together from different teams. AI can take over much of that administrative effort, which should give delivery leaders more space to focus on risks, dependencies, flow, and stakeholder alignment.

Operations, in my view, could see an even bigger shift. AI agents can already handle common incidents, reset passwords, correlate alerts, summarize issues, and respond to routine support requests. That likely means large L1 support and NOC teams will shrink over time, while the need for people focused on automation, reliability, and AIOps will keep growing.

That’s why I don’t think the future IT organization will simply be a smaller version of what we have today. And personally, I don’t think companies should look at AI only through the lens of headcount reduction.

For me, the real opportunity is to redesign teams around value flow instead of functional silos. The goal should be to remove friction, speed up delivery, and keep knowledge closer to the people actually creating value.