Every role has a next chapter.
a map for the new org chart
Every role on the chart has a next chapter
Gartner's new org chart merges twelve roles into six, and six into three teams. It reads like a list of cuts. It's a map: for each role, what AI takes, what stays human, and the first skill to learn.
Look at the left column of Gartner's new org chart and you'll find most of today's tech job titles. Product manager. QA tester. Business analyst. Scrum master. Twelve of them, merging into six, rolling up into three new teams.
It's easy to read as a list of jobs going away. I read it as a map.
Every arrow is a change in what someone does all day. Gartner is explicit that it isn't a change in how many people do it: "not a reduction of overall headcount." Three in four software engineering leaders expect their teams to hold or grow.
What changes in each role
- AI takes
- First drafts of code and docs
- Stays human
- Deciding what to build, and why
- AI takes
- Variants and test cases
- Stays human
- Knowing what good looks like
- AI takes
- Meeting notes, first-draft stories
- Stays human
- Knowing how the business really works
- AI takes
- Boilerplate integration
- Stays human
- Sitting with the customer
- AI takes
- Routine runbooks
- Stays human
- Guardrails and reliability
- AI takes
- Model plumbing
- Stays human
- Proving it actually worked
Read the cards one after another and the same thing happens in every one.
That's why the business analyst card is my favourite. In OpenAI's GDPval study, the best model matched or beat industry experts on 47.6% of tasks, and those tasks were precisely specified up front. The spec is what made the output. Whoever writes it well becomes the multiplier.
None of this is only about software. In a health plan it's the claims analyst who writes the rules an agent applies. In a bank, the credit analyst who decides what an agent should flag.
Ninety days to put a team on the map
- After week 3, everyone can say in a sentence what they'll do that AI won't.
- After week 8, one thing has shipped from a new pairing: a product engineer with an AI requirements analyst, say.
- After week 12, you have cycle time and one quality number, before and after, written down.
I don't know how fast this lands. Gartner's own date is 2029, and predictions about org charts age badly. The direction is clear enough to plan for, though, and Gartner expects 80% of engineers to need new skills by 2027 either way.
AI does more of the making. People do more of the deciding.
Which card is your team furthest from?
- CIO Dive — Gartner: engineering teams shrink as AI takes on responsibilities (10 Jul 2026)
- PwC 2026 Global AI Jobs Barometer (15 Jun 2026)
- OpenAI — GDPval: evaluating AI on real-world, economically valuable tasks (Oct 2025)
- CDO Magazine — Gartner: 80% of software engineers must upskill through 2027
- Microsoft Work Trend Index 2025 — the year the Frontier Firm is born
Every figure here describes something measured and committed. See the measurements · read the method