Here is a question worth sitting with before the next sprint planning session: If entry-level work is increasingly being performed by AI, where do the senior engineers of 2035 come from?
Nobody seems to be asking this out loud. Organizations are racing to deploy AI, cutting headcount to show short-term efficiency gains, and declaring transformation. But here is what the data is actually showing—and it is a three-part story that nobody is putting together in one place.
First, the cuts are real, and they are accelerating. Companies explicitly cited AI as the reason for 54,694 planned U.S. job cuts in 2025 alone, according to Challenger, Gray & Christmas, the firm that has tracked layoff announcements for decades. By March 2026, AI had become the single leading stated reason for workforce reductions. These are not projections. These are companies going on record.[1,2,3] Amazon's CEO said it plainly in June 2025: the company would need "fewer people doing some of the jobs" as AI produces efficiency gains, and expected headcount to shrink over the next few years.[4]
Second, the youngest workers are taking the hardest hit. A 2025 Stanford study using ADP payroll data—actual employment records, not a survey—found that employment for workers aged 22 to 25 in AI-exposed roles like software development and customer support dropped 16% since late 2022, while employment in those same roles grew for older, more experienced workers.[7] Stanford economist Erik Brynjolfsson called it the fastest, broadest change he has seen in labor markets. When entry-level jobs disappear, the learning pipeline that produces tomorrow's senior engineers disappears with them. Transformation requires someone on the other side of it who knows what good looks like. That someone does not appear overnight. They grow—through years of doing entry-level work, making mistakes, learning systems from the inside, accumulating the kind of knowledge that never makes it into a Jira ticket.
Third—and this is the part that should stop every leader in their tracks—companies are rehiring the people they let go. Ford Motor Company’s recent quality turnaround highlights the power of adaptive leadership and human-centered innovation. Recognizing that cutting-edge AI reaches its full potential only when paired with deep human expertise, Ford swiftly pivoted its strategy by bringing back 350 veteran engineers to mentor rising talent and fine-tune its digital tools. This agile decision to blend institutional knowledge with advanced technology paid off immediately, propelling Ford to top the J.D. Power Initial Quality Study for the first time since 2010 [5]. Research firm Robert Half found that 29% of organizations that cut staff for AI have already rehired for those exact same positions, at 20 to 35% higher salaries, because now those employees also need to manage the AI.[6]
Three chapters. Fire. Discover the gap. Rehire at a premium. That is not transformation. That is an expensive lesson in what happens when organizations discard the humans who know things before understanding what they are actually discarding.
Scrum practitioners have a name for that knowledge. It is called tacit knowledge. And right now, some organizations are quietly dismantling the pipeline that creates it.
The layoff logic that does not add up
Look at what is happening in the market. Companies are announcing AI strategies and workforce reductions in the same press release. The implicit message is clear: AI will do what people used to do, so fewer people are needed.
There is a version of this that is true. AI is genuinely capable of automating repetitive, rules-based tasks at a scale no team can match. Code generation, test automation, summarization, data analysis—these are real productivity gains that practitioners are experiencing right now.
But there is a version of this that is dangerously incomplete. New hires are not a budget line to eliminate. They are the investment. That college intern picking up their first real ticket, asking questions that seem basic, making mistakes that feel small; they are the senior engineer of 2032. Skip them today and there is nobody to promote in five years. There is nobody who knows the system from the inside when the people who built it are gone. Experience is not something that can be purchased off the street or replicated by a tool. It has to be earned, through time, repetition, and an environment where learning is still possible.
Some organizations already understand this. While much of the industry was cutting entry-level roles, Shopify went the other direction entirely. In 2024, they hired 25 engineering interns. In 2025, they hired 1,000. Farhan Thawar, Shopify's VP of Engineering, was deliberate about why: they wanted people with AI-native thinking, people who grew up with the tools and could help shape how the organization uses them. That is not just an internship program. That is a pipeline strategy.[11]
That history cannot be hired. It can only be grown, protected, and transferred.
When organizations remove entry-level roles in the name of efficiency, they are not becoming leaner. They are consuming their own future.
The pipeline problem nobody is talking about
Here is the part of the AI conversation that gets almost no attention: entry-level jobs are disappearing faster than the workforce can adapt. Revelio Labs, tracking over 100 million job postings, found that U.S. entry-level openings are down more than 35% since January 2023, which is over 100,000 fewer new positions posted every single month.[12] SignalFire's analysis of 600 million employee profiles found a 50% decline in new role starts for people with less than one year of experience between 2019 and 2024, consistent across every business function: engineering, sales, marketing, finance, and operations.[10] A Harvard University study examining 62 million workers across 285,000 firms put it plainly: AI is eroding the bottom rungs of career ladders.[13] The World Economic Forum's Future of Jobs Report 2025—drawing on surveys of over 1,000 employers across 55 economies—found that 40% of employers expect to reduce their workforce in areas where AI can automate tasks.[9][1]
That is not a prediction. It is a stated intention, already in motion.
This matters enormously for the long-term health of every technical and agile team. A junior developer joining a team today learns by doing: picking up tickets, asking questions, pairing with senior engineers, and slowly building the mental model of how things fit together. A junior business analyst learns by sitting in refinement sessions, hearing why requirements get challenged, and watching a product owner navigate stakeholder tension in real time.
If AI absorbs those entry-level tasks entirely, the learning ladder gets pulled up. The junior practitioner has nothing concrete to do. The senior practitioner has no one to mentor. The knowledge transfer that scrum events are specifically designed to facilitate—the daily standup, the retrospective, the sprint review—has fewer participants to pass knowledge between.
Agile works because it creates conditions for continuous learning across a team. Remove the learners and the system stops working—no matter how sophisticated the AI tools are.
This is not an argument against AI adoption. It is an argument for intentional AI adoption. There is a significant difference.
This is not a fringe concern. Harvard Business School professors Amy C. Edmondson and Tomas Chamorro-Premuzic made the same case in their September 2025 Harvard Business Review article, arguing that eliminating entry-level jobs is short-sighted because these roles are crucial for developing future leaders, fostering innovation, and building the organizational knowledge that sustains performance over time. The argument in the research mirrors the one practitioners are living on the ground: the learning pipeline matters, and dismantling it quietly is not a cost saving—it is a liability.[8]
What scrum already knows
The agile community has been working on this problem for decades, long before AI entered the conversation.
The Agile Manifesto values individuals and interactions over processes and tools. Not because tools are unimportant, but because tools without the right people and the right interactions produce the wrong outcomes. A sprint retrospective is not a process formality. It is a structured knowledge transfer event. A sprint review is not a status update. It is a feedback loop that connects the people building the product to the people using it.
These events create the conditions for tacit knowledge to move through an organization. They are the mechanism that prevents expertise from becoming siloed in a single person or lost when that person leaves.
AI adoption done well should strengthen these mechanisms, not bypass them.
Consider what augmentation looks like in practice. A scrum master uses AI-generated sprint data to spot patterns in team performance, then brings those patterns into the retrospective as conversation starters, not conclusions. A product owner uses AI to process large volumes of user feedback before refinement, arriving with clearer themes and more time to explore the nuances with the team. A developer uses AI to handle boilerplate, then spends the time saved pairing with a junior team member on the architecture decision that actually requires judgment.
In every one of these cases, the human is being freed to do the work that only humans can do: think, teach, challenge, decide, and connect.
The question organizations should be asking
Most AI transformation conversations start with the wrong question. The question being asked is: How much of this work can AI do?
The more useful question is: What does this person do best, and what would they be able to do if the repetitive parts of their job were handled for them?
That shift in framing changes everything. It moves the conversation from replacement to investment. It treats existing practitioners as assets to be developed rather than costs to be reduced.
Real transformation is not measured by how many roles were eliminated. It is measured by how much better the remaining people can do the work that machines cannot.
This requires something harder than deploying a tool. It requires deliberate upskilling, time, and space for practitioners to learn new capabilities without abandoning the institutional knowledge they already carry, and leaders who are honest about uncertainty. It requires teams with enough psychological safety to raise concerns when AI output does not look right.
These are all things agile teams already know how to build. The frameworks exist. The ceremonies exist. The principles exist. The challenge is applying them with the same rigor to AI adoption that practitioners apply to software delivery.
What sustainable AI adoption looks like
Organizations getting this right share characteristics recognizable to any experienced agile practitioner.
They protect the learning pipeline. Junior practitioners still have meaningful work. AI handles the volume; humans handle the judgment. The mentorship relationship stays intact because senior practitioners are still actively teaching, not just supervising AI output.
They treat knowledge transfer as a sprint commitment, not an afterthought. Retrospectives explicitly surface where knowledge is concentrated in too few people. Documentation sprints happen before workforce changes, not after.
They define the human-AI boundary clearly. Every AI-assisted workflow has a documented answer to: at what point does AI output require human review before moving forward? That boundary is agreed upon by the team and revisited as the technology evolves.
They measure what actually matters. Efficiency metrics go up with AI adoption—that almost always happens. Sustainable teams also track quality over time and team capability growth. A team that ships faster but learns nothing is not performing better. It is running down a clock.
The answer was already here
The anxiety around AI is real. The technology is moving fast, the workforce implications are significant, and the gap between what organizations are promising and what they actually understand is wide.
But for agile practitioners, this moment does not require a new framework. It requires applying the one that already exists.
Center the people. Protect the learning. Build in feedback loops. Be honest about what is known and what is not. Treat knowledge as a team asset, not an individual one.
The scrum master in the room is not becoming obsolete. In many ways, the role has never mattered more.
The organizations that will look back on this period as genuine transformation are the ones that augmented their people. The ones that replaced them will spend the next decade figuring out what they lost.
References
[1] Challenger, Gray & Christmas. (2026, January 8). 2025 Year-End Job Cut Report: Highest Q4 Layoffs Since 2008. https://www.challengergray.com/blog/2025-year-end-challenger-report-highest-q4-layoffs-since-2008-lowest-ytd-hiring-since-2010/
[2] Challenger, Gray & Christmas. (2026, April 2). Challenger Report: March cuts rise 25% from February, AI leads reasons. https://www.challengergray.com/blog/challenger-report-march-cuts-rise-25-from-february-ai-leads-reasons/
[3] Iacurci, G. (2025, December 21). AI was responsible for almost 55,000 layoffs in the U.S. in 2025. CNBC. https://www.cnbc.com/2025/12/21/ai-job-cuts-amazon-microsoft-and-more-cite-ai-for-2025-layoffs.html
[4] Leswing, K. (2025, June 30). Amazon CEO says AI will mean fewer people doing some of the jobs. CNBC. https://www.cnbc.com/2025/06/30/amazon-ceo-says-ai-will-mean-fewer-people-do-jobs-that-get-automated.html
[5] Wayland, M. (2026, June 25). Ford has been rehiring quality inspectors after AI fell short. Bloomberg. https://www.bloomberg.com/news/articles/2026-06-25/ford-has-been-rehiring-quality-inspectors-after-ai-fell-short
[6] Blanco, O. (2026, June 5). Why some companies are rehiring employees they laid off due to AI. Fast Company. https://www.fastcompany.com/91554983/ai-boomerang-why-some-companies-are-rehiring-employees-they-laid-off
[7] Fried, I. (2025, August 26). AI is already taking jobs away from entry-level workers. Axios. https://www.axios.com/2025/08/26/ai-entry-level-jobs
[8] Edmondson, A. C., & Chamorro-Premuzic, T. (2025, September 16). The perils of using AI to replace entry-level jobs. Harvard Business Review. https://hbr.org/2025/09/the-perils-of-using-ai-to-replace-entry-level-jobs
[9] Leopold, T. (2025, April 30). How AI is reshaping the career ladder. World Economic Forum. https://www.weforum.org/stories/2025/04/ai-jobs-international-workers-day/
[10] Doshay, H., & Bantock, A. (2025, September 7). AI is not just ending entry-level jobs. It's the end of the career ladder as we know it. CNBC. https://www.cnbc.com/2025/09/07/ai-entry-level-jobs-hiring-careers.html
[11] CoderPad. (2026, April 23). In the AI era, Shopify is investing in junior engineers — not cutting them. https://coderpad.io/blog/hiring-developers/in-the-ai-era-shopify-is-investing-in-junior-engineers-not-cutting-them/
[12] Revelio Labs. (2025, August 4). Is AI responsible for the rise in entry-level unemployment? https://www.reveliolabs.com/news/macro/is-ai-responsible-for-the-rise-in-entry-level-unemployment/
[13] Hosseini, S. M., & Lichtinger, G. (2026, May). Generative AI as seniority-biased technological change: Evidence from U.S. résumé and job posting data. Harvard University working paper. Covered in: HRD Australia (2026, May 19). The entry-level job is disappearing. https://www.hcamag.com/au/specialisation/hr-technology/the-entry-level-job-is-disappearing/575695
--
A note from Scrum Alliance:
Ready to guide your agile team through AI adoption without losing human judgment and decision-making?
Implementing AI responsibly requires agile mindsets, intentional human-AI collaboration, and continuous learning. Explore Scrum Alliance's AI & Emerging Practices courses to gain practical skills for blending AI efficiency with proven agile principles.
About the author
