The Four-Question Role Screen
The AI-and-jobs question lands on every operator’s desk eventually. A board member asks. An investor asks. The CFO asks. The pressure is the same whether you run a company of five people or fifty thousand. Which jobs here are actually at real risk, and what should we do about it.
Most of the tools people reach for to answer that question are wrong. The popular ones count which tasks could be done by AI and add up the percentages. That math doesn’t tell you which roles in your specific company are in trouble. Consulting decks aren’t much better. They list percentages by function and call it analysis.
This page is the alternative. A four-question scoring tool. Twenty minutes per role. Run by your own people, not outsiders. The output is a clear read on which roles are at meaningful displacement risk in the next three to five years, which need augmentation investment now, and which are fine for the moment. The original version of this screen ran on LinkedIn in April 2026. Since then, different operators have applied it to their companies. One was kind enough to give me permission to share this broadly. He analyzed twenty-five roles inside a typical mid-market US ground brokerage in the freight brokerage industry and shared the full matrix with permission to publish the scoring and the insights. That worked example is below.
This is not a strategy framework or a transformation initiative. It is a working tool for management teams trying to make good capital and people decisions while the noise around AI and jobs is louder than the signal.
Published April 2026. Harry Glorikian.
What this is
A four-question scoring tool. Pick a role inside your company. Score it on the four questions below. Place it in one of three zones. Red means transition plan needed. Yellow means augmentation investment. Green means rerun in six to twelve months.
The screen is mine. The underlying economics are not. Both matter. Credit details at the bottom.
How to run it
Pick a role. Score each question one to five. Add the four numbers. Read the zone.
Twenty minutes per role. The right people in the room are the ones with real knowledge of the work and real responsibility for the cost. At a large company that’s the CHRO, the CFO, and a board or audit committee member. At a fifty-person company it’s the founder, the head of operations, and whoever owns the budget. At a five-person company it’s a Saturday morning and a whiteboard.
Don’t hire outsiders to do this. Do it yourself if you want tangible insights.
Why this screen instead of an exposure index
The smartest people writing about AI and jobs disagree with each other at a fundamental level. Once I started reading past the headlines, the reason became clear. They aren’t answering the same question.
Dario Amodei runs Anthropic. He’s inside the lab. He says entry-level white-collar work gets hit hard and soon. Daron Acemoglu won the Nobel in 2024 and pegs the total US productivity impact from AI at well under one percent of GDP over ten years, a rounding error at the macro level. A Microsoft researcher I interviewed who is also a trained anthropologist reminded me that in two hundred years of industrial history, technology has never net eliminated the jobs humans evolve into. Andrej Karpathy, also close to the frontier, built a viral dashboard ranking 342 US occupations by AI exposure, then pulled the source code down when it went sideways on X, because exposure was never the same thing as displacement.
Many serious camps. Many different altitudes. Many different answers.
The exposure-index approach is the one that gets the most airtime, and it’s the one that breaks first under scrutiny. Exposure scores are built on the assumption that tasks are separable. Automate task A, task B is unaffected. Add up the percentage of tasks that can be automated, get an exposure number. Joshua Gans and Avi Goldfarb at the Rotman School of Management showed in their January 2026 NBER paper that this approach is mathematically inconsistent with how actual production works. Tasks complement each other. Productivity is multiplicative, not additive. Small quality improvements on the remaining tasks multiply through the entire job. When AI absorbs some of the tasks in a high-dimensionality job, the worker reallocates to what’s left and becomes more valuable, not less. Gans and Goldfarb call this the focus effect.
Once you see that, the disagreements start to make sense.
Amodei is probably right about what the technology will do in the next thirty-six months for the workers in low-dimensionality jobs. The anthropologist is probably right for the workers in high-dimensionality jobs, where the focus effect raises wages instead of cutting headcount. Acemoglu is probably right on the aggregate macro number. They’re describing different parts of the same animal.
That means the macro debate about whether AI destroys jobs in the aggregate is, for anyone with fiduciary responsibility, a distraction from the question in front of us. The question is which specific roles inside which specific companies. Exposure indices don’t answer that. Three variables do: dimensionality of the role, demand elasticity for the output, and the firm’s incentive to eliminate the full wage bill. Add competitive pressure and you have the screen.
Picking sides between Amodei and Acemoglu was the wrong move. Picking variables is the right one.
The four questions
1. Dimensionality
How few essential tasks make up this job?
One essential task is a five. Two or three is a four. Four to six is a three. Seven to ten is a two. More than ten is a one.
A job built around one thing is the job most at risk of being eliminated whole. A job built around many complementary tasks is the job where AI absorbs some of them and the worker becomes more valuable on what’s left.
2. Demand elasticity
What happens to demand if the price of this role’s output drops thirty percent?
Translation: if this role’s work becomes thirty percent cheaper to deliver, do your customers want more of it, the same amount, or less.
Flat or declining demand is a five. Stable is a four. Modest growth is a three. Meaningful growth is a two. Explosive latent demand is a one.
If your customers won’t buy more when the product gets cheaper, productivity gains turn into layoffs. Software developers grew when the cloud made them more productive. Bank tellers didn’t, when ATMs and mobile banking made the branch optional. Bank teller employment grew from roughly 300,000 in 1970 to over 600,000 by the early 2000s as ATMs cut the cost of opening a branch and banks opened 43 percent more of them. Then mobile banking arrived and teller employment has been declining since. The Bureau of Labor Statistics now projects another 13 percent decline by 2034.
3. ROI to full automation
How clear and large is the ROI for automating this role completely?
Both clarity and size matter. A five requires a clean integration path and recovery of the full wage bill.
Full wage bill savings with a clear integration path is a five. A strong partial business case is a four. A muddy partial case is a three. Only possible piece by piece is a two. Technically infeasible within five years is a one.
The sharper the total ROI, the harder capital pushes. Automating one task inside a twenty-task job rarely justifies a major integration project. Automating the last manual task inside a two-task job often does.
4. Competitive pressure
How fast is competitive pressure forcing the move?
The anchors below describe what fast looks like in your competitive landscape.
Competitors already automating at scale is a five. Competitors piloting publicly is a four. A credible new entrant building around it is a three. Nobody has moved is a two. Protected by regulation or credentialing is a one.
If you don’t automate and your competitor does, your cost structure gives you about eighteen months before you become uninvestable.
Scoring and zones
Add the four numbers. Range is 4 to 20.
16 or higher: Red zone. Transition plan needed ASAP. The role is at meaningful risk of being eliminated within three to five years.
12 to 15: Yellow zone. Augmentation investment. Tools, training, role redesign so humans spend their time on the parts AI can’t touch.
Below 12: Green zone. Rerun every six to twelve months. Scores move.
If everything comes up red, you’re scoring for advocacy. If everything comes up green, you’re not paying attention.
Two workers
A management consultant. Roughly eight things in the job. Research, data analysis, client communication, slide construction, strategic reasoning, team coordination, relationship management, implementation support. AI helps with most of them. By any exposure index you can find, her job looks highly exposed.
A long-haul truck driver. Roughly one thing. Move the truck safely from point A to point B. Logistics, loading, dispatch are someone else’s jobs. By any exposure index, his job looks lightly exposed because he isn’t the one using AI day to day.
Which one is more likely to lose her or his job in the next three to five years?
The truck driver. It isn’t close.
The consultant scores low on dimensionality and low on ROI to full automation, because no firm is going to spend integration money to automate one task inside an eight-task job. The driver scores high on both, because the entire wage bill is on the table when one task is automated, and Aurora and Kodiak are already running commercially.
The data backs the consultant case up. AmLaw 100 revenue grew about 13 percent in each of the last two years and lawyer headcount grew 7.7 percent in 2024. Radiologist compensation rose 9 percent in 2025 to an average of $571,000. The American Medical Association reported physician AI use jumped from 38 percent in 2023 to 81 percent in early 2026. None of those professions is shrinking. The technology is absorbing tasks, not eliminating jobs.
The high-exposure worker keeps her job and may even get a raise. The low-exposure worker loses his.
A worked example: freight brokerage
Ram Dhan Yadav Katamaraja, founder of Colaberry, who is building AIXFreight, an invisible interface for the freight brokerage industry, applied the screen to twenty-five roles commonly found in and around a typical mid-market US ground brokerage. Mid-market here is $5M to $250M in revenue. A freight brokerage matches shippers who have goods to move with trucking carriers who move them. They don’t own trucks. They make money on the spread between what the shipper pays and what the carrier charges. He shared his full matrix with me, and I am sharing it here with you, with all the permissions to do so.
Twenty-five roles. Eight red. Eight yellow. Nine green.
| # | Role | Essential Tasks | Dim | Elast | ROI | Comp | Total | Zone |
|---|---|---|---|---|---|---|---|---|
| 1 | Load coordinator / ops rep | Inbox triage, RFQ extraction, rate lookup, quote send, order entry, follow-up | 3 | 3 | 4 | 4 | 14 | Yellow |
| 2 | Track & trace / check-call rep | Call/text driver, update TMS, email customer | 5 | 4 | 5 | 5 | 19 | Red |
| 3 | Dispatcher | Carrier assignment, appt setting, exception handling, driver comms, TMS updates | 3 | 3 | 3 | 3 | 12 | Yellow |
| 4 | Carrier sales rep | Load posting, carrier outreach, rate negotiation, rate con, booking | 3 | 3 | 4 | 4 | 14 | Yellow |
| 5 | Customer sales / BD rep | Prospecting, qualifying, pitching, negotiating, relationships, QBRs, contracts, expansion | 2 | 2 | 2 | 3 | 9 | Green |
| 6 | Billing / AR clerk | Invoice generation, POD/BOL validation, customer billing, dispute intake | 4 | 4 | 5 | 4 | 17 | Red |
| 7 | Carrier settlement / AP clerk | Carrier invoice processing, rate con match, payment exec, quick-pay, factoring recon | 3 | 4 | 5 | 4 | 16 | Red |
| 8 | Claims / disputes coordinator | Claim intake, documentation gathering, carrier/shipper comms, resolution, GL filing | 3 | 4 | 3 | 2 | 12 | Yellow |
| 9 | Pricing analyst (rate desk) | Market analysis, customer rules, lane pricing, RFP responses | 4 | 3 | 4 | 4 | 15 | Yellow (high) |
| 10 | QA / post-load auditor | Audit load data, flag discrepancies, verify charges, report | 4 | 4 | 5 | 3 | 16 | Red |
| 11 | Operations manager | Supervision, escalation, KPI mgmt, coaching, customer escalation, process, reporting, hiring | 2 | 2 | 2 | 2 | 8 | Green |
| 12 | Broker-agent / branch manager | P&L, customer relationships, strategic pricing, team leadership, BD, vendor mgmt, compliance oversight | 2 | 2 | 1 | 1 | 6 | Green |
| 13 | Compliance officer | Carrier vetting, insurance check, CSA monitoring, denied-party screening, DOT compliance | 3 | 4 | 5 | 5 | 17 | Red |
| 14 | Long-haul OTR driver | Move truck A→B safely | 5 | 4 | 4 | 4 | 17 | Red |
| 15 | Regional / local driver | Drive, multi-stop delivery, customer handoff, dock exceptions, paperwork | 3 | 4 | 2 | 2 | 11 | Green |
| 16 | Owner-operator | Drive, dispatch, billing, compliance, maintenance, taxes, customer relationships | 2 | 4 | 2 | 3 | 11 | Green* |
| 17 | Fleet dispatcher (carrier-side) | Driver assignment, HOS mgmt, customer comms, appt setting, exception handling | 3 | 3 | 3 | 3 | 12 | Yellow |
| 18 | Safety / compliance manager (carrier) | FMCSA filings, DOT audits, DQ files, CSA mgmt, incident investigation, training | 2 | 2 | 2 | 2 | 8 | Green |
| 19 | Carrier back-office clerk | Invoice generation, IFTA, permits, factoring submission, payment tracking | 3 | 4 | 5 | 4 | 16 | Red |
| 20 | Transportation coordinator (shipper) | Carrier selection, tendering, spend analysis, routing, performance tracking | 3 | 3 | 3 | 3 | 12 | Yellow |
| 21 | Freight pay / audit clerk (shipper) | Invoice audit, payment approval, accrual, GL coding | 4 | 4 | 5 | 5 | 18 | Red |
| 22 | Logistics manager (shipper) | Carrier RFPs, relationships, network design, budget, team mgmt, sourcing | 2 | 2 | 2 | 2 | 8 | Green |
| 23 | Customs broker / trade compliance | HTS classification, entry filing, denied-party screening, duty calc, regulatory filing, client advisory | 3 | 3 | 3 | 3 | 12 | Yellow |
| 24 | EDI / integration analyst | Map EDI connections, troubleshoot feeds, onboard partners, monitor, custom integrations | 3 | 1 | 2 | 2 | 8 | Green |
| 25 | Freight forwarder coordinator | Booking, documentation, customs coordination, tracking, client comms, regulatory | 2 | 3 | 2 | 3 | 10 | Green |
*Owner-operator scored Green but flagged structurally threatened. Matrix and analysis by Ram Dhan Yadav Katamaraja, Colaberry Inc., used with permission.
Four of the cases are worth pulling out.
Track & trace at 19. One essential task. Get status, relay status. Stable shipper demand. Mature voice and messaging AI vendors already in market. The matrix calls this the single cleanest displacement case in brokerage. A typical broker runs three to eight reps per shift at $45K to $65K loaded.
Billing / AR clerk at 17. Deterministic work against validated documents. Clear path to integrate with the systems brokerages already use to manage loads and run their books. The matrix cites a real case, called CES, where four roles were eliminated and roughly $240,000 a year, or $22 per load, came out of labor cost. AI-native entrants are already moving in the category at the enterprise tier.
EDI analyst at 8. EDI stands for electronic data interchange, the plumbing that lets one company’s systems talk to another’s. This role scored a 1 on demand elasticity, the only role in the matrix with that score. Green not because AI can’t touch the work but because demand for integration work explodes precisely as AI platforms proliferate. Every new AI overlay needs another integration. Every new logistics platform needs another connector. The more successful the category, the more this role grows.
Long-haul driver at 17. Two years ago that role was yellow. Aurora ran ten driverless trucks commercially in December 2025 and is targeting more than 200 by year end. Kodiak’s Atlas Energy deployment in the Permian is running up to twenty-four hours a day. Yellow to red in twenty-four months. Exactly why the screen says rerun every six to twelve.
The matrix author also flags two observations that change how a reader should use the screen.
Score the elasticity question at the right level. A broker that automates track-and-trace and billing wins share. Faster and cheaper than the broker down the street. Green-flag growth story at the firm level. But brokerage as an industry sits on top of physical goods movement, and physical goods movement is inelastic in aggregate. Individual brokers grow load volume through share gains, while total broker ops headcount across the industry shrinks. Cloud demand was elastic when it arrived, and developer headcount expanded. Mobile banking demand was inelastic, and teller headcount collapsed. Both stories are true at the same time. The same role scores differently at the firm level versus the industry level. Score both. The gap is the part most management teams miss.
The distribution tells you who buys, who gets augmented, and whose wage bill is on the table. Green-zone roles are the operations manager, the branch manager, the logistics manager, the carrier-side safety manager. Those are the people who buy the product. Red-zone roles are whose wage bill gets eliminated. Yellow-zone roles are whose jobs change shape. A pitch that treats those three groups the same fails. The load coordinator who hears “we’re going to replace you” kills the pilot. The same load coordinator who hears “we’re going to take the inbox grind off your desk so you can focus on exceptions” champions it. That distinction isn’t marketing spin. It’s what the focus effect predicts for yellow-zone roles, and it’s what makes the screen useful as segmentation rather than a displacement threat.
The matrix author also flags two roles to watch. Regional driver could move from green to yellow within three years on the same trajectory long-haul just took. Customs broker could move from yellow to red if cross-border licensing loosens.
What this matrix shows is what a careful operator can produce in one industry with the screen and a few hours of work. The distribution is defensible. The variance tracks what someone with deep freight knowledge would tell you. The cases pull in real vendors, real wage data, real numbers. That’s useful insight for any board or management team thinking about their own roles. Run it on your own industry and you’ll get the same kind of result.
One more thing from the matrix. Don’t deliver it. Co-produce it. Run the four questions against the prospect’s own org chart on the whiteboard and let them reach the red-zone conclusion themselves. The variance is what creates the urgency. A pre-cooked answer doesn’t.
On scoring discipline. Scores reflect a typical mid-market freight brokerage in the $5M to $250M revenue range. Bigger logistics companies that own warehouses or trucks would score differently. Role specialization raises dimensionality scores. Regulatory accountability lowers the ROI-to-full-automation score. Smaller brokerages under $5M often consolidate multiple roles into single seats, which raises effective displacement risk because automating one consolidated seat captures multiple wage bills. Score for the company you have, not the company in the org chart template.
What the screen is not
Not a forecast. It’s a structured way for a management team to think, built on a January 2026 NBER working paper and a March 2026 Substack synthesis.
Not a substitute for industry expertise. Someone who knows the industry will score more accurately than someone who doesn’t.
Not a one-time exercise. Scores move. Long-haul truck driver was yellow two years ago and red today. Rerun the screen every six to twelve months.
Not a replacement for human judgment about specific roles, specific people, or specific transition plans. It ranks risk. It doesn’t make decisions about individual employees.
Credit
The core economics in the screen come from Joshua Gans and Avi Goldfarb, “O-Ring Automation,” NBER Working Paper 34639, January 2026, which itself builds on Michael Kremer’s 1993 “O-Ring Theory of Economic Development.”
Alex Imas (University of Chicago Booth) and Soumitra Shukla (Harvard Business School and Burning Glass Institute) brought the paper into wider circulation with their March 2026 Substack post “How Will AI-driven Automation Actually Affect Jobs?” The two-workers framing and the dimensionality language are theirs.
The four-question screen, the scoring anchors, the zone thresholds, and the boardroom application are mine.
The freight matrix was built and analyzed by Ram Dhan Yadav Katamaraja, Founder and CEO of Colaberry Inc. and the team building AIXFreight, and is reproduced with permission.
References
Acemoglu, Daron. “The Simple Macroeconomics of AI.” NBER Working Paper 32487, May 2024. Published in Economic Policy, vol. 40 no. 121, 2025.
Amodei, Dario. Public statements on AI capability and white-collar displacement, 2024 to 2026.
Bessen, James. Research on technology and bank teller employment. Boston University.
Bureau of Labor Statistics. Occupational Outlook Handbook. Bank teller employment projections through 2034.
Gans, Joshua, and Avi Goldfarb. “O-Ring Automation.” NBER Working Paper 34639, January 2026. Rotman School of Management, University of Toronto.
Imas, Alex, and Soumitra Shukla. “How Will AI-driven Automation Actually Affect Jobs?” Substack, March 2026.
Karpathy, Andrej. “AI Job Exposure Map.” karpathy.ai/jobs, March 15, 2026. Interactive analysis of 342 US occupations from the Bureau of Labor Statistics Occupational Outlook Handbook. The accompanying GitHub repository was removed by the author within hours of publication after the visualization was widely misinterpreted as a displacement forecast.
Kremer, Michael. “The O-Ring Theory of Economic Development.” Quarterly Journal of Economics, August 1993.
Provenance
Originally published as a LinkedIn article on April 20, 2026. This page is the canonical version. Cite this URL.
Do the analysis now. Not after the pressure arrives.
About the author
Harry Glorikian is Managing General Partner at Scientia Ventures and a Visiting Researcher at the MIT Media Lab. His forthcoming book The Invisible Interface: How AI Turns Intentions Into Actions, And Who Wins (Simon & Schuster, June 2026) examines how AI is reshaping the layer between people and the work they get done. Available for pre-order. He hosts The Harry Glorikian Show and is the author of MoneyBall Medicine and The Future You.