COOKED
The core loop — pull the account file, dial the debtor, deliver scripted disclosures, negotiate a payment plan inside pre-set authority limits, log the disposition, mail the validation letter — is exactly the text-and-voice work current AI performs at usable quality, and self-service payment portals already remove many calls entirely. FDCPA and state agency licensing constrain how collection happens but almost never require a licensed individual on the call, so there is no personal liability shield. What survives is thin: high-balance or disputed accounts, skip tracing with fragmented evidence, hardship judgment calls, and compliance oversight of the automated contact channel itself.
Most of this decline happened after 2021 — it is not the pandemic dip.
Median pay $37,000 → $47,030 +1.7% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
-10.5% 166,900 → 149,400 on the projections basis
Exposed, and shrinking
Both signals point the same way: the tasks are largely automatable and the BLS projects -10.5% by 2034. This is the case where the score and the forecast agree, and it is the one worth taking seriously.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~13,700 openings a year on average, including replacing people who leave.
ChaserDunnerCollectorRepossessorBill CollectorData CollectorDebt CollectorRent CollectorCar RepossessorField CollectorClaims CollectorCollection AgentCollection ClerkCredit SpecialistInstallment AgentMedical CollectorPayment CollectorAccounts CollectorCredit CoordinatorRepossession AgentCollections AnalystCollections OfficerInsurance CollectorTelephone Collector
Holding it up: judgment & accountability . Weakest point: embodiment .
Core tasks are already automatable Dialing from a queue, reading the mini-Miranda, taking a card payment, setting up a 6-month arrangement inside the authority matrix, and coding the account NOAN or PTP are all fully scriptable steps that voice AI and IVR/portal flows already run end to end — the 5 rather than 0 reflects genuinely hard skip tracing on stale addresses and untangling a debtor who claims identity theft or a bankruptcy stay.
Fully desk- and screen-based The entire shift is a headset, a dialer, and a CRM screen in a call-center seat; nothing in the job requires leaving the chair — repossession and field visits belong to other roles entirely.
No licence, no signature requirement State collection-agency licenses and bonds attach to the agency, not to you; FDCPA and TCPA violations land on the employer or the creditor, and you can be trained onto the floor in weeks with no exam, so the 3 is only for the collector-registration and background-check requirements a few states impose on individual employees.
Executes defined procedures on defined inputs Settlement percentages, hardship deferrals, and interest waivers come off a pre-approved matrix with anything unusual escalated to a supervisor, and the consequential calls — charge-off, litigation referral, credit reporting — are made by the creditor, leaving you discretion over tone, timing, and when to flag a dispute.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (5/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (3/20) is whether the law requires a licensed human to sign. Trust premium (5/20) is whether buyers specifically pay for a person. Judgment and accountability (6/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 14 of this occupation's 20 points (70%).
Embodiment (1/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 37/100 — EXPOSED.
Genuine two-tier structure: if bots absorb first-contact, payment-plan-within-authority, and portal deflection, the residual role is disputed-debt investigation, skip tracing on fragmented/conflicting identity evidence, FDCPA dispute validation under §1692g, and bankruptcy/deceased/cease-communication handling. Watch for BLS/agency staffing mixes where the collector headcount falls but median wage rises — the signature of tier collapse upward.
TCPA enforcement treating AI voice agents as 'artificial or prerecorded voice' (the FCC's Feb 2024 AI-voice declaratory ruling already says AI-cloned voices are covered), so calls without prior express consent carry per-call statutory damages. That pushes outbound dialing back onto live humans for any account lacking documented consent.
A state licensing regime that names a personally accountable individual for collection communications — e.g., NY DFS Part 1 debt collector rules or Massachusetts Div. of Banks licensing amended to require a designated licensed compliance manager to attest to, and be personally liable for, every automated/AI-generated dunning contact and validation notice. Today licensure attaches to the agency, not a person on the call.
Hardship and ability-to-pay determinations becoming a regulated, documented decision: e.g., CFPB medical-debt rules and state hospital financial-assistance statutes (CA AB 1020) that require a documented affordability determination before referral or continued collection, with the determiner named in the file.
A CFPB supervisory expectation or consent order requiring a human reviewer to approve any AI-recommended escalation to litigation, credit furnishing, or wage garnishment — parallel to the human-review conditions appearing in recent UDAAP consent orders.
The limit. Trust premium has no realistic route — no debtor pays extra to be dunned by a human, and creditors buy recovery rate and compliance exposure, not humanness. Embodiment is structurally near zero. Even with every lever above, the realistic ceiling is a much smaller occupation of compliance-officer-like specialists rather than a restored 158k-person call floor: the levers raise per-worker defensibility, not headcount.
| Dallas-Fort Worth-Arlington, TX | 8,950 | $47,790 +2% |
| Los Angeles-Long Beach-Anaheim, CA | 6,650 | $57,440 +22% |
| New York-Newark-Jersey City, NY-NJ | 6,210 | $54,370 +16% |
| Phoenix-Mesa-Chandler, AZ | 5,140 | $47,740 +2% |
| Houston-Pasadena-The Woodlands, TX | 4,660 | $44,670 -5% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 3,990 | $46,870 +0% |
| Chicago-Naperville-Elgin, IL-IN | 3,950 | $49,390 +5% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 3,840 | $49,110 +4% |
| San Jose-Sunnyvale-Santa Clara, CA | 660 | $77,360 +64% |
| Napa, CA | 80 | $77,160 +64% |
| San Francisco-Oakland-Fremont, CA | 1,450 | $72,430 +54% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 20. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.