EXPOSED
The desk half of this job — public-records pulls, skip tracing, social media sweeps, credit and criminal database checks, and writing up the investigative report — is exactly what aggregation platforms and LLMs now do faster and cheaper, and that half has been shrinking for a decade. What survives is physical: sitting on a subject for six hours, following a vehicle through traffic, knocking on a neighbor's door, testifying about the chain of evidence. Most states license PIs and courts care who collected the evidence, which puts a real floor under the field-work tier even as the research tier commoditizes.
Roughly flat across the period, with year-to-year wobble.
Median pay $50,510 → $51,220 -18.9% 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
+6%
Percentage only. The projection counts a different population from the 35,580 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Growing, and only partly exposed
The BLS expects +6% more of these jobs by 2034, and at 56/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
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.
~3,900 openings a year on average, including replacing people who leave.
DetectivePrivate EyeSkip TracerInvestigatorSkip LocatorBonding AgentHouse OfficerDistrict AgentHouse DetectiveStore DetectiveSecurity AnalystUndercover AgentCase InvestigatorPrivate DetectiveAlarm InvestigatorField InvestigatorFraud InvestigatorIntelligence AgentSurveillance AgentSecurity ConsultantUndercover OperatorPrivate InvestigatorSpecial InvestigatorSurveillance Manager
Holding it up: embodiment . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier Half your billable hours — TLO and IRB searches, courthouse record pulls, asset and lien checks, drafting the surveillance summary — are now a subscription and a prompt away, while the other half (mobile surveillance in traffic, pretext canvassing at a door, catching the claimant lifting a cooler on video) can't be run remotely, which is why this lands at 10 and not 5 or 15.
Hands-on in uncontrolled environments Insurance and infidelity work means hours in an unair-conditioned vehicle in a residential zone you don't control, moving on foot when the subject does, dealing with being made by a neighbor or a patrol car — uncontrolled and unpredictable, but you're not climbing towers or entering confined spaces, which is what separates 14 from 18.
Licensed human required and personally liable Most states (CA BSIS, TX DPS, FL Chapter 493) require a licensed agency plus individual registration with hour requirements, and an unlicensed investigator's product gets excluded and can trigger criminal exposure — but the licence covers conduct and evidence handling, not a protected scope of practice a non-licensee is barred from performing, so it sits mid-band rather than at physician or PE level.
Meaningful discretion You decide in the moment whether to break off a follow, whether a pretext crosses into the state's prohibited conduct, what goes in the report and what stays out — and that call can taint a case or expose the client — but the assignment scope and legal boundaries are set by the retaining attorney, which caps this at 11.
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 (10/20 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 (11/20) is whether the law requires a licensed human to sign. Trust premium (10/20) is whether buyers specifically pay for a person. Judgment and accountability (11/20) 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 32 of this occupation's 56 points (57%).
Embodiment (14/20) 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 73/100 — SAFE.
Genuine two-tier shift: as records pulls and social sweeps commoditize, the residual day becomes pretext interviews, source development, testimony prep, and deciding what a contradictory record set actually means — work current models cannot do at usable quality. No law needed; already visible in the shift of PI billing from research hours to surveillance and interview hours.
State licensing boards (e.g., California BSIS, Texas DPS PSB) tightening rules so that any investigative report sold to a client — including one assembled from AI-aggregated data — must be signed by the licensed PI of record, who is personally accountable for accuracy; plus FCRA enforcement treating AI-generated background dossiers used for employment/tenant screening as consumer reports requiring a licensed, liable human verifier. CFPB's 2024 guidance on background-screening data accuracy is the visible hook.
Court evidentiary rules (state analogs to FRE 901/902 authentication) requiring a live human witness who personally observed or collected surveillance material before video/GPS evidence is admitted, explicitly excluding automated collection affidavits — being litigated now as deepfake-driven authentication challenges spread.
Insurer SIU contracts and corporate internal-investigation protocols (e.g., post-Sarbanes/whistleblower workflows) naming a licensed investigator as the person who decides whether a claim is referred for fraud prosecution or an employee allegation is substantiated — a consequential call under ambiguity with a named owner rather than a scored output.
Anti-stalking and data-broker statutes (Illinois BIPA-style biometric limits, state ALPR restrictions, EU-influenced state privacy bills) closing off automated remote surveillance, forcing location verification back into physical presence.
Defense-side and family-law clients paying specifically for a human who can be deposed and cross-examined about method — buyers already choose PIs partly for courtroom presentability, and that widens if AI-sourced evidence draws routine reliability challenges.
The limit. The research tier is not coming back; most of the upside is confined to the field and testimony tier, which is a minority of current headcount. Even with every lever, this stays a smaller occupation — resistance rises per surviving worker, not for the 35,580.
| Los Angeles-Long Beach-Anaheim, CA | 1,520 | $61,680 +20% |
| Dallas-Fort Worth-Arlington, TX | 1,070 | $61,460 +20% |
| New York-Newark-Jersey City, NY-NJ | 1,040 | $59,140 +15% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 1,030 | $54,720 +7% |
| Atlanta-Sandy Springs-Roswell, GA | 980 | $45,080 -12% |
| Chicago-Naperville-Elgin, IL-IN | 970 | $68,310 +33% |
| Charlotte-Concord-Gastonia, NC-SC | 860 | $73,430 +43% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 790 | $62,300 +22% |
| Memphis, TN-MS-AR | 310 | $99,680 +95% |
| Salem, OR | 180 | $81,070 +58% |
| Portland-Vancouver-Hillsboro, OR-WA | 270 | $78,330 +53% |
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 56. 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.