← JOURNAL / LEIMAN LAB
FIELD NOTE 009BUSINESS & PRODUCTIONRESEARCH / ECONOMICS OF WORK

The New Economics of Finding Work: Can AI Discover Income Opportunities Before We Do?

How to discover monetary opportunities beyond job listings, separate prize amounts from likely earnings, and evaluate the current Income Intelligence pilot.

Editorial illustration for Income Intelligence — Mapping Opportunity
LEIMAN LAB / 009   ·   EDITORIAL ILLUSTRATION

People usually search for work by typing the name of their profession. Yet income is often found at the intersection of skills, market demand and access to a specific opportunity. A filmmaker may qualify for a grant, a researcher for a paid advisory contract, and a designer for an international project. A single job-title search misses much of that terrain.

I examined labour-market evidence, official job and grant directories, public APIs and research on algorithmic hiring. The result is the Income Intelligence research concept, now accompanied by a public beta. Whether the system actually helps users earn more remains untested.

The working Income Intelligence beta is separate from this research article. A cash prize is conditional on winning, while a grant supports an eligible project under restricted terms. Neither is guaranteed personal income. OPEN INCOME INTELLIGENCE BETA ↗

01 / The market is changing. AI income promises remain unproven

The World Economic Forum's Future of Jobs Report 2025 surveyed more than 1,000 major employers. It projects substantial restructuring of jobs and skills through 2030, with AI and big data, cybersecurity and adaptability among the areas gaining importance. Those are employers' expectations, not evidence that any individual applicant will find a job more easily. WEF, Future of Jobs 2025 ↗

An IMF study published in 2026 associates the appearance of new skills in job postings with employment and wage growth in particular US labour markets. Association is not a causal benefit for every worker. Upwork's 2026 skills report also points to demand for people who combine specialist expertise with AI tools, but its evidence comes from its own marketplace rather than the entire global economy. IMF, Bridging Skill Gaps (2026) ↗ · Upwork, In-Demand Skills 2026 ↗

02 / Employment is only one route to income

A conventional aggregator asks which employer needs a worker. A better question for someone with several skills is: who is prepared to pay for an outcome I can deliver, under what conditions, and how soon? That question extends well beyond a vacancy.

  • Employment: local, international and remote positions with disclosed pay, links to primary sources and advertised location restrictions. Actual hiring conditions and eligibility still require verification.
  • Freelance and contracts: paid assignments, tenders, project roles and short advisory engagements.
  • Grants and research funding: project support subject to eligibility, institutional affiliation and geographic restrictions.
  • Competitions, awards and residencies: cash awards, production support, funded programmes or purely non-cash benefits.
  • Commercial leads: business needs, licensing, partnerships, teaching and product niches, each requiring independent evidence of demand.

A grant is not necessarily personal income. Some funds may only be spent by an eligible institution or on a narrowly defined project. Awards can carry taxes, rights transfers or exclusivity. If a platform shows the headline amount but hides those restrictions, its ranking is economically misleading.

03 / A credible data layer starts with primary sources

There are practical starting points. Greenhouse makes public GET endpoints available for company job boards, while Lever provides its Postings API. These sources can make employer-originated information easier to retrieve. They do not imply a universal right to replicate every jobs database, nor complete coverage of the global market. Greenhouse Job Board API ↗ · Lever Postings API ↗

Grants.gov and the EU Funding & Tenders Portal document APIs for funding opportunities. EURAXESS publishes research jobs, grants and hosting positions; USAJOBS specifies conditions for using its federal jobs API. Each feed needs its own licence review, refresh strategy, eligibility extraction and rules for republication. Grants.gov API ↗ · EU Funding & Tenders APIs ↗ · EURAXESS ↗ · USAJOBS API terms ↗

The first pipeline is deliberately boring: ingest authorised records, normalise fields, deduplicate, recheck status, extract hard requirements, then rank for a user. Every result must retain its original source, direct URL, retrieval time and most recent verification. When an announcement closes, the system must stop presenting it as an open opportunity.

04 / An opportunity score should measure fit, not destiny

Instead of a fabricated '93% chance of success', I propose a transparent Opportunity Fit Score from 0 to 100. It describes how well an opening fits a person's verified circumstances and stated goals. These initial weights are a design hypothesis and must be tested.

  • 25 points: match between demonstrated skills and explicit requirements.
  • 20 points: eligibility, including jurisdiction, work authorisation, credentials and language.
  • 15 points: economics, including expected net pay or usable funding relative to time and direct costs.
  • 15 points: source credibility and evidence that the opportunity remains open.
  • 10 points: time available and realistic preparation burden before the deadline.
  • 10 points: alignment with preferred working arrangements and career direction.
  • 5 points: documented information about competition or comparative advantage, when actually available.

Hard disqualifiers should be applied separately. If a candidate cannot legally apply, a high aggregate score must not conceal that fact. Every card should distinguish verified, unknown and needs-human-review fields. Missing information cannot silently become zero or a confident invention.

Estimating the probability of getting hired or winning a grant would require dependable application histories, a calibrated model, selection-bias controls and prospective validation. None of that is available in version one. So we publish relevance rankings, not pseudo-scientific percentages.

05 / Three illustrative cases

Consider a filmmaker experienced in generative video. Searching 'AI filmmaker' retrieves some roles. A broader system would also check AI film competitions, funded residencies, production tenders, branded-video commissions, short art-direction contracts and creator support programmes. Geography, rights and mandatory-model requirements will disqualify some results. Leiman Lab guide: AI film competitions 2026–2027 ↗

A psychologist who also conducts research might explore behavioural-science grants, UX research contracts, teaching engagements and research roles, alongside conventional job listings. The system must distinguish academic or advisory work from regulated clinical practice, where professional licensing is decisive.

A product designer could compare full-time employment, fixed-scope agency work, design-system contracts and team training. Beyond advertised pay, the meaningful factors include workload, milestone payments, jurisdiction, copyright and substantiated payment-risk indicators. These three examples are scenarios, not a list of verified live vacancies or invented fees.

06 / The CV is an input, not a commodity

The first intake should accept a text CV, declared skills, country of residence, languages, eligible jurisdictions, working preferences, available hours and minimum financial requirements. A later release may support PDF/DOCX uploads, portfolios and evidence of completed work. AI can suggest extracted skills, but the person must be able to correct every inference.

  • Separate, understandable permission for CV analysis and for optional opportunity notifications.
  • Clear retention periods and the ability to export, amend and delete personal data.
  • No public indexing of CVs and no resale of personal profiles without a separate, specific choice.
  • No sending applications, CVs or personal data to third parties without human confirmation.
  • Avoid collecting protected or sensitive attributes for scoring unless a specific, lawful necessity is demonstrated.

Research on algorithmic hiring documents both systematic bias and opaque decision-making. Under the EU AI Act, many recruitment and selection systems are treated as high-risk; GDPR imposes lawful-processing requirements and data-subject rights. The safer initial product is a candidate-side research assistant that explains its suggestions, not an autonomous employer-side gatekeeper. Survey: Fairness and Bias in Algorithmic Hiring ↗ · European Commission, AI Act employment ↗ · European Commission, GDPR rights ↗

07 / Proposed free service and subscription economics

The free service should produce immediate value: a skills profile, a basic CV assessment, a small set of verified leads, explained rankings and links to primary sources. Charging people for a mysterious opportunity to get rich would merely recreate the industry's least attractive habit.

  • PROPOSED, NOT LIVE: an individual profile, an initial CV review, 10 relevant records per week and a saved-opportunity history.
  • Pro, test price $12–19/month: personal monitoring, broader source coverage, deadline alerts, net-income comparison and tailored application materials.
  • Professional, test price $39–59/month: multiple professional profiles, grant and tender portfolio, collaboration and application analytics.
  • Paid placement must never secretly change the editorial score. Sponsored opportunities and referral links require clear disclosure.

These are proposed prices for possible future experiments, not active subscription plans. They should be evaluated against actionable leads, deadline accuracy, acquisition and inference costs, and real economic outcomes, rather than the number of cards displayed.

08 / Test the economic result, not the interface

The minimum credible experiment compares ordinary manual search with assisted search over a 30-day period. Participants should cover several occupations and experience levels. Record their existing networks, geography and hours spent. The aim is to test opportunities that can actually be acted on, not the visual appeal of a dashboard.

  • Primary metric: verified, personally eligible opportunities discovered per hour of search.
  • Secondary metrics: response rates, interviews, funded applications, preparation costs, and source freshness.
  • Longer-term metric: money actually received, net of unavoidable costs, measured separately from headline award sizes.
  • Safety metrics: mistaken eligibility, stale listings, invented deadlines, unwanted contact and privacy complaints.

09 / From research proposal to public beta

As of 19 November 2026, a public beta provides vacancy discovery, selected cash-prize competitions, limited monetary grant and tender searches, preliminary requirement matching and illustrative execution plans. Key parts of the vision remain unbuilt: automatic CV analysis, complete legal eligibility checks, independent payout verification, application submission, autonomous delivery and paid subscriptions. The next test is whether the system helps people obtain real economic results, not simply whether it displays more opportunities.

Research snapshot: 18 November 2026. Product status updated on 19 November 2026. Job, grant, competition and API terms change; verify the original source immediately before applying.

SOURCES12 LINKS
  1. World Economic Forum — The Future of Jobs Report 2025 ↗RESEARCH
  2. IMF — Bridging Skill Gaps for the Future (2026) ↗RESEARCH
  3. Upwork — In-Demand Skills 2026 ↗MARKETPLACE REPORT
  4. Greenhouse — Job Board API ↗OFFICIAL DOCUMENTATION
  5. Lever — Postings API ↗OFFICIAL DOCUMENTATION
  6. Grants.gov — API Resources ↗OFFICIAL DOCUMENTATION
  7. European Commission — Funding & Tenders APIs ↗OFFICIAL DOCUMENTATION
  8. EURAXESS — Jobs & Funding ↗OFFICIAL DIRECTORY
  9. USAJOBS — Developer Portal ↗OFFICIAL DOCUMENTATION
  10. ACM — Fairness and Bias in Algorithmic Hiring ↗PEER-REVIEWED SURVEY
  11. EU AI Act Service Desk — Employment ↗LEGAL REFERENCE
  12. European Commission — GDPR rights ↗LEGAL REFERENCE

Sources were reviewed in November 2026. Proposed scoring weights, possible subscription prices and functionality beyond the deployed beta are Leiman Lab hypotheses, not verified commercial results.

COMMENTS / 0

SELECT TEXT TO DISCUSS A PASSAGE