Discovery Lens
C Combination Innovation
Two separate worlds finally connect — and the intersection is a product
USCIS published a comprehensive machine-readable dataset of adjudication outcomes in 2023, and the UK Home Office followed with similar transparency. For the first time there is enough structured case outcome data to train meaningful predictive models — precisely as the immigration backlog hit historic highs above 9 million pending cases.
One-Liner
an AI-powered immigration case management tool that predicts outcomes, optimizes applications, and guides applicants through the process.
The Journey
◆Origin
Immigration processing delays have reached historic levels — average USCIS processing times for employment-based categories now exceed 2-3 years. Applicants and their sponsoring employers desperately need better guidance on strategy, documentation strength, and timing optimization. Immigration attorneys are bottlenecked and expensive ($3,000-$15,000 per case), creating a massive underserved middle tier of applicants who need more than a form-filing service but cannot afford full legal representation.
⚡The Breakthrough
Historical immigration case outcome data — USCIS public datasets, FOIA releases, court records — intersects with LLMs' ability to parse complex regulatory requirements and generate plain-language guidance. The combination enables case-strength analysis and outcome forecasting at scale, previously only available through expensive attorney consultation during billable hours.
☠Almost Killed
Unauthorized Practice of Law regulations nearly killed this idea. In most US states, providing immigration advice for compensation without a law license is a federal crime — the DOJ EOIR has actively prosecuted non-lawyers for exactly this. The survival path: position as a decision-support tool for immigration attorneys and corporate HR departments (B2B, not direct-to-consumer), provide case analytics rather than legal advice, and require attorney sign-off on all guidance output. The tool saves 80% of attorney research time; the attorney does the legal work.
⏰Why Now
USCIS published its first machine-readable outcome dataset under a 2023 transparency initiative. The backlog also hit record highs in 2024, creating overwhelming demand for efficiency tools from both applicants and law firms. Meanwhile, several immigration law firms began using AI document review successfully in 2024, demonstrating regulatory tolerance for the tool-not-advice positioning and opening the door for a dedicated analytics layer.
The Surprising Insight
Outcome prediction models trained on USCIS historical data can predict approval probability with 70-80% accuracy — yet 90% of applicants file without any data-driven guidance, relying on attorneys who rarely share success statistics for their specific case profiles.
Kill Reason
Critical weakness: Regulatory risk
What do you think?
Related ideas you can explore free:
killed: WiFi signal-based spatial mapping does not have the precision required for building compliance enforcement, and the primary customer — municipal governments and inspectors — moves too slowly for startup sales cycles. Any IoT company can replicate the hardware stack from commodity ESP32 chips; there is no proprietary data asset that accrues over time.
killed: Government information AI assistants are being built by civic tech organizations, state governments directly, and commercial players — USA.gov, state agency chatbots, and legal aid platforms. Consumer willingness to pay for government information access is near zero since citizens expect these services to be free, making the business model dependent on grant funding rather than commercial revenue.
killed: In most jurisdictions, retired government officials face strict post-employment ethics rules — equivalents of the US Hatch Act — that severely restrict paid consulting on matters related to their former duties, gutting the core supply side of the marketplace. LinkedIn could replicate the expert directory concept in a single product sprint.