Seedlabs

DeepRule: Infinite-Depth Eligibility Engine for Policies and Contracts

A rules engine that answers eligibility questions from insurance policies, tax codes, and contracts by recursively decomposing them into logical steps inside the call stack, so even hundreds of nested conditions resolve deterministically with a full audit trail.

Computer ScienceAdvanced Neural Network Applications
Cross-Corpus Speech Emotion RecognitionInsurance claims and tax/benefits eligibility determination

Concept

A rules engine for organizations that answer eligibility questions against long, deeply nested documents — insurance policies, tax codes, benefits rules, contracts. When a query arrives, the engine applies recursive query fragmentation: the question is decomposed via logical operators ("if", "and", "therefore") into primitives resolved recursively within the program call stack, so nesting depth is limited by compute, not by a transformer's context window. The output is a deterministic yes/no determination plus a full tree showing exactly which conditions were evaluated, making every decision explainable to regulators and customers. An LLM is used only at the edges — translating document language and phrasing the final explanation — never for the determination itself.

Why now

The new paper [0] introduces recursive query fragmentation specifically to "process infinite logical depth within the call stack itself, bypassing the degradation typically seen in standard transformer context windows" — precisely the failure mode when 200-page policies with layered conditions are stuffed into an LLM context. The corpus also shows the industry moving toward deterministic control over LLM output, e.g., SOSA-MVU's minimal verification units for deterministic output control [5]; this product applies the same determinism principle to the rule-determination step itself rather than to post-hoc checking.

Written by GLM · 2 Oct 2026

AI assessment

Backed by 8 papers77

A high-utility application of formal logic translation to high-stakes compliance, though it overlaps significantly with the 'Logic Coprocessor' concept.

Evidence strength
4/5
The idea is well-supported by a convergence of papers on translating natural language to temporal/formal logics (nl2spec, TR2MTL) and the need for deterministic governance boundaries (ETA).
Market pull
5/5
Insurance and tax compliance are multi-billion dollar industries where deterministic, auditable 'yes/no' decisions are a regulatory requirement, not a luxury.
Novelty & moat
2/5
The core mechanism—offloading logic from an LLM to a deterministic mathematical engine for auditability—is nearly identical to the 'Logic Coprocessor' idea.
Feasibility
3/5
While the logic engine is feasible, the 'edge' (translating complex 200-page legal documents into perfect formal logic) remains a significant technical hurdle.
Wedge clarity
4/5
Focusing specifically on insurance eligibility and tax codes provides a sharp, high-value entry point compared to a general-purpose logic engine.
Simplicity / focus
5/5
The product is a single, focused tool: a deterministic eligibility engine with a clear input (policy/query) and output (verdict/audit trail).

Scored by Gemma 4 · 10 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Business analysis

The PESTEL analysis reveals a strong strategic alignment with current regulatory demands for AI explainability and the technical shift toward deterministic governance. While the technological and legal drivers are highly positive, the primary risks lie in the high economic cost of initial formalization and the social friction of replacing human-led policy interpretation.

Political2

Economic2

Social2

Technological2

Environmental2

Legal2

The product's viability is heavily dependent on legal and regulatory requirements for deterministic audit trails in insurance and tax law. · Generated 2026-10-10 by Gemma 4AI-generatedFull PESTEL Analysis →

Who benefits

  • Lemonadecompany

    Its model depends on instant, defensible claims decisions; resolving deeply nested policy conditions deterministically with a per-decision audit trail would speed payouts while reducing disputed denials.

  • Intuitcompany

    TurboTax must resolve extremely nested tax conditionals; a call-stack rules engine that returns the determination plus the exact chain of rules applied directly powers its 'why do I owe this?' explanations.

  • Icertiscompany

    Its contract-intelligence platform analyzes conditional obligations and entitlements across large contract portfolios; infinite-depth logical parsing of nested clauses is a core unmet capability.

  • Centers for Medicare & Medicaid Services (CMS)organization

    Benefits eligibility rules are notoriously deep and drive a high volume of appeals; auditable, deterministic determinations with a visible rule trace would reduce errors and administrative burden.

Research it builds on

  1. Abstract Functional Language Logic: A Competitive Mixture of Experts Architecture for Paradox-Free Reasoning and Adaptive Intelligence
    Torres H., Juan P. · 2025 · 588 citations
    All ideas from this paper →
  2. nl2spec: Interactively Translating Unstructured Natural Language to Temporal Logics with Large Language Models
    Matthias Cosler, Christopher Hahn, Daniel Mendoza et al. · 2023 · 122 citations
    All ideas from this paper →
  3. Formalization of natural language requirements into temporal logics: a survey
    Igor Buzhinsky · 2019 · 36 citations
    All ideas from this paper →
  4. TR2MTL: LLM based framework for Metric Temporal Logic Formalization of Traffic Rules
    Kumar Manas, Stefan Zwicklbauer, Adrian Paschke · 2024 · 31 citations
    All ideas from this paper →
  5. Formal Consistency Checking over Specifications in Natural Languages
    Rongjie Yan, Chih‐Hong Cheng, Yesheng Chai · 2015 · 29 citations
    All ideas from this paper →
  6. Execution-Time Authorization for AI Agents: A Formal Framework for Deterministic Governance Boundaries
    Edward Meyman · 2026 · 18 citations
    All ideas from this paper →
  7. SOSA-MVU: Formal Validation Layer for Deterministic Output Control in LLM Systems
    Michele Bottino · 2026 · 8 citations
    All ideas from this paper →
  8. Execution-Time Authorization for AI Agents: A Formal Framework for Deterministic Governance Boundaries
    Edward Meyman · 2026 · 3 citations
    All ideas from this paper →
Leave feedback
feasibility