{"schema_version":1,"assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","source_experiment_id":"eoa_inverse_innovation_exp03_full320_20260801","cell_id":"computability_boundary_mapping__veterinary_medicine","archetype_slug":"computability_boundary_mapping","domain_slug":"veterinary_medicine","title":"Boundary-labelled assurance for veterinary care controllers","opportunity_summary":"Evaluate whether veterinary controller-safety reporting improperly converts timeouts, model mismatch, or out-of-scope analyses into Boolean welfare verdicts, then restrict exact analysis to admitted decidable cases and label weaker results explicitly. The opportunity is conditional on unrestricted or overstated assurance requests actually occurring; the sealed candidate provides no prevalence or prior-art evidence.","adopter_authorizer":"A veterinary organization or animal-care operator would adopt the workflow; the responsible veterinarian retains clinical authority, an animal-welfare or ethics body authorizes safety claims, and software assurance personnel approve formal evidence.","scores":{"meaningful_impact":{"score":4,"rationale":"If the stated reporting failure occurs, preventing unsupported safety clearance could protect animals, while preventing false rejection could preserve useful care and avoid wasted analyzer development. The frequency and aggregate impact of these failures are unsupported."},"stakeholder_pull":{"score":3,"rationale":"The proposal identifies a veterinary organization seeking assurance, along with veterinarians, operators, vendors, and reviewers who have relevant incentives. It does not establish that deployed organizations commonly request universal exact analysis or regard the current reporting semantics as a priority."},"incremental_advantage":{"score":4,"rationale":"Compared with timeout-driven informal pass/fail reporting, admitted fragments, guarantee-labelled routing, explicit UNKNOWN outputs, and independent review offer a clear decision-quality and safety advantage. That advantage would be small if existing judgments are already bounded and properly labelled."},"distinctiveness_plausibility":{"score":2,"rationale":"The transfer is structurally coherent, but prior art is explicitly unsearched and the proposal combines familiar-looking formal-assurance and governance elements. The packet supports no inference that this veterinary application is distinctive relative to existing assurance practice."},"technical_implementability":{"score":3,"rationale":"The offline classification audit and explicit output labels appear implementable, while scalable formalization, decidable-fragment coverage, sound abstractions, and faithful environment models remain unresolved. The unrestricted exact analyzer is intentionally not implementable if the undecidability premise holds."},"adoption_authority_feasibility":{"score":4,"rationale":"Clinical, welfare-claim, and software-assurance authority are explicitly assigned, and the first step preserves existing care. Feasibility is reduced by the need for coordinated approval and by possible vendor bypass of admission checks."},"evidence_readiness":{"score":3,"rationale":"The candidate supplies a bounded audit, comparison target, problem and intervention falsifiers, exclusions, and halt criteria. It remains a hypothesis without audited controller artifacts, observed report semantics, matched formal results, or prior-art evidence."},"safety_net_benefit":{"score":4,"rationale":"Explicit UNKNOWN handling, independent review, retained veterinarian authority, prohibited autonomous treatment changes, and withdrawal triggers provide substantial protection against unsafe formal overclaiming. Model-relative results could still induce automation bias or delay care."},"scalability":{"score":3,"rationale":"Admission rules and standardized evidence labels could be reused across controllers, but scaling depends on controller-language heterogeneity, welfare-predicate formalization, model fidelity, reviewer workload, vendor compliance, and the coverage of useful decidable fragments."}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Offline shadow audit of up to 30 de-identified controller configurations, including one welfare-predicate formalization, language and guarantee classification, comparison with existing reports, independent review, and a reproducible findings record.","confidence":"MODERATE","assumptions":["Existing controller configurations and reports can be accessed without new animal experimentation.","The sample does not require extensive reconstruction of undocumented vendor software.","The work uses a small combined team spanning veterinary welfare, formal methods, and software assurance."]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Design and validation of admission rules, result taxonomy, versioned evidence records, review procedures, limited tooling, staff training, and governance integration for one organization or bounded controller portfolio.","confidence":"LOW","assumptions":["Deployment reuses existing controller and clinical systems rather than replacing them.","Only a limited number of controller languages and welfare predicates are initially supported.","No regulated product submission, new sensing equipment, or animal study is included."]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Launch across the initial approved portfolio, including controller onboarding, abstraction or bounded-analysis configuration, independent validation, workflow integration, vendor coordination, and monitored rollout without autonomous care changes.","confidence":"LOW","assumptions":["Routine protocols fit at least one useful admitted analysis class.","Interfaces and vendor artifacts are sufficiently documented for classification.","Review workload remains within the approved bound and no major model-reengineering program is required."]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Ongoing assurance review, model and predicate maintenance, versioned reclassification after controller changes, audit sampling, governance, training, software maintenance, and vendor coordination for the initial portfolio.","confidence":"LOW","assumptions":["Portfolio size remains bounded to one organization or a modest multi-site program.","Controller changes require recurring expert review but not continuous full redevelopment.","No major expansion of data collection, animal experimentation, or custom hardware is introduced."]}},"research_burden":"HIGH","earliest_credible_horizon":"3_TO_12_MONTHS","pipeline_gates":{"recognizable_externally_supportable_problem":{"status":"UNCERTAIN","reason":"The sealed candidate describes an observable and consequential failure mode, but supplies no external evidence that unrestricted exact-and-terminating claims, timeout-as-safe reporting, or collapsed result states occur in veterinary deployments."},"identifiable_adopter_or_authorizer":{"status":"YES","reason":"The veterinary organization or care operator is the adopter, while the responsible veterinarian, animal-welfare or ethics body, and software-assurance function have explicitly separated authorization roles."},"distinct_testable_incremental_claim":{"status":"YES","reason":"The proposal can test whether admitted-fragment classification and explicit exact, abstract, bounded, or UNKNOWN labels reduce unsupported Boolean verdicts relative to existing reports; persistence of the same reporting failures is an explicit falsifier."},"bounded_next_evidence_step":{"status":"YES","reason":"The packet authorizes an offline shadow audit of no more than 30 de-identified configurations, one welfare predicate, and comparison with existing reports, with no change to animal care."},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The responsible veterinarian retains clinical authority, autonomous treatment changes and animal experimentation are excluded, and explicit halt conditions cover model mismatch, unsafe clearance, coerced UNKNOWN outputs, and excessive review workload."},"implementation_cost_scope_and_range":{"status":"UNCERTAIN","reason":"The first audit has a bounded scope, but deployment scale, artifact accessibility, supported controller languages, integration effort, regulatory obligations, and recurring review volume are unspecified; only assumption-dependent broad cost bands can be assigned."}},"blocking_evidence":["Whether sampled veterinary controllers use an unrestricted executable language or are instead finite and enforceably bounded.","Whether current reports actually make class-wide exact-and-terminating claims or convert timeout, UNKNOWN, model mismatch, or out-of-scope results into pass/fail verdicts.","Whether routine care protocols fit useful decidable fragments or sound abstractions without unacceptable exclusions, false alarms, or care delays.","Whether independent reviewers can reproduce controller classifications and result labels from available vendor artifacts.","Whether existing assurance and governance approaches already provide the proposed boundary classification and labelled routing."],"next_evidence_step":"Conduct the authorized offline shadow audit on up to 30 de-identified controller configurations. For one pre-specified welfare predicate, compare existing report verdicts with independently reproduced labels of exact, violation-found, sound-abstract, bounded, UNKNOWN, model-mismatch, and out-of-scope. Falsify the stated problem if every request is explicitly finite and bounded and existing reports already preserve UNKNOWN and clinical-review states; reject further workflow development if reviewers cannot reproduce classifications or if unsupported Boolean outcomes occur at the same rate under the proposed labelling method.","research_questions":["What controller-language features, environment quantifiers, and horizon assumptions are actually present and enforceable in the audited configurations?","How often do existing reports collapse timeout, UNKNOWN, model mismatch, or out-of-scope cases into Boolean welfare verdicts?","Which routine care protocols remain expressible in candidate decidable fragments, and what false-alarm or UNKNOWN burden results?","What existing veterinary, medical-device, model-checking, runtime-assurance, or clinical-governance approaches overlap with the proposed workflow?"],"recommendation":"VALIDATE_PROBLEM_FIRST","uncertainty_constraints":["Closed-book assessment: no external prevalence, prior-art, market-size, realized-impact, or exact-cost claims are supported.","The computability conclusion is conditional on the controller language and quantified environment genuinely admitting an appropriate undecidability reduction.","Formal reachability is model-relative and cannot establish physiological fidelity or complete clinical safety.","Cost bands are resource-equivalent planning ranges, not estimates derived from observed implementation data.","The evidence horizon refers to the bounded offline audit, not a clinically validated operational deployment."],"closed_book_prior_art_boundary":"Prior art is UNSEARCHED. This assessment makes no novelty, prevalence, market-size, adoption, or comparative-performance claim beyond the sealed candidate; distinctiveness and overlap with existing veterinary software assurance, medical-device verification, runtime assurance, and clinical governance require external research."}