{ "schema_version": 1, "builder_plan": { "scenario": "An agent looks good in demos, but you cannot tell which product tasks it fails.", "change": "Use the study's evaluation idea to inspect a small set of tasks with human-checked answers.", "baseline": "Your current evaluator and a manually reviewed reference set.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Agreement with human review, missed failures, false alarms, and evaluation cost per task.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2610.06587", "title": "Mind the Accent Gap: British Accent Robustness in Speech-Driven Financial Voice Assistants", "summary": "CavaBench replaces ad hoc or American-English-only ASR selection with an accent-stratified, end-to-end evaluation of ASR models plus downstream LLM tool-calling on financial voice queries. The cost is building or replicating an internally collected accent-labelled speech benchmark and running both WER and task-level evaluations per candidate model, since WER alone is an insufficient proxy under latency and memory constraints. Adoption can fail if the team's accent distribution or acoustic conditions differ from the benchmark's, because the best ASR model by WER is not guaranteed to be the best by tool-calling accuracy.", "rationale": "The reader can apply the evaluation methodology directly when choosing among API or 24GB-deployable ASR models, but the specific benchmark is internal, so results must be validated on their own workload and accent mix.", "technique": "Accent-robustness benchmark for ASR-LLM voice pipelines", "family": "avaliacao_de_agente", "recommendation": "testar", "published": "2026-10-05", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.06587", "independent_implementations": 0, "total_implementations": 1, "stars": 1, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "CavaBench replaces ad hoc or American-English-only ASR selection with an accent-stratified, end-to-end evaluation of ASR models plus downstream LLM tool-calling on financial voice queries. The cost is building or replicating an internally collected accent-labelled speech benchmark and running both WER and task-level evaluations per candidate model, since WER alone is an insufficient proxy under latency and memory constraints. Adoption can fail if the team's accent distribution or acoustic conditions differ from the benchmark's, because the best ASR model by WER is not guaranteed to be the best by tool-calling accuracy.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.06587", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "WER strongly predicts downstream tool-calling accuracy (r = -0.93), but WER can fail to reflect task-level performance; accent-related failures vary substantially across ASR models and acoustic conditions.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.06587", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }