AI-Drafted Subpoenas for Medical Records in Workers Comp

Every workers' comp practitioner knows the drill. A critical medical record sits locked behind a non-compliant provider, your subpoena deadline is tomorrow, and your paralegal is buried in three other urgent files. The fastest firm wins — and right now, most firms are losing time on a document that AI can draft in seconds.

Subpoenas for medical records are the unglamorous backbone of workers' compensation litigation. Whether you're defending a claim under Labor Code § 4600 or building an applicant's case around QME findings, the ability to rapidly compel accurate and complete medical records can determine whether you settle strong or go to trial blind. Yet the drafting process remains stubbornly manual at most firms. It is prone to jurisdictional errors, missing provider identifiers, and formatting inconsistencies that invite objections and delay production. [SOURCE_1]

This guide breaks down exactly how AI is transforming the medical records subpoena workflow in workers' comp. We cover the statutory framework, common drafting pitfalls, and how to deploy purpose-built legal AI tools that compress hours of document prep into minutes — without sacrificing the precision your cases demand.

The Role of Medical Records Subpoenas in Workers' Comp Litigation

Medical records are the evidentiary foundation of virtually every workers' comp dispute. From injury causation to apportionment under Brodie v. WCAB, the records you compile — or fail to compile — determine the trajectory of a case. No QME evaluation, no AME panel report, and no trial brief stands on solid ground without a complete underlying medical history.

The stakes cut both ways. Applicant attorneys need complete treatment histories to establish industrial causation under the AOE/COE standard. Defense attorneys need those same records to identify apportionment — the allocation of permanent disability to non-industrial causes under Labor Code § 4663 — and to surface pre-existing conditions that limit the employer's liability [SOURCE_2].

Subpoena vs. Authorization: Knowing Which Tool to Deploy

A voluntary HIPAA authorization suffices in many situations. If the applicant cooperates and the provider is compliant, an authorization is faster and cheaper. But a formal subpoena duces tecum becomes necessary when a provider refuses to produce, the applicant revokes consent, or you need a hard production deadline backed by court authority.

Some facilities require a subpoena even when an authorization would legally suffice. They do this to protect themselves from liability. Knowing this dynamic saves time — issue the subpoena first for known problem providers rather than cycling through an authorization that will be rejected.

The strategic advantage of a subpoena is the enforceable deadline. It creates a record for sanctions motions under Labor Code § 5813 if the provider ignores it. That leverage is something a HIPAA release can never provide.

Common Records Targets in Workers' Comp Cases

The provider universe in a typical workers' comp file is wide. You will typically subpoena primary treating physician records under the Labor Code § 4600 treatment framework, QME and AME panel physician records, emergency room and hospital records tied to the date of injury, pharmacy records relevant to apportionment arguments, and prior employer medical surveillance records or pre-employment physicals.

Mental health providers and chiropractors add complexity. Both categories involve additional privacy overlays. They often require specific record-type language in the subpoena to compel complete production. Vague descriptions give these providers grounds to under-produce.

California Statutory and Regulatory Framework for Medical Subpoenas

California Code of Civil Procedure §§ 1985–1997 governs subpoenas duces tecum in workers' comp proceedings [SOURCE_3]. Labor Code § 5710 grants the WCAB independent subpoena authority in workers' comp adjudication. It operates separately from the civil court framework and has its own procedural requirements.

Discovery in WCAB proceedings is also governed by 8 CCR § 10530 et seq. These rules differ from civil court discovery in important ways. The timelines are different. The service requirements are different. Using a civil court subpoena template in a WCAB proceeding is one of the most common — and costly — formatting errors practitioners make.

HIPAA intersects with state subpoena law in ways that create additional procedural requirements. These requirements apply before a subpoena compels production of protected health information. California's Confidentiality of Medical Information Act adds another layer beyond HIPAA. Compliance becomes a two-statute exercise every time you draft a medical records subpoena.

Procedural Requirements Every Subpoena Must Satisfy

Every compliant subpoena must properly identify the custodian of records and the specific facility or provider entity. DBA issues — where a provider operates under a trade name different from the legal entity name — are a frequent source of misdirected subpoenas.

Consumer notice under CCP § 1985.3 requires you to notify the injured worker at least ten days before the subpoena compels production of personal records. Miss this window and opposing counsel has a ready-made objection [SOURCE_3]. Proof of service must comply with WCAB rules, including acceptable service methods and documentation requirements.

The production deadline and location must be specified clearly. WCAB proceedings use deposition officer protocols that differ from civil court practices. Fee and cost requirements for medical record production should also be addressed in the subpoena to avoid provider objections based on unpaid copy fees.

HIPAA Compliance Overlay in Workers' Comp Subpoenas

The HIPAA Privacy Rule requires a qualified protective order — or assurances that one is in place — before a subpoena compels disclosure of protected health information in litigation [SOURCE_4]. This is not optional. Providers who receive a subpoena without proper HIPAA compliance language have a legal basis to refuse production.

Drafting a notice to consumer that satisfies both HIPAA and the CMIA simultaneously is a precise exercise. Generic subpoena templates routinely fail this test. They either omit the federal qualified protective order language or fail to meet California's stricter state-law notice requirements. Either gap gives a provider grounds to object — and gives opposing counsel a delay tactic.

Where Traditional Subpoena Drafting Breaks Down

The manual drafting bottleneck is real. Attorneys and paralegals spend 45 to 90 minutes per subpoena on tasks that add zero strategic value. These tasks include hunting for provider addresses, verifying entity names, inserting statutory citations, and formatting consumer notice blocks. Multiply that by dozens of subpoenas per month across a high-volume caseload and you are losing days of productive attorney time [SOURCE_5].

Jurisdictional formatting errors are the most common failure mode. A practitioner pulls a civil court subpoena template and uses it in a WCAB proceeding. The format triggers an objection. The clock resets. Meanwhile, the QME panel appointment that depended on those records gets pushed back weeks.

Incorrect provider entity names are a close second. Hospital systems operate through subsidiaries. Imaging centers operate under brand names different from their legal entities. A subpoena served on the wrong entity — even the wrong floor of the same building — can result in no production at all.

The High Cost of Subpoena Errors in Workers' Comp

Delayed production has a cascading effect. A stalled subpoena pushes back QME scheduling. A delayed QME evaluation pushes back the WCAB hearing date. A delayed hearing date extends the life of the claim — and the costs that come with it.

Sanctions exposure under CCP § 1987.2 applies to procedurally defective subpoenas. Labor Code § 5813 authorizes sanctions for frivolous or abusive litigation conduct. Systematically defective subpoena practice can fall into that category [SOURCE_6]. The malpractice dimension is equally serious. If critical records are never obtained because a defective subpoena went unchallenged, and those records would have changed the case outcome, you have a professional liability problem.

Opposing counsel in high-value cases will exploit every procedural defect they can find. A defective consumer notice becomes a motion to quash. A motion to quash buys the other side another 30 days. In a case where time and information are leverage, that delay can be decisive.

Why Generic AI Tools Are Dangerous for Legal Drafting

General-purpose large language models are trained on broad internet data. They have no workers' comp specialization. They routinely conflate civil procedure rules with WCAB-specific requirements because those distinctions don't exist clearly in their training data.

The hallucination problem is acute in legal drafting. Generic AI tools fabricate Labor Code section numbers and invent case citations. Under time pressure, an attorney reviewing a draft may not catch every citation before it goes out. A hallucinated citation in a filed document is a professional responsibility issue [SOURCE_7].

Generic AI tools also have no awareness of 8 CCR regulatory updates, En Banc WCAB decisions, or jurisdiction-specific local rules. The regulatory landscape in California workers' comp changes regularly. A tool trained on static data produces stale output. Worse, general AI vendors accept zero liability when their hallucinated content causes a procedural failure. The attorney eats that risk alone.

How AI Transforms the Medical Records Subpoena Workflow

Purpose-built workers' comp AI does what generic tools cannot. It ingests case-specific data — claim number, ADJ number, provider details, injury date, adjuster information — and generates a jurisdiction-compliant subpoena draft in seconds. The output reflects actual WCAB procedural context, not civil court defaults.

Automated population of consumer notice language is one of the highest-value features. The AI inserts CCP § 1985.3-compliant notice blocks without attorney intervention. The HIPAA overlay language is included automatically. The result is a draft that would take a paralegal 45 minutes to produce manually — generated in under 30 seconds.

Integration with case management systems takes this further. When the AI pulls provider data, claim identifiers, and hearing dates directly from the file, the error rate from manual data entry drops to near zero. Version control and audit trails built into AI platforms also create a defensible record of document generation — a practical asset for ethics compliance under California Rule of Professional Conduct 1.1 [SOURCE_8].

From Case File to Draft in Seconds: The AI Drafting Workflow

The workflow is straightforward. Step one: input case identifiers — claim number, ADJ number, provider entity, injury date, and record date range. Step two: the AI selects the appropriate subpoena format based on the proceeding type — records-only versus deposition subpoena, WCAB versus civil court.

Step three: the AI inserts statutory authority, consumer notice, and proof of service blocks automatically. Step four: the AI flags missing information — an incomplete provider address, an ambiguous record category — before the draft is finalized. Step five: the attorney reviews and approves the draft in under five minutes, compared to 30 minutes or more with a manual template.

That compression is not marginal. Across a firm handling 200 subpoenas per month, the time savings fund real capacity — more files, more strategic work, faster resolution.

AI-Powered Cross-Referencing: Subpoenas Informed by Case Intelligence

The most powerful feature is not drafting speed. It is case intelligence. AI can analyze an existing QME or AME report and identify which providers, dates of treatment, and record types should be subpoenaed. This closes the gap between medical findings and discovery strategy.

Consider a QME evaluator who cites 14 prior treaters in their report. How many of those providers have been subpoenaed? Which ones produced records? Which records did the physician rely upon that you haven't seen yet? Manual cross-referencing of a 200-page QME report against a subpoena log takes hours. AI does it in seconds.

This capability also flags apportionment-relevant prior providers mentioned in the medical-legal report that haven't been subpoenaed yet. Under Brodie v. WCAB and Labor Code § 4663, apportionment requires medical evidence of causation — and that evidence lives in the records of providers you may not have reached yet [SOURCE_2].

Key Features to Look for in a Workers' Comp AI Drafting Tool

Not all legal AI tools are equal. For workers' comp practitioners, the baseline requirement is workers' comp specialization. The distinction matters enormously in output quality. Look for a platform built ground-up for the practice area — not a general legal AI retrofitted with a workers' comp module.

Training data integrity is the first technical criterion. The AI must be trained on actual WCAB decisions, En Banc opinions, Labor Code provisions, and 8 CCR regulations. Hallucination resistance follows directly from training quality. Purpose-built vertical AI dramatically reduces fabricated citations compared to general large language models [SOURCE_7].

HIPAA-compliant infrastructure is non-negotiable. Every subpoena workflow touches protected health information. The platform must hold documented HIPAA compliance certifications and handle protected data within a secure environment. California-specific procedural nuances — WCAB local rules, DWC forms, hearing-specific document requirements — must be embedded in the tool's core logic.

Evaluating AI Tools: Questions Every WC Practitioner Should Ask

Start with the basics. Is the AI trained exclusively on workers' comp law, or is it a general tool with a workers' comp skin? Ask the vendor directly. If they can't answer with specificity about their training corpus, that's your answer.

How does the platform handle updates to 8 CCR regulations and new En Banc WCAB decisions? A tool that updates its training data quarterly is meaningfully better than one updated annually — California workers' comp law moves fast. Ask what the vendor's liability posture is when AI-generated content contains errors. Get this in writing.

Does the platform offer document review capabilities beyond drafting? The ability to analyze QME reports and subpoena returns in the same environment — without switching tools — is a significant workflow advantage. SOC 2 Type II certification and a HIPAA Business Associate Agreement are the minimum data security standards to require.

CompFox: Purpose-Built AI for Workers' Comp Practitioners

CompFox is trained exclusively on California workers' comp case law, Labor Code, and WCAB precedent. It is not retrofitted from a general legal AI. That distinction is why CompFox produces subpoena drafts with accurate statutory citations and WCAB-correct procedural blocks — not plausible-sounding but incorrect content.

Hallucination-resistant architecture means the citations and procedural requirements in a CompFox draft reflect actual current law. The platform compresses legal research, document drafting, and QME/AME report analysis into a single workers' comp operating system. CompFox speaks the language of workers' comp natively — apportionment, QME panels, Labor Code § 4664, En Banc decisions — without requiring practitioners to prompt-engineer around general legal training.

If you're ready to stop losing hours to manual drafting and start operating at the speed the practice demands, start researching with CompFox today and experience what a purpose-built workers' comp AI actually delivers.

Practical Implementation: Integrating AI Subpoena Drafting Into Your Firm

Start with a subpoena volume audit. Count how many subpoenas your firm drafts monthly across all open files. For a mid-size defense firm handling 300 active files, that number is often 80 to 150 per month. At 60 minutes average drafting time, that's 80 to 150 attorney and paralegal hours per month spent on a task AI can handle in seconds. The ROI calculation is not complicated.

Build a firm-standard AI subpoena workflow with clear ownership. Who initiates the draft? Who reviews it? What does the quality-check checklist include? The answers should be documented and consistent. AI drafting reduces errors, but attorney oversight remains an ethical requirement under California Rule of Professional Conduct 1.1 — the duty of competence includes supervising AI-generated work product [SOURCE_8].

Train paralegals and legal assistants to use the AI drafting tool effectively. The goal is maximum throughput without sacrificing attorney oversight. Paralegals initiate and populate the AI draft. Attorneys review and approve. That workflow scales in ways that pure manual drafting never can.

Workflow Integration for Defense Firms and TPAs

Claims adjusters and legal ops leads at self-insured employers can use AI drafting tools to prepare subpoena packages for outside counsel review. This compresses the back-and-forth between in-house teams and retained counsel. Outside counsel receives a near-complete draft rather than a blank intake form — saving billable time on both sides.

Defense firms handling high-volume carrier accounts need AI throughput to maintain profitability on flat-fee and per-file billing arrangements. When every subpoena costs 60 to 90 minutes of staff time under a fixed-fee structure, margin erosion is constant. AI drafting restores that margin without requiring additional headcount.

TPA workflows benefit from integration at key case milestones. Configure AI subpoena drafting triggers at QME assignment, trial setting, and lien resolution stages within the claims management system. The subpoena package is prepared automatically when the trigger fires — turning a manual task into an automated workflow step. Learn more about AI Drafting Tool for Workers Comp Trial Briefs.

Workflow Integration for Applicant-Side Practitioners

Applicant attorneys often operate with leaner staff than defense firms. A solo practitioner or two-attorney firm handling 150 active files doesn't have the paralegal bandwidth to manually draft 50 subpoenas a month. AI drafting is a force multiplier that makes a small firm operate like a larger one. Learn more about Best AI Tool for Workers' Comp Trial Briefs.

The applicant-side use case for AI cross-referencing is particularly powerful. Medical records often contain references to prior treaters that aren't in the applicant's own file. AI can identify every provider mentioned across a set of medical records and generate a systematic subpoena strategy. This ensures no link in the treatment chain goes undocumented before the QME evaluation. Learn more about The Best AI Legal Research Tool for Workers' Comp Attorneys: Why Generic Tools Are Costing You Cases.

Speed matters on the applicant side too. Moving faster than defense counsel in securing records that establish industrial causation gives you informational leverage before the QME examiner ever picks up the file. The practitioner with complete records shapes the QME narrative. The practitioner with gaps cedes that ground. Learn more about AI Tools for Workers' Comp Medical Summaries.

Subpoena Strategy Beyond Drafting: AI as a Litigation Intelligence Tool

The subpoena is not the end of the workflow — it's the beginning of the intelligence-gathering phase. The records you obtain shape deposition strategy, QME cross-examination, and trial brief arguments. AI transforms returned records from raw document sets into structured strategic assets. Learn more about The Best AI Legal Research Tool for Workers' Comp Attorneys in 2026.

When a subpoena return arrives — often hundreds of pages of hospital records, treatment notes, and imaging reports — AI can extract diagnosis codes, treatment dates, and physician opinions on causation [SOURCE_5]. It organizes those findings in a format ready for integration into trial briefs or settlement demand letters. What used to take a paralegal a full day takes AI minutes. Learn more about How AI Reduces Malpractice Risk in Legal Research.

Cross-referencing subpoena returns against QME or AME report citations is where AI earns its place as a litigation intelligence tool. QME evaluators cite records they reviewed. If your subpoena return doesn't include records the evaluator cited, either the provider under-produced or there's a gap in your discovery. AI flags both scenarios automatically. Learn more about Legal Technology Adoption for Workers' Comp Law Firms.

From Records to Trial Brief: The AI-Powered Litigation Pipeline

The pipeline works like this. AI extracts key findings from returned medical records — ICD codes, treatment dates, prescribing history, and physician opinions on causation and work restrictions. It organizes those findings by provider and date. Then it cross-references them against the QME report to identify what the evaluator saw, what they missed, and what was cited but not produced. Learn more about AI Is Reshaping Workers' Comp Defense in 2026.

Automated gap flagging is especially valuable before a WCAB hearing. If a provider referenced in the QME report didn't produce records, you need to know before the hearing — not during cross-examination. AI surfaces that gap in the record review stage, when you still have time to act.

For legal research connecting subpoena strategy to hearing preparation, AI research tools find controlling authority on disputed medical evidence issues — for example, the admissibility of treating physician opinions versus QME findings under cases like Minniear v. WCAB [SOURCE_9]. Workers' comp AI trained on WCAB panel decisions and En Banc opinions can surface that authority in seconds rather than hours.

Speed in this pipeline creates settlement leverage. Carriers and opposing counsel move faster when they know your firm has complete records, identified the gaps in their case, and has controlling case law ready. That posture — informed, prepared, and fast — is what AI-powered firms project. Firms still running manual workflows project the opposite.

Key Takeaways

Using AI to draft subpoenas for medical records in workers' comp isn't a futuristic concept. It is a live competitive advantage that purpose-built platforms are delivering today. From eliminating the jurisdictional errors that plague manual drafting, to compressing the entire subpoena workflow from an hour to minutes, to connecting record production to QME analysis and trial preparation — AI transforms the most repetitive task in workers' comp discovery into a strategic asset.

The practitioners who adopt vertical AI tools purpose-built for workers' comp — not generic AI retrofitted with legal prompts — are defining the standard of practice for the next decade. The statutory framework is complex. The procedural requirements are exacting. The volume is relentless. And the cost of errors — delayed production, sanctions exposure, malpractice risk — is too high to absorb.

Your competitors are already evaluating these tools. The question is whether you move first or spend the next three years catching up. Start researching with CompFox and see how a workers' comp operating system built for practitioners who refuse to lose on speed actually performs in your workflow.

Frequently Asked Questions

Q: What is a subpoena duces tecum and when is it needed in workers' comp cases?

A subpoena duces tecum is a legal command requiring a person or organization to produce specific documents — in workers' comp cases, typically medical records. While a voluntary HIPAA authorization is often faster and cheaper when a provider cooperates, a formal subpoena becomes necessary in several situations: when a provider refuses to produce records, when an applicant revokes consent, or when you need a hard production deadline backed by court authority. Some facilities require a subpoena even when an authorization would legally suffice, simply to protect themselves from liability. The key strategic advantage of a subpoena is its enforceability — it creates a record for sanctions motions under Labor Code § 5813 if a provider ignores the demand. That leverage is something a HIPAA release simply cannot provide. For known problem providers, experienced workers' comp practitioners often skip the authorization step entirely and issue the subpoena first to avoid wasting time on a document that will be rejected.

Q: How is AI used to draft a subpoena for medical records in workers' comp cases?

Using AI to draft a subpoena for medical records in workers' comp cases involves deploying purpose-built legal AI tools that can generate jurisdiction-compliant subpoena documents in minutes rather than hours. Instead of manually compiling provider identifiers, formatting the document, and cross-referencing statutory requirements, attorneys and paralegals can input case-specific details and let the AI produce a ready-to-review draft. AI tools are particularly valuable for ensuring consistency across multiple subpoenas in a single file — for example, when targeting a primary treating physician, QME panel physicians, hospital records, and pharmacy records simultaneously. By automating the repetitive structural elements of drafting, AI reduces the risk of jurisdictional errors, missing provider information, and formatting inconsistencies that commonly invite objections and delay production. The result is a faster workflow that frees legal staff to focus on strategy rather than document preparation.

Q: What are the most common drafting mistakes in workers' comp medical record subpoenas?

The most common drafting mistakes when preparing subpoenas for medical records in workers' comp cases include jurisdictional errors, missing or inaccurate provider identifiers, vague record-type descriptions, and formatting inconsistencies. Vague language is especially problematic when subpoenaing mental health providers or chiropractors, who face additional privacy requirements — an imprecise description of the records requested gives these providers legal grounds to under-produce. Missing provider identifiers can cause delays in service and compliance. Formatting errors that deviate from California Code of Civil Procedure §§ 1985–1997 requirements may result in objections that push production timelines back significantly. These errors are largely preventable through careful drafting checklists or, increasingly, through AI-assisted drafting tools that flag common omissions and apply consistent formatting across all documents in a file.

Q: Why are medical records so important in workers' compensation litigation?

Medical records are the evidentiary foundation of virtually every workers' comp dispute. They inform injury causation, treatment history, and apportionment — the allocation of permanent disability to non-industrial causes under Labor Code § 4663. Without a complete medical history, no QME evaluation, AME panel report, or trial brief stands on solid ground. Applicant attorneys rely on complete treatment histories to establish industrial causation under the AOE/COE standard — meaning the injury arose out of and occurred in the course of employment. Defense attorneys need the same records to identify pre-existing conditions and apportion liability accordingly, potentially limiting the employer's financial exposure. Cases like Brodie v. WCAB have made apportionment central to permanent disability determinations, making thorough records collection a strategic necessity for both sides of a workers' comp case.

Q: Which types of providers are typically subpoenaed for medical records in workers' comp cases?

The provider universe in a typical workers' comp file is broad. Practitioners commonly subpoena primary treating physician records under the Labor Code § 4600 treatment framework, QME and AME panel physician records, emergency room and hospital records tied to the date of injury, pharmacy records relevant to apportionment arguments, and prior employer medical surveillance records or pre-employment physicals. Mental health providers and chiropractors are also frequent targets but add complexity — both involve additional privacy overlays and require specific, precise record-type language in the subpoena to compel complete production. Failing to tailor the subpoena language to these specialties gives providers a basis to withhold records or produce incomplete files, which can undermine case preparation and delay resolution.

Q: What California laws govern subpoenas for medical records in workers' comp proceedings?

Subpoenas for medical records in California workers' comp proceedings are primarily governed by California Code of Civil Procedure §§ 1985–1997, which sets out the general requirements for subpoenas duces tecum. Labor Code § 5710 provides additional authority specific to workers' compensation proceedings, granting the Workers' Compensation Appeals Board the power to compel testimony and document production. Labor Code § 4600 governs the medical treatment framework that often forms the basis for records requests, while Labor Code § 4663 addresses apportionment of permanent disability — a key reason defense teams pursue comprehensive medical histories. Labor Code § 5813 provides the sanctions mechanism that gives a properly issued subpoena its enforcement teeth when a provider fails to comply. Understanding this layered statutory framework is essential for drafting a subpoena that will withstand objection and compel timely production.

Q: Can AI replace a paralegal or attorney when drafting a workers' comp medical records subpoena?

AI is best understood as a force multiplier for legal staff, not a replacement. Using AI to draft a subpoena for medical records in workers' comp cases dramatically compresses the time required for document preparation — turning what might take a paralegal an hour into a minutes-long task — but human review remains essential. Attorneys and paralegals must still verify that provider identifiers are accurate, confirm the statutory citations are current and jurisdiction-appropriate, and ensure the record descriptions are precise enough to compel full production. AI tools reduce the risk of common errors like formatting inconsistencies and missing fields, but they don't replace the strategic judgment required to decide which providers to subpoena, in what order, and with what scope. The practical benefit is that legal professionals can handle higher case volumes without sacrificing accuracy — the AI handles the repetitive structural work while the attorney focuses on case strategy.

Q: What is the difference between using a HIPAA authorization and a subpoena to obtain medical records in workers' comp?

A HIPAA authorization is a voluntary release signed by the patient (the injured worker) that permits a provider to disclose records. It is faster, cheaper, and sufficient in many workers' comp situations where the applicant cooperates and the provider is compliant. A subpoena duces tecum, by contrast, is a court-backed legal command that does not depend on the applicant's ongoing consent. A subpoena is necessary when a provider refuses to produce records, when the applicant has revoked their authorization, or when you need a binding deadline with enforcement options. Crucially, a subpoena creates a documented record that supports sanctions motions under Labor Code § 5813 if the provider fails to comply — leverage that a HIPAA release cannot provide. For providers known to resist voluntary releases, experienced practitioners often issue the subpoena at the outset rather than cycling through an authorization that is likely to be rejected.