Medical Coders and Health Records Specialists

Impact

Computer-assisted coding has matured into autonomous coding: language models read the clinical documentation and assign diagnosis and procedure codes for routine encounters in radiology, pathology, emergency and outpatient settings with no human coder on the chart. Other systems check documentation for specificity, generate physician queries, prepare claims and work denials. The specialist's work shifts from reading notes and assigning codes to auditing the system's output, coding the complex inpatient and surgical cases it cannot handle confidently, managing documentation-improvement programmes and overseeing the integrity, release and retention of the record itself.

Risk

Very high exposure: routine code assignment is automating now; value shifts to audit, complex cases and records governance.

The core task of reading documentation and assigning codes is being automated for routine encounters today, along with charge capture, claim scrubbing, abstracting and much of the release-of-information workflow. What remains human is coding complex inpatient stays, surgical cases and unusual presentations, auditing automated output for compliance and revenue accuracy, writing and managing physician queries, interpreting payer rules that change constantly, and the governance of patient records, including privacy, amendments and legal requests. Observed usage of language models by this occupation is among the highest of any measured role and official measures place it in the top exposure tier, which is why the score sits in the very high band. Over the 0-3 year window expect sharp reductions in production coding headcount at health systems that adopt autonomous coding, with remaining roles reclassified towards auditor, documentation specialist and data-integrity titles.

Sector readiness

Autonomous Coding Moving Into Production

Large health systems and revenue-cycle outsourcers have moved from computer-assisted coding to autonomous coding for high-volume specialties, with vendors such as Nym, Fathom and CodaMetrix in production alongside the established Solventum (formerly 3M) and Optum platforms, and Epic embedding AI into its revenue-cycle tools. Smaller hospitals and physician practices still rely on human coders working with encoder software, but their billing companies are adopting the same tools. Payer-side automation of claim review is advancing in parallel, which raises the stakes on documentation accuracy.

1. 01

   Become the auditor. Pursue an auditing credential and ask to review samples of autonomously coded encounters. Every organisation that automates coding needs people to check it, and that work pays better than production coding.

2. 02

   Go where the machine is weakest. Complex inpatient, cardiothoracic, oncology and trauma coding involve judgement across long, messy records. Build depth there and you buy years.

3. 03

   Learn clinical documentation integrity. Querying physicians to document specificity and severity is a growing speciality that sits between clinical and coding worlds and is harder to automate.

4. 04

   Understand how the model codes. Ask vendors and your revenue-cycle leaders how confidence thresholds and exception routing work. Knowing why a chart was sent to you helps you work faster and spot systematic errors.

5. 05

   Pivot towards records governance. Privacy, legal requests, amendments and data quality are health information management functions that expand with digitisation. A degree-level credential in HIM opens them.

6. 06

   Do not wait. The window for this role is the shortest in this batch. Start the credential, the conversation with your manager and the move towards audit or specialty work this year, not after the next restructuring.

1. 01

   Autonomous coding engines. Nym, Fathom and CodaMetrix read clinical documentation and assign codes for routine encounters at confidence levels that let organisations skip human review entirely.

2. 02

   Computer-assisted coding and encoder platforms. Solventum 360 Encompass and Optum systems suggest codes and documentation gaps across the chart, raising coder productivity and reducing the number of coders needed.

3. 03

   EHR-embedded AI. Epic and other electronic health record vendors are building coding suggestions, charge capture and denial prediction into the systems coders already work in.

4. 04

   Ambient clinical documentation. Tools that draft the physician note from the visit also structure the content for coding, making downstream automation easier.

5. 05

   Revenue-cycle cost pressure. Hospitals under margin pressure outsource and automate coding aggressively, and outsourcers compete on automation rates.

6. 06

   Payer automation and audit. Payers automate claim review and denials, forcing providers to invest in accuracy, auditing and documentation improvement, which are the human tasks that remain.

Large health systems

Earliest adopters of autonomous coding for radiology, emergency and outpatient volume; production coder roles are shrinking fastest here while audit and CDI roles grow.

Physician practices and clinics

Smaller practices still rely on human coders or billing companies, but those companies are deploying the same autonomous tools across their client base.

Revenue-cycle outsourcers

Compete on automation rate and offshore labour, so onshore production coding is under the most pressure and quality-assurance roles are the main survivors.

Payers and government programmes

Use coding expertise for audit, risk adjustment and policy, a growing destination for experienced coders who want to move away from provider-side production.

1. 01

   Coding audit and compliance. Validating automated and human coding against documentation and regulation is the central human task now. Auditing credentials from AAPC or AHIMA are the standard route.

2. 02

   Complex inpatient and specialty coding. Long, multi-condition records with procedural nuance are where autonomous systems defer to people. Deepen expertise in a demanding specialty.

3. 03

   Clinical documentation integrity. Reviewing records concurrently and querying physicians for specificity requires clinical knowledge and tact. CDI certification is a recognised next step for coders.

4. 04

   Health information governance. Privacy, release of information, legal holds and data quality are growing responsibilities. A health information management credential opens these roles.

5. 05

   Payer rules and denial management. Understanding why claims are denied and how to appeal them is work that changes faster than models are retrained. Track payer policy updates in your specialty.

6. 06

   Data literacy and reporting. Coded data feeds quality measures, risk adjustment and research. Being able to query it and explain anomalies positions you as a data professional, not a clerk.

Autonomous radiology codingExample 1

How

A health system routes all radiology reports through an autonomous coding engine that codes the large majority without review and sends the rest, flagged by low confidence, to the coding team.

Gain

Coder time concentrates on genuinely ambiguous reports and turnaround on clean claims drops to hours.

Audit of automated outputExample 2

How

A senior coder samples autonomously coded emergency encounters weekly, scores accuracy against documentation and feeds systematic errors back to the vendor and the compliance team.

Gain

The organisation gets the speed of automation with documented controls that satisfy auditors.

AI-generated physician queriesExample 3

How

A CDI specialist uses a documentation platform that drafts specificity queries from the chart, reviews and personalises them, and tracks physician response rates.

Gain

More documentation gaps are closed before billing with less specialist time spent writing queries.

Data Entry KeyersMore exposed · exposure 83

AI impact

Pure data entry from documents is being eliminated by extraction and automation with none of the clinical judgement or compliance oversight that remains in coding.

Work moves to

Data entry roles are converting to exception handling and quality checking where they survive at all.

Healthcare AdministratorsDifferent skills, growing · exposure 42

AI impact

AI streamlines reporting and operations, but managing people, budgets and regulatory relationships keeps the role human and demand is growing.

Work moves to

Coders with revenue-cycle and compliance knowledge are a natural pipeline into health information and revenue-cycle management roles.

Medical AssistantsComplementary, less exposed · exposure 26

AI impact

AI handles documentation and scheduling tasks, but patient-facing clinical support remains hands-on and in demand.

Work moves to

Medical assistants increasingly capture the structured information that coding systems depend on, linking the two roles more closely.

Our report is one reading of the evidence. This section is the other dataset: what people who do or know this work make of it. Nobody has scored this role yet. Sign in to add yours.