What Helps and What Hurts
By Evan M. Gwilliam DC MBA CPC CCPC CPMA CPCO
Senior Vice President at Practisync.
Artificial intelligence is no longer a “someday” issue for chiropractic. It is already showing up in electronic health record (EHR) templates, predictive text, claim scrubbers, coding prompts, clearinghouse edits, prior authorization systems, payer analytics, and even ambient listening tools that can draft a visit note from the treatment room conversation.
That does not mean every chiropractor needs to become a tech expert. But, as health care providers, chiropractors do need to be proficient in compliance and documentation, and AI is impacting those fields in ever changing ways.
The promise is real. AI can help organize SOAP notes, clean up clunky language, identify missing documentation elements, summarize prior visits, draft appeal letters, build training materials, and help clinics create policies for HIPAA, OSHA, and OIG compliance. Used correctly, it can make a good compliance workflow faster.
But AI does not create compliance. The clinic’s workflow does. That is the point too many vendors skip while they are busy sprinkling AI glitter on everything like it came from a technology craft store. AI can improve clinical narratives, goal tracking, documentation completeness, and workflow efficiency, but it also creates risks such as hallucinations, cloning, overdocumentation, unsupported coding, and compliance violations.
Advertisement
The first major risk is hallucination. Large language models do not “know” facts the way clinicians do. They predict likely text. That means an AI tool can confidently insert an orthopedic test, range of motion measurement, treatment response, spinal level, symptom duration, or patient statement that never happened. In October 2024, the Associated Press reported concerns that one major AI company’s transcription tool, used in some medical settings, could fabricate text in transcripts. That is not a small problem when the transcript becomes the starting point for a legal medical record. (AP News)
The second risk is bias. AI models reflect their training data. If the broader healthcare data ecosystem underrepresents chiropractic, conservative care, physical medicine, and functional improvement models, AI output may lean toward medical or pharmaceutical assumptions. That could affect treatment suggestions, documentation language, coding prompts, or utilization review logic. The safeguard is simple: use AI to improve documentation quality and administrative efficiency, not to replace clinical reasoning.
HIPAA is the next big issue. If a vendor receives, maintains, creates, or transmits PHI on behalf of a covered entity, the practice generally needs a Business Associate Agreement. HHS guidance on cloud services states that covered entities and business associates must obtain satisfactory assurance through a BAA when a cloud service provider handles ePHI. The same basic logic applies when an AI vendor touches PHI. (HHS.gov)
That means staff should not paste patient names, dates of birth, visit details, or chart excerpts into consumer AI tools unless the tool is approved for that use and a BAA is in place. “But I only did it once” is not a compliance strategy. It is a confession.
Ambient listening deserves special attention. These tools can record the patient encounter and generate a structured draft note. The upside is obvious: less typing, better eye contact, and less after-hours charting. Research published in NEJM Catalyst in 2024 reported that ambient AI scribes reduced documentation burden and improved physician-patient communication in a large implementation. (Kaiser Permanente Division of Research)
But the compliance guardrails matter. Clinics should obtain clear patient consent, offer a non-recorded option, verify state recording laws, confirm BAA coverage, understand retention settings, and require provider review before signing. The AI draft is not the chart. It is a draft. The signed note is the chart, and the provider owns every word.
Documentation and coding are where chiropractors need to be especially careful. AI can help format subjective complaints, tie symptoms to activities of daily living or functional deficiencies, organize objective findings, create measurable goals, and make the medical necessity story clearer. That is useful. It can also suggest CPT codes, improve ICD-10 specificity, help with modifiers, and draft appeal language.
But coding suggestions must be treated as suggestions, not instructions. AI should never be allowed to upcode a visit, add unsupported services, invent time, or use modifiers creatively to bypass edits. The sequence should be: document what happened, determine what is supported, then verify the code. Not the other way around.
Payers are moving fast too. CMS has clarified that Medicare Advantage organizations may use algorithms and AI to assist coverage determinations, but those tools may not override applicable coverage and medical necessity rules. (AAMC) CMS has also pushed electronic prior authorization modernization, emphasizing better exchange of information among payers, providers, and patients. (Centers for Medicare & Medicaid Services) In 2026, UnitedHealthcare announced it would remove prior authorization requirements for about 30% of services, with reporting indicating that some chiropractic visits may be included. That sounds encouraging, but it does not mean scrutiny is going away. It means payers are getting more selective and more data-driven. (The Wall Street Journal)
So what should a chiropractic clinic do now?
Start with an AI inventory. Identify every AI-enabled tool in the practice, including EHR features, dictation, scheduling, claim scrubbers, coding tools, chatbots, analytics, and anything staff use informally. Then create a written policy (ask AI to create a draft for you) that answers five questions: which tools are approved, whether PHI is allowed, whether a BAA exists, who reviews AI output, and how errors are handled.
Next, use AI where it strengthens compliance. Ask it to help draft HIPAA training, build an OSHA checklist, create an OIG compliance training outline, identify vague treatment goals, flag cloned language, or summarize documentation gaps in a self-audit. Keep PHI out unless the tool is approved and covered. Do not use AI to “fix” old notes. Use it to improve future behavior.
Finally, connect documentation to billing. Clean billing starts with clean documentation. If the note does not support the service, the claim has a problem before it ever reaches the payer. Practisync, a third-party billing service that is an affinity partner to your state association, keeps its billing work focused around that reality. Claims are not just pushed out the door. Practisync helps clinics create a cleaner, more defensible revenue cycle.
AI is not going away. The smart move is not to panic or pretend it does not exist. The smart move is to use it like a sharp instrument: carefully, intentionally, and with both hands on the handle.
Evan M. Gwilliam DC MBA CPC CCPC CPMA CPCO is Senior Vice President at Practisync. Practisync provides third-party billing and compliance-oriented revenue cycle support for chiropractic practices, including discounted options for members of participating state association affinity partners. To learn more, visit Practisync.com and complete the discovery form.
