Every acquired site brings its own systems and habits
Different practice management systems, phone setups and referral routines mean no two sites report the same measure the same way. Leadership cannot compare them with confidence.
MSOs & Groups
One set of workflows and one reporting view across every site
We help management services organizations and multi-site groups standardize patient access, referral handling and reporting after acquisitions, with clinicians and site managers shaping the playbook.
Short answer
How can AI help an MSO or multi-site practice group?
AI helps a group close the gap between its best-run site and the rest: shared call handling, referral and fax intake, comparable reporting and denial analysis. It works only after the group agrees common definitions and owners. Clinical review stays with site clinicians, and each tool needs central governance, because tool sprawl across acquired sites is a risk.
Who this covers
Organizations in this segment
- Management services organizations supporting physician-owned or investor-backed practices
- Dental support organizations and orthodontic groups with several locations
- Multi-specialty medical groups growing through acquisition
- Private-equity-backed platforms integrating newly acquired practices
- UK private clinic groups and multi-site primary care operators
Challenges
The business problems that hold this segment back
Tool sprawl with no owner
One site buys a scribe, another a reminder tool, a third a chatbot. Nobody holds the inventory, the business associate agreements differ, and patient data flows are unmapped.
Central patient access can feel like a threat
Consolidating calls promises consistency, yet front desk teams fear replacement and patients expect their own practice to answer. Poorly handled, central access hurts the relationship it was meant to protect.
Revenue cycle variation hides inside the averages
Denial patterns, coding habits and payer terms differ by site, so days in accounts receivable and first-pass claim acceptance are not comparable until definitions are shared.
Accountability sits across several entities
Privacy, clinical and billing responsibility must land at the correct entity, which is harder when records, billing and referral workflows are centralized.
Where AI fits
Suitable AI, and how mature each application is
Established means widely deployed and well documented. Emerging means real deployments exist but results vary with your data and workflow. Early means pilots or research, not a safe default.
Shared patient access desk
Bounded routing and booking support has evidence from single practices, and a central team using AI-assisted routing of messages and calls can serve several sites, though multi-site results are not yet shown. Local scheduling rules per site must be configured first.
Referral and fax intake
Document intelligence can classify scanned referral packets, extract fields and match patients, with an exception queue for items it cannot match. Supplier-reported matching rates show why that queue is essential.
After-hours voice across sites
Voice agents are credible for verification, routine questions and narrow appointment actions, but autonomous booking is still emerging. Each site's slot rules and a human handoff must be tested first.
Reporting definitions and exception summaries
AI can help map different site fields onto shared definitions and write plain-language summaries of what changed. The mapping must be reviewed by whoever owns the metric, because data quality sets the ceiling.
Denial pattern analysis and appeal drafts
Cross-site analysis can show which payers and codes drive denials, and draft appeals for a billing specialist to edit, as billing companies do. Only about one in five providers used AI for denials management in a February 2026 HFMA report, so check what your vendors already offer before you build.
Configuration drift and AI-use monitoring
Watching for sites that quietly change scripts, tools or settings is attractive, but proven drift monitoring across acquired networks is thin. Start with a simple register and periodic human audit.
What stays human
Responsibilities that stay with your people
Chief operating officer or operations lead
Decides what is centralized and what stays local, and owns the common definitions every site reports against.
Site medical director or lead clinician
Reviews clinical content, approves local workflow changes and can stop a tool that fits their specialty poorly.
Compliance and privacy officer
Keeps the AI inventory, checks business associate agreements and risk analyses, and reviews referral or marketing payment arrangements.
Revenue cycle director or billing manager
Approves rules, samples audited claims and signs off appeals that AI has drafted.
Practice manager at each site
Runs local adoption, reports exceptions and tells the center when a process does not match reality.
Improvements
Operational changes you can expect to work on
Consistent speed of answer across sites
Lead response time and call abandonment are measured the same way everywhere, so weak sites are visible early.
One view of performance
No-show rate, referral conversion and first-pass claim acceptance sit on shared definitions that leadership and lenders can trust.
Faster onboarding of a new location
A tested playbook of tools, scripts, permissions and training shortens the time before a new site works like the rest.
Less referral leakage
Every referral is logged from receipt to first appointment, so lost referrals are noticed and chased.
Commercial value
Growth from variation you can see
In a group, commercial value usually comes from narrowing the gap between sites: same-site growth through faster access and rebooking, referral capture through consistent intake, and less revenue lost to billing variation. Revenue operations makes the gap measurable, and people have to set the standard first.
- Make lead response time and no-show rate comparable site to site before judging any tool.
- Capture referrals centrally, then track conversion by referring practice and site.
- Run central marketing with local booking, using data integrations that keep patient details out of ad platforms.
- Give investors and lenders reporting built on common definitions rather than site spreadsheets.
Rules and limits
US and UK rules that apply, and what is not ready yet
General information for planning, not legal advice. Statuses and dates are as published at the time of writing.
United States
- HIPAA (existing rules apply now): each AI tool processing PHI needs a permission basis, a risk analysis and, where the vendor is a business associate, an agreement. The January 2025 Security Rule proposal is not final, with a July 2027 target in the regulatory agenda.
- HHS-OIG Advisory Opinion 26-15 (June 2026) found a payment arrangement for online referral software unfavorable under the federal Anti-Kickback Statute. It turned on particular facts and does not cover all MSO software, but referral and marketing payments deserve review.
- Texas SB 1188 (EHR storage provisions from 1 January 2026): covered EHRs must be physically maintained in the US or a US territory, which matters for shared services using offshore support or hosting.
- CMS-0057-F (payer duties from 1 January 2026, APIs from 1 January 2027) changes what prior authorization timelines to expect from affected payers.
United Kingdom
- UK GDPR as amended by the Data (Use and Access) Act 2025 (most changes commenced 5 February 2026): a shared service needs an Article 6 basis and an Article 9 condition for health data, plus a DPIA where required.
- ICO guidance on outsourced AI: assess controller and processor roles and document subcontractor arrangements. A central team handling site patient data is exactly this case.
- CQC (AI expectations page updated 21 May 2026) does not approve individual AI products but expects human oversight, staff training, risk management and governance from registered providers. Its March 2026 draft sector frameworks remain draft.
Not mature or not suitable yet
- Proven post-acquisition margin gains. A September 2025 MGMA poll found 68% of responding medical groups added or expanded AI, but that is medical-group evidence, not a private-equity roll-up result.
- One AI product for every specialty. Site-by-site integration, clinical review and drift monitoring are harder to verify than a license or a pilot.
- Agents from different vendors coordinating work across sites on their own. Cross-vendor agent protocols are established, but healthcare use is still emerging.
Services
The services that fit this segment
01AI StrategyChooses which workflows to standardize first and which tools to keep, retire or never buy.02AI GovernanceBuilds one inventory, one risk tiering method and one set of vendor terms for every site.03Data & IntegrationsConnects mixed practice management systems, phones and forms so reporting means the same thing everywhere.04Revenue OperationsStandardizes lead handling, referral tracking, recall and reporting across locations.05AI TrainingGives front desk, billing and site managers role-based training so the new routine sticks.
Where to start
Where we would begin
- 01
Inventory sites, systems and tools
One page per site: systems, phone setup, AI tools in use, vendor agreements and who owns each. Duplicates and gaps become obvious.
- 02
Agree the common measures
Set shared definitions for lead response time, no-show rate, days in accounts receivable, first-pass claim acceptance and referral conversion, and compute them the same way at every site.
- 03
Pilot a shared service with two or three sites
Referral intake or after-hours handling is a good first choice. Compare pilot sites against a baseline before extending the service, and keep dental practices and medical sites separate if their workflows differ.
Questions
Frequently asked questions
Should an MSO pick one AI vendor for every site?
Not by default. Standardize the workflow and the data first, then choose tools that fit it. A single vendor can still be wrong for a specialty, and switching later is costly.
How does AI help when sites run different EHRs?
It helps with document intake and reporting if the integration layer is built carefully. Integration is the most-cited barrier to scaling AI in a McKinsey survey, so we plan it before any tool.
Does central patient access mean replacing site front desks?
No. It moves repeatable work such as overflow calls and after-hours messages to a shared team, while site staff keep relationships and in-person care. We involve front desk leads in the design.
Is there proof that AI improves margins after acquisitions?
Not yet. Evidence for private-equity roll-ups is thin, so we measure site by site against a baseline instead of promising a margin result.
How do you govern AI across several legal entities?
With one inventory and risk tiers, plus clear notes on which entity is responsible for privacy, clinical and billing decisions. Our AI governance work draws on NIST AI RMF and ISO/IEC 42001 as reference frameworks, alongside your counsel.
Pick the first shared workflow
Tell us how many sites you run and where they differ most. We will suggest one shared service worth piloting and the measures to judge it.