Home/AI Suite/AI for Eligibility Verification
Continuous 270/271 · Sub-300 ms P95 · Mid-cycle change detection

Eligibility you check continuously, not at intake.

A one-time eligibility check at intake tells you what was true on the appointment date. By the time the claim ships 30 days later, the plan may have terminated, the deductible may have reset, or the patient may have migrated. Our AI runs continuous 270/271 sweeps on every active patient, surfaces material coverage changes the day they happen, and refreshes the verification the night before every appointment. The result is 99.4 percent of active coverage caught at the moment of service.

Get a free eligibility audit → See the AI
1,200+ payer connections 30 Medicaid state adapters Per-service-code benefit detail
What the 271 returns

Per service code. Parsed and structured.

Most eligibility tools give you an eligibility-yes-or-no result. Our model parses the full 271 response into structured benefit detail per service code: in-network and out-of-network deductible status, copay amounts, coinsurance, out-of-pocket max, prior-auth requirements, specialist referral rules, mental-health and substance-use parity exceptions, and ABA-specific authorization rules where applicable.

The card to the right is the live parsed view a front-desk team sees on intake. Parents see their actual financial responsibility before session one, not after the first statement arrives in 35 days.

Benefit detail · 97153 ABArefreshed 14 min ago
Coverage status
StatusActive
Policy end2026-12-31
Deductible · individual
Amount$1,500
Met YTD$842
Cost share · 97153
Copay$25
Coinsurance20%
Authorization · 97153
RequiredYes
Active auth320 units
Out-of-pocket max
Amount$6,500
Met YTD$1,624
Per-CPT detail · cached freshness window 24h
The capabilities

Six capabilities. Coverage stays current.

Each capability is real production code, validated by certified intake and billing specialists, accountable to a written SLA. Eligibility is the source-of-truth that everything downstream depends on; we treat it that way operationally.

01 · Continuous sweep

Daily background 270/271.

Daily background sweep on every active patient. Material coverage changes surfaced within 24 hours of payer-side update. Pre-appointment refresh runs the night before every scheduled visit so the front-desk team has current detail at check-in.

02 · Service-code detail

Per-CPT benefit parse.

The 271 is parsed into structured fields per service code: deductible status, copay, coinsurance, OOP max, prior-auth, referral rules, parity exceptions. Front-desk teams see the financial responsibility per CPT, not just eligibility-yes-or-no.

03 · Mid-cycle catch

Plan migrations surfaced fast.

Patients migrate between policies as employer brokers change. The AI surfaces material coverage changes within 24 hours and rebills any in-process claims under the corrected policy. Self-pay AR reclassification typically lifts 6 to 11 percent annually.

04 · Medicaid adapters

State-specific response parsing.

State-specific adapters for the top 30 Medicaid programs (92 percent of national enrollment). Translates raw 271 responses into normalized fields. Schema updates roll within 72 hours when a state changes its response format. Generic tools that treat Medicaid as homogeneous miss this entirely.

05 · Auth surfacing

Prior-auth requirements caught.

The 271 carries prior-auth flags by service code. Our model surfaces auth requirements at intake so the front-desk team obtains authorization, not just verifies coverage. Eliminates the single most common auth-related denial: service rendered without active auth on file.

06 · Discovery loop

Self-pay rescan monthly.

Self-pay accounts get scanned against 1,200+ payer rosters monthly via Insurance Discovery. Discovered coverage flows into the continuous 270/271 sweep so it stays current. The discovery hit rate on previously declared self-pay is typically 11 percent in our active book.

The sweep cadence

A patient's eligibility, every day this week.

The view to the right is one active patient's eligibility activity in a single week. Daily sweep runs are quiet when nothing changes. A material change (deductible reset, plan migration, auth status update) emits an alert and queues a downstream action. The front-desk and billing teams see what happened and what to do about it; they do not have to remember to re-run a check.

Most eligibility tools treat the 270 as a one-shot intake step. We treat it as the continuous data feed that everything downstream depends on. The cost difference at scale is roughly 6 to 11 percent of annual AR.

Patient eligibility · activity stream · last 7 days
Mon · 09:14
Daily sweep · no changeActive commercial, deductible at $842
Auto
Tue · 09:14
Daily sweep · no changeActive commercial, deductible at $842
Auto
Wed · 09:14
Daily sweep · auth status updateAuthorization unit balance decremented to 312/320
Alert
Thu · 09:14
Daily sweep · plan migration detectedMember migrated to plan ID PPO-2026A, effective Thu
Action
Thu · 09:42
Auto-rebill triggeredIn-process claims rebilled under new policy
Auto
Fri · 09:14
Daily sweep · new policy confirmed$1,500 deductible, $25 copay, prior auth required
Auto
Sat · 23:30
Pre-appointment refresh · Sunday visitCoverage active, auth balance verified
Auto
7 sweeps · 1 plan migration caught · 1 auto-rebill
What clients see

Measured outcomes from continuous eligibility.

Across the active book. Anonymized; individual results depend on payer mix and prior intake discipline.

99.4%
Coverage caught at intake
Active coverage correctly identified at the moment of service across the active book. The remaining 0.6 percent is typically payer-side data latency on plan migrations that resolves within 24 hours.
<300ms
270/271 P95 latency
Through our gateway with caching layer. Clearinghouse baseline runs about 1.2 seconds. The 4x speedup matters at scale; eligibility runs hundreds of times per day per location.
11%
Self-pay AR recovered
From Insurance Discovery rescan on previously declared self-pay accounts. Coverage flows into the continuous 270/271 sweep so AR reclassifies cleanly to the correct payer for back-billing.
Common questions

Frequently asked questions: AI for eligibility.

What is 270/271 and why does it matter for AI eligibility?
270 is the X12 eligibility inquiry sent to a payer; 271 is the response. The pair is the standardized way to verify a patient's active coverage, benefit detail, deductible status, copay, and authorization requirements. The challenge is not running a single 270/271; it is running them continuously as plans change, deductibles reset, and patients migrate between coverage. Our AI runs continuous 270/271 sweeps with intelligent reschedule logic instead of one-shot intake checks.
How is continuous eligibility different from a one-time check at intake?
A one-time check tells you what was true on the appointment date. By the time the claim ships 30 days later, the plan may have terminated, the deductible may have reset, the prior-auth requirement may have changed, or the patient may have migrated to a different policy. Continuous eligibility runs a daily background sweep on every active patient, surfaces material changes the day they happen, and re-runs the 270 the day before every scheduled appointment. The result is a 99.4 percent rate of catching active coverage at the moment of service.
What is the round-trip latency on a 270/271?
Under 300 ms at the 95th percentile through our gateway. The clearinghouse-level latency varies by payer, but our caching layer holds recent responses and emits them synchronously when the question is repeated within the freshness window. Stale entries get re-run in the background, never blocking the user. CFOs care about this number because every second of latency in the patient-intake workflow compounds across thousands of appointments per month.
What kind of benefit detail comes back?
Active coverage status, policy effective and end dates, in-network and out-of-network deductible status, copay amounts, coinsurance percentages, out-of-pocket maximums, prior-auth requirement by service code, specialist referral requirements, mental-health and substance-use parity exceptions, and ABA-specific authorization rules where applicable. Per appointment, per service code, parsed and structured. Not just an eligibility-yes-or-no.
How do you handle Medicaid eligibility, which is messy by state?
Each state Medicaid plan has its own quirks, error patterns, and eligibility refresh cadence. We carry state-specific adapters for the top 30 Medicaid programs (covering 92 percent of national Medicaid enrollment) that translate raw 271 responses into normalized fields. Updates roll out within 72 hours when a state changes their response schema. Generic clearinghouse tools that treat Medicaid as one homogeneous source miss this entirely.
What about mid-cycle coverage changes and plan migrations?
This is the highest-value catch. A patient on commercial coverage in January often migrates to a different policy by July as their employer changes brokers or their personal circumstances shift. The AI surfaces material coverage changes within 24 hours of a payer system update and rebills any in-process claims under the corrected policy. The dollar value of this single capability is typically 6 to 11 percent of self-pay AR reclassified to billable annually.
Does this connect to your eligibility-discovery model?
Yes. Patients who present as self-pay are run through Insurance Discovery, which scans 1,200+ payer rosters monthly to find active coverage the patient did not declare. Discovered coverage then flows into the same continuous 270/271 sweep so it stays current. The discovery hit rate on previously declared self-pay accounts is typically 11 percent in our active book, which is a material self-pay AR reduction.
How does AI prevent eligibility-related denials specifically?
Three ways. First, the 270 is re-run before every claim submission with a freshness gate; stale eligibility data does not ship. Second, the AI surfaces prior-auth requirements by service code at intake so the front-desk team knows to obtain auth, not just verify coverage. Third, the model classifies eligibility responses as deterministic-fail (clear termination), deterministic-pass (clear active), or ambiguous (payer system delay, missing fields), and routes ambiguous cases to specialists rather than letting them ship.

Send the appointment file. We send back the coverage gaps.

A free eligibility audit. Drop 30 days of appointments. We run continuous 270/271 sweeps across your payer mix and return a four-page audit covering coverage capture rate at intake, discovered self-pay coverage with dollar values, plan migrations missed, and a 90-day fix plan. A senior partner on the call.

Get my free eligibility audit See the tech