Case study / Employer outcome

Databricks Claims Forensics / Referral Leakage

Claims intelligence translated into service-line, provider-network, and account-target decisions.

Evidence register

What this case can support.

Evidence class: Employer outcome
Claim boundary: Aggregate results are reported in the September 12, 2026 executive resume. The $125.9M leakage figure is a modeled opportunity and $22M downstream LTV is an estimate, not realized or booked revenue. Account-level and patient-level records remain private.
Source basis: September 12, 2026 executive resume

Case architecture

Ecosystem thesis

The leakage work was not a data exercise. It was an ecosystem translation problem: claims data had to become market structure, provider behavior, payer logic, service-line priority, and field action.

System path

01Data lake
02Leakage map
03Growth thesis
04Account targets

Executive decision brief

CEO question

What system did this work make more launchable, fundable, or scalable?

Operating answer

Claims analytics becomes valuable when it changes commercial action: which accounts to pursue, which services to launch, and where referral leakage can be captured.

Proof to inspect

The leakage signal matters because it was connected to action: claims-record context, service-line focus, provider targeting, activated patient flow, and downstream value logic.

Ecosystem context

The outcome only makes sense inside the system around it.

Healthcare leakage is rarely visible from one system. It hides across referral patterns, payer mix, specialty utilization, out-of-network behavior, access gaps, and disconnected provider relationships. A static dashboard may describe the problem but still fail to change where the team spends time.

The strategic value came from turning claims records into a shared commercial map: where patients were leaving, which specialties mattered, which provider corridors could be influenced, what revenue was addressable, and which accounts deserved a different operating motion.

For founders, this is the difference between analytics as reporting and analytics as GTM infrastructure. The data only matters if it changes segmentation, ICP, account prioritization, service-line focus, field cadence, and proof of value.

Stakeholder tensions

The people and pressures shaping the work.

Executive Team

Needed a financeable view of opportunity size and prioritization.

Opportunity dollars are only useful when they can be converted into accountable decisions.

Service-Line Leaders

Needed to know which specialties and patient flows could support growth.

Not every high-volume category is strategically attractive or operationally capturable.

Provider Network

Created or blocked referral capture through real-world relationship and access patterns.

Leakage is often a relationship and workflow problem, not only a data problem.

Payers

Shaped reimbursement, steering logic, network behavior, and value proof.

The commercial story had to connect to payer economics, not just internal revenue goals.

Field Teams

Needed account targets, talk tracks, and service-line priorities they could act on.

Insight dies when it is not translated into a usable operating motion.

Outcome record

The proof signals attached to the case.

analyticsClaims intelligence

Leakage identified

$125.9M modeled claims-leakage opportunity; opportunity sizing, not realized revenue.

analyticsEnterprise context

Claims scale

3B+ claims records, 1,362 provider relationships, and 163% referral growth in three months.

growthPatient flow

Patient activation

Approximately 4,000 patients activated through referral routing and access redesign.

growthValue proof

Downstream LTV

Estimated $22M downstream lifetime-value opportunity; not booked revenue.

Interoperability map

How the layers connect.

The case is designed as an operating ecosystem: signal, economics, workflow, proof, and expansion are connected rather than treated as separate workstreams.

Data Signal

Where is leakage occurring?

Claims records were structured into leakage, utilization, specialty, payer, and referral-corridor views.

Market Map

Which leakage is addressable?

The work separated raw opportunity from plausible capture based on service line, provider behavior, and payer context.

GTM Translation

Who should the team pursue?

Claims intelligence became account priorities, provider targets, and service-line growth theses.

Proof Loop

How do leaders know the strategy is working?

Patient activation, downstream value, and referral capture became the evidence layer behind the commercial motion.

Operating record

The work, the sequence, and the strategic read.

The record separates the conditions, operating moves, interpretation, and repeatable lessons so the result can be evaluated without flattening the work into a headline.

Challenge

Referral leakage and specialty opportunity were invisible across payer segments, provider corridors, and care pathways.

Approach

Built a Databricks claims-intelligence framework across 3B+ records, then translated modeled leakage into strategic accounts, referral corridors, and service-line priorities across Article 28 care and adjacent home/community services.

Founder takeaway

Claims analytics becomes valuable when it changes commercial action: which accounts to pursue, which services to launch, and where referral leakage can be captured.

Strategic read

The high-level point is that data lakes do not create strategy by themselves. The operator job is to turn raw signal into a decision architecture that executives, field teams, and service-line owners can use without needing to become data scientists.

Proof interpretation

The leakage signal matters because it was connected to action: claims-record context, service-line focus, provider targeting, activated patient flow, and downstream value logic.

Operator moves

Structured claims data into usable commercial intelligence instead of static reporting.
Mapped leakage by specialty, payer segment, and referral corridor.
Converted insights into named account targets and service-line opportunities.
Connected claims opportunity to downstream LTV, contribution margin, and provider outreach priority.
Built MarketScape / 90-day GTM architecture for RoadL home visits and HouseCalls, connecting AWV, TCM, HEDIS/Stars, risk capture, and homebound primary care.
Supported $3.2M+ value-based agreements and translated Article 28/CON, HCBS, home health, 340B/APG, licensure, payer, workforce, and capital constraints into expansion and M&A decisions.

Expansion path

01
Define the leakage question in business language before modeling.
02
Separate total opportunity from capturable opportunity.
03
Translate leakage into service-line, provider, payer, and account lanes.
04
Give field teams a focused target list and operating cadence.
05
Measure whether referral capture, patient activation, and downstream economics move.

What I would do again

Start with the decision model before building dashboards.
Separate verified dollars from opportunity ranges.
Use claims intelligence to govern field motion, not just inform strategy slides.

What this proves

Azis can turn a healthcare data lake into a GTM operating system founders can use.

Evidence objects

Proof artifacts

Current executive resume

Start a serious conversation

Build the wedge. Prove the motion. Scale what repeats.

For Series A/B teams that need sales, partnerships, implementation, payer logic, and revenue intelligence to become one operating system.

Contact Azis Download Resume