Unnat Patel

Work · Brief · 2 minutes

The decisions behind the work.

Explore claims reconciliation, pharmacy operations and procurement adoption through the decisions and evidence. Cases omit internal names and confidential details. Outcome figures are real and stated in headline form. Retrospective analysis is labeled.

What

The missing claims: 300,000 claims and about $10M found by asking what never reached the report

While testing claim files I noticed some never reached the audit table. I extended the analysis to three years, 20 million claims and 50 GB of data, found 300,000 missing claims worth about $10M, and put a daily reconciliation audit in place.

Healthcare projects · From my experience

Three projects. Three different decisions.

Supporting healthcare documents →

Claims intake automation

400,000+ paper claims converted to EDI 837; an 80-person team redeployed; about $1M a year saved.

From scanned CMS-1500 and UB-04 forms to validated EDI transactions and auto-adjudication.

Provider resolution

A match the next system can trust.

Provider matching moved to the provider master, with a 99% match in the proof of concept.

An integration platform · the legacy claims integration replaced

A legacy claims integration replaced.

I led the replacement of Microsoft BizTalk with homegrown EDI shredding into a normalized datamart before adjudication. Complete claim data reduced enhancement turnaround by 50% and improved received-to-paid mapping and CMS reporting.

Who

Applied AI · Featured technical case

Agentic customer care

The technical design: access and authorization before customer context or AI, routing, retrieval, authorized actions, both escalation paths, reconciliation and reviewed learning.

How

Automating state controlled-substance reporting: rules for the known, people for the exceptions

Every controlled-substance dispense must be reported to state monitoring programs, and the reporting was manual. I led the automation: 2,250 daily reports to all 50 states and DC, about 5 hours of manual work a day eliminated, and zero compliance gaps.

So what

Case studies

Rebuilding the drug-safety check to run in real time

Drug utilization review is a clinical-safety check in the pharmacy workflow. I co-led the product work to replace the asynchronous path with a synchronous service, with historical replay and staged rollout described in the launch account.

A safety check before dispense, launched through a phased rollout

I led the requirements and phased rollout of a safety check closer to dispense, aligning readiness with clinical and operational partners.

Deep dive · 3 minutes

Work: Deep dive

Cutting alert noise without weakening the safeguard

I analyzed alert categories and designed eleven proposed optimizations across three themes: reduce avoidable alerts, surface useful clinical context, and consolidate review. The proposal models reduced workload while preserving clinical coverage. Those modeled benefits are separate from the narrower dose-check fix reported as shipped. The broader program is in progress.

One console for pharmacy customer care: about three minutes saved per contact in the pilot

Amazon Pharmacy customer care worked across a legacy agent tool plus chat and voice systems, while pharmacists and technicians juggled many separate tools. I was the product manager for a unified care console: I flipped the launch sequence to protect continuity of care, piloted it, and then shipped the capabilities that removed the biggest sources of transfers and hold time. The pilot saved about 3 minutes per contact across more than 6,000 contacts a month. AI-assisted routing saved 21 minutes per routed contact and eliminated about 1,000 transfers a month.

Sequencing an MDM migration around the systems that depend on it

Leadership wanted a source-by-source migration from a legacy customer book-of-record to the enterprise master-data platform. I brought the dependency data (500K daily validation calls, 100K+ provider transactions), won approval to integrate critical consumers first, and then led the initiative across both platforms.

Enrollment that completes within an hour

I chose to redesign the enrollment flow on the new streaming infrastructure rather than patch the legacy batch. Nine months later members had digital access, claims and service within an hour of enrolling, and a seven-days-before-go-live performance crisis was fixed at ten times the throughput.

Punchout: a five-lever turnaround across more than 100 e-procurement systems

Punchout lets an enterprise buyer leave their procurement system, shop on Amazon Business and return the cart for approval. I owned Punchout across more than 100 e-procurement systems, but growth had slowed and too few connected users were buying. I wrote and ran the go-to-market turnaround across sales, services, product, marketing and partners. Amazon reported new Punchout adopters up 50% in 2023.

Integrated Quoting: we were optimising the wrong engine

Integrated Quoting lets enterprise buyers request quotes from Amazon Business inside their e-sourcing platform. After basket-level quoting launched in 2023, five months of funnel data showed why it was not converting: most quotes had product matches, but almost all requests failed the quote engine’s eligibility rules, and the prices we did return looked wrong because of a unit-of-measure default. I wrote the version 2 technical strategy: a marketplace offer with a price lock, a unit-of-measure fix and automation of the manual steps.

Hosted Catalog: designing the change, not just the publication

Enterprise buyers wanted Amazon Business’s planned-spend catalog inside their own e-procurement systems next to other suppliers. Keeping a published catalog current took manual refreshes. I took the product from framing through eight requirement versions and a pilot, with change management designed in from the start.

Moving pharmacy point-of-sale off the back-room server: a business-model conversion

At McKesson I led the migration of the retail point-of-sale and pharmacy management systems that process in-store payment and prescription transactions. $25M budget, a seven-person team.

Punch-in: partner readiness is not adoption

Punch-in lets a buyer start on Amazon Business and send the cart into their procurement system as a requisition. Partners went green. Customer adoption came in well below goal. This is the miss, what I changed, and why it reshaped how I set goals.