Facets work for health plans that can't afford a bad claims weekend.
We help health plans run TriZetto Facets, the core administration platform from Cognizant. That means integrations, upgrade testing, data extracts and access controls around the claims, enrollment and provider data it holds.
Workflow Agentic Processing GA for Facets and QNXT (28 Sep 2026)
MCP tools
Library of more than 100 MCP tools for claims operations
release watch
Workflow Agentic Processing for Facets and QNXT28 Sep 2026
TriZetto AI Gateway06 Aug 2025
connects with
in short·15-second read
TriZetto Facets is Cognizant's core administration platform for health plans, and it holds claims, membership, billing and provider data.
We work alongside Facets teams. That means integrations, regression testing for upgrades, reporting extracts into Snowflake or Databricks, and access governance for anyone, or anything, touching member data. We also help plans size up Cognizant's newer agentic features, like Workflow Agentic Processing for pended claims.
you probably need this if
A Facets upgrade is on the calendar and most of the regression suite is still manual.
There's no reporting copy, so analysts have been querying the production Facets database.
Service accounts with access to member data have no clear owner.
06 Aug 2025TriZetto AI Gateway Existing TriZetto clients got this with the August 2025 release. One access point for TriZetto software and AI agents, with SSO, role-based access and telemetry. source ↗
28 Sep 2026Workflow Agentic Processing for Facets and QNXT GA now. Agents work from a plan's standard procedures and clear routine pended claims that qualify. Denials and complex cases still go to a person. More than 100 MCP tools ship with it. source ↗
today · Oct 22, 2026
the story · trizetto facets estate
From the first call to a setup your own team can run.
Here's what usually prompts a call about your TriZetto Facets estate, what we do about it and in which order, and where you end up.
01 · where it hurts02 · what we change03 · how it runs04 · where you end up
incident board · trizetto facets estate3 open
openA Facets upgrade is on the calendar and most of the regression suite is still manual.
recurringThere's no reporting copy, so analysts have been querying the production Facets database.
blockedService accounts with access to member data have no clear owner.
exposure
Upgrade and regression testingIntegrations around FacetsData extracts and analyticsAccess governance for member dataAgentic claims evaluation
startrehearsehandover
Environment and interface map
Regression suite from your…
Reporting copy
Access controls around…
Narrow agentic pilot
we are here
ready to hand over
Upgrades go out with a regression result the business has signed.
Compliance agreed the review rules up front and there's a measured baseline.
Regression suite built from your own claims, enrollment and billing scenarios
Interface inventory with owner, format and failure handling per feed
chapter 01 · where it hurts
Why teams reach out
A Facets upgrade is on the calendar and most of the regression suite is still manual.
There's no reporting copy, so analysts have been querying the production Facets database.
Service accounts with access to member data have no clear owner.
chapter 02 · what we change
How we solve it
Upgrade and regression testing
Integrations around Facets
Data extracts and analytics
Access governance for member data
Agentic claims evaluation
chapter 03 · how it runs
Our approach
Environment and interface map
Regression suite from your scenarios
Reporting copy
Access controls around member data
Narrow agentic pilot
chapter 04 · where you end up
What changes
Upgrades go out with a regression result the business has signed.
Compliance agreed the review rules up front and there's a measured baseline.
Regression suite built from your own claims, enrollment and billing scenarios
Interface inventory with owner, format and failure handling per feed
what we do
Our TriZetto Facets work
5 kinds of TriZetto Facets work, each run by engineers who have done it on live tenants.
Test plans built around your own claims, enrollment and billing scenarios, run in lower environments before each Facets upgrade or patch is promoted.
Interfaces between Facets and your CRM, care management, payment and broker systems, including EDI file flows and the APIs Cognizant exposes to licensed clients.
Facets data moved into Snowflake, Databricks or Microsoft Fabric on a schedule, modeled so finance, actuarial and operations teams can report from one place.
Role design, access reviews and service account controls for Facets and its databases, connected to SailPoint or Entra ID so access follows the employee.
A structured pilot plan for Cognizant's AI features on Facets, covering which pended-claim queues to start with, how human review works and what to measure.
capability 01 / 05in production
Upgrade and regression testing
Test plans built around your own claims, enrollment and billing scenarios, run in lower environments before each Facets upgrade or patch is promoted.
Interfaces between Facets and your CRM, care management, payment and broker systems, including EDI file flows and the APIs Cognizant exposes to licensed clients.
Facets data moved into Snowflake, Databricks or Microsoft Fabric on a schedule, modeled so finance, actuarial and operations teams can report from one place.
Role design, access reviews and service account controls for Facets and its databases, connected to SailPoint or Entra ID so access follows the employee.
A structured pilot plan for Cognizant's AI features on Facets, covering which pended-claim queues to start with, how human review works and what to measure.
We pieced these together from TriZetto Facets situations that come up a lot. They're patterns. You won't find a specific client in them.
case file 01
Facets upgrades tested with screenshots and a spreadsheet
composite · not a client story
Beforethe situation
Each Facets upgrade is tested by claims examiners working through a spreadsheet and pasting screenshots. Coverage of pricing, COB and accumulators is thin, and the last upgrade surfaced claim errors in production weeks later.
What we dothe work
We turn the spreadsheet and recent production claims into a regression suite organized by business process, with expected results for pricing and accumulators. It runs in the lower environment on every patch and upgrade. Failures go to named owners with the claim, the expected result and the actual one.
Afterwhere you end up
Upgrades go out with a regression result the business has signed. If pricing or accumulators break, you find out in test rather than from member complaints.
case file 02
Pended claims piling up and an agentic pilot on the table
composite · not a client story
Beforethe situation
Pended-claim queues grow every month-end, and leadership wants to try Cognizant's Workflow Agentic Processing. Compliance is uneasy about AI near adjudication. And nobody has a baseline for how examiners clear each queue today.
What we dothe work
We measure volume and handling steps per pend code, then pick two queues with written procedures and outcomes that are easy to check. Every denial and exception in the pilot gets human review, and agent access is reviewed like any other service identity.
Afterwhere you end up
Compliance agreed the review rules up front and there's a measured baseline. Whatever happens, it happened on your own queues.
We don't start a step until the last one has produced something you can check.
step 01
Environment and interface map
We write up your Facets environments, the batch schedule, EDI flows, APIs and database consumers. After that, everyone can see what depends on what.
step 02
Regression suite from your scenarios
Test cases come from your own claims, enrollment and billing scenarios, and we run them in lower environments before each upgrade or patch gets promoted.
step 03
Reporting copy
Facets data reaches Snowflake, Databricks or Fabric through extracts, replication or change data capture, depending on your hosting model and version.
step 04
Access controls around member data
Roles, service accounts and database access are reviewed and connected to SailPoint or Entra ID, so access ends when the job does.
step 05
Narrow agentic pilot
If you try Workflow Agentic Processing, start with one pended-claim queue. Keep human review on the exceptions, and measure what the agent does against your own examiners.
ai on trizetto facets
Where AI fits on TriZetto Facets, and where it needs guardrails.
Cognizant ships two AI pieces for Facets today. Workflow Agentic Processing, generally available since 28 September 2026, clears routine pended claims, and the TriZetto AI Gateway connects agents to TriZetto software. For an engineering team the work is picking the queues and scoping what agents can reach. Then you measure results against your examiners.
AI agentasks for data or an action
Scoped identityonly the access it needs
MCP serverthe one door into the platform
Logged, approveda person signs off on writes
TriZetto Facetsdoes the work
With Workflow Agentic Processing, Cognizant publishes a library of more than 100 MCP tools. The TriZetto AI Gateway adds SSO, role-based access and telemetry on top. We'd give each agent its own service identity and scoped roles and log its calls, and a human has to approve before a claim is denied or changed.
Workflow Agentic Processing for pended claims
Agents follow a plan's own standard procedures to clear eligible routine pended claims, and route denials and complex cases to people. It needs written procedures per pend code and agreed review rules. We measure each queue first, then pilot one or two queues with examiner review on every exception.
TriZetto AI Gateway
Shipped in the August 2025 TriZetto release, this is a single access point for AI agents, covering SSO, role-based access and telemetry. Our work is deciding which roles each agent gets, connecting sign-in to your identity provider, and getting the telemetry into your SIEM.
The MCP tool library for claims
Workflow Agentic Processing ships with a library of more than 100 MCP tools. We review every tool an agent can call and separate the ones that read claim data from the ones that change it. The second group stays behind human approval.
AI work around Facets
Outside Cognizant's features, AI helps with the work around Facets. We use it to draft regression cases from production claims, and to summarize batch and EDI error logs. It also maps Facets data to the FHIR resources behind the CMS-0057-F APIs. Every output gets reviewed by a person.
How claims, member data and access move around Facets
Enrollment and claims come in as EDI files and API calls, and Facets adjudicates them. Cognizant's agents can handle routine pended claims, and examiners get the exceptions. Analytics runs off a reporting copy. SailPoint governs access.
These are working documents, written against your configuration, so your team can run them without us.
tests01
Regression suite built from your own claims, enrollment and billing scenarios
yours to keep
inventory02
Interface inventory with owner, format and failure handling per feed
yours to keep
design03
Facets-to-warehouse data model for finance, actuarial and operations
yours to keep
plan04
Facets role design and access review schedule, vendor accounts included
yours to keep
plan05
Pilot plan for Workflow Agentic Processing with human review rules
yours to keep
talent and delivery · trizetto facets
Certified engineers, wherever you need them.
Some clients want a project team, some want a couple of engineers on theirs. Either works. Onsite, nearshore or offshore, with handovers across time zones that keep things moving.
Onsite
Engineers at your sites in the US
Nearshore
Overlapping hours with US teams
Offshore
Overnight progress on your backlog
All time zones
Coverage handed over across time zones
24-hour coverage · US Central time
OnsiteUS business hours
Nearshoreoverlapping US hours
Offshoreovernight progress
Combinedhandover, not hand-off
00:0006:0012:0018:0000:00
roles we staff
Facets configuration analyst (benefits, pricing)
Facets claims subject matter expert
Facets technical developer (extensions, batch)
Facets upgrade engineer
Facets EDI and integration engineer (834, 837)
Facets QA analyst
due diligence · trizetto facets
Questions to ask any partner, including us.
Put these to anyone bidding on your TriZetto Facets work. The ones who've done it answer differently from the ones who've read about it.
Q1
How will you build our regression suite, and whose claims will it use?
A vendor's generic scripts don't know your benefit configuration or your pricing. Upgrade defects show up when your own claims run through your accumulators, so the suite should start there.
Q2
Which of our pend codes suit agentic processing, and why those?
Cognizant aims Workflow Agentic Processing at routine pended claims, and denials and complex cases go to people. You can only measure a pilot on queues that have written procedures and outcomes someone can check.
Q3
How will vendor and offshore Facets accounts be reviewed and removed?
BPO staff and implementation contractors rarely appear in HR, so their Facets accounts never get a leaver event. Each needs an owner, an end date and a line in the access review.
Q4
How will Facets data get into our warehouse, and what does Cognizant actually support for the version we're on?
Hosting model and version decide which of extracts, replication or change data capture you can use. Guess wrong and it's stale data, or a setup Cognizant won't support.
Q5
Before 1 January 2027, how will you test the Facets feeds behind the CMS-0057-F APIs?
The Provider Access, Payer-to-Payer and Prior Authorization APIs draw on Facets claims and enrollment data. Feed defects found in January turn into compliance findings.
01Do you have access to Facets release notes and documentation?
Release notes and product documentation for Facets go to licensed clients. So we work from your copy, under your agreements, and in public we stick to what Cognizant has published.
02Can you help us decide whether to use Workflow Agentic Processing?
Yes, but narrow. We look at pended-claim volumes, the procedures examiners follow and compliance requirements, then propose a pilot with a human reviewing every exception. Cognizant says early-adopter results come later, so we measure yours instead of waiting.
03How do you get Facets data into a modern data platform?
It depends on your hosting model and what Cognizant supports for your version. Usually it's scheduled extracts, database replication into a reporting copy, or change data capture into Snowflake, Databricks or Fabric. We'd pick one after looking at your environment, not before.
04Do you also work on QNXT?
Facets is our focus. Much of the integration, testing and data work carries over, though, so tell us what you run and we'll say plainly where we can help and where we can't.
05Can you help with the CMS-0057-F APIs that draw on Facets data?
Yes. The Provider Access, Payer-to-Payer and Prior Authorization APIs have to be in place for impacted payers by 1 January 2027. We map the Facets data each one needs, then build and test the feeds and put identity in front of the API layer.
services
Related services
Projects on TriZetto Facets usually end up touching one of these, so they're worth a look.
