PENDRAL
Request a demo

Proven where
it is hardest
to be wrong.

Built and proven inside a payer organization before it was ever sold to anyone else.

Request a demo See the platform
Vocalis · calls

Calls

3,247 today · 412 suppressed pre-dial
TimeContactCampaignDurOutcomeSentiment
10:41Member · claim statusInbound reception4:12Met0.82
10:39Provider · roster checkRoster verification2:48Met0.61
10:38Member · benefitsInbound reception6:03Met0.74
10:36Member · wellness visitAnnual wellness3:21Met0.68
10:34Member · post-dischargeDischarge follow-up5:47Transferred0.44
10:31Member · ID cardInbound reception1:58Met0.79
10:29Provider · credentialingRoster verification4:36Met0.57
10:27Member · wellness visitAnnual wellness2:12Callback0.66
10:24Member · prior authInbound reception7:19Met0.71
10:22Member · wellness visitAnnual wellness0:41Suppressed—
Inbound · member services+1 (555) 014-2288LIVE · 02:14
CallerHi, I’d like to check on my claim from last month.
AgentClaim 4821 for the visit on 12 July — it was approved Tuesday.
CallerGreat. Do I owe anything on it?
AgentYour share is $40, applied to your deductible. Nothing due today.
CallerCan someone call me back about my new card?
AgentBooked — Thursday at 2:00 PM with member services.
2017Founded as Evry Health
YC W18Y Combinator
0.5M+Documents processed
500K+Voice calls completed

Nobody else can
claim this heritage.

We built these tools to run a health plan, on live member traffic, under audit. Then we sold them. Every competitor’s compliance story starts with a roadmap. Ours starts with an incident log.

Four products.
One governance
model.

VocalisVoice agents

Enterprise voice AI that answers, calls, and knows when to hand off.

EvryAIAI workspace

One governed workspace for chat, agents, artifacts and tool access.

InlineCode review agent

An autonomous reviewer on every pull request — and the supply chain behind it.

LakefrontDeployment platform

Ship to the cloud accounts you already own, without the architecture PhD.

VocalisVoice agents

The phone line, staffed by an agent that reads the file first

Enterprise voice AI that answers, calls, and knows when to hand off.

Vocalis · campaigns

Campaigns

Outbound work, its objectives, and what it achieved.

New campaign
CampaignAgentProgressObjectiveStatus
Annual wellness outreach7,961 of 8,420 contactsWellness v495%71%Running
Post-discharge follow-up1,104 of 1,240 contactsDischarge v289%88%Running
Provider roster verification1,482 of 3,110 contactsRoster v748%64%Running
Prior-auth status callbacks620 of 2,400 contactsAuth v126%59%Paused
Open-enrollment reminder0 of 12,800 contactsEnrollment v10%—Scheduled
Care-gap closure — diabetic eye2,210 of 2,210 contactsCare gap v3100%76%Complete

412 numbers suppressed before dialing — federal registry, two state lists, and the plan’s own opt-outs.

<500ms

agent response

500K+

calls completed

AES-256

at rest, US-only

Conversation

  • Vocalis Realtime Model
  • Knowledge bases
  • Agent workflow builder
  • Simulator

Operations

  • Campaigns and contact lists
  • Call centers, IVR and agent groups
  • Human handoff and transfer
  • Help desks

Compliance

  • Multi-source DNC
  • TCPA support
  • Safety escalation
  • Evidence capture

Intelligence

  • Objective analysis
  • Sentiment and issue detection
  • Coaching pipeline
  • Reporting and export
Works withSalesforceSIP / carrier failoverLiveKitWebhooksREST API

EvryAIAI workspace

Every model your team needs, behind one set of rules

One governed workspace for chat, agents, artifacts and tool access.

EvryAI · workspace
Eligibility reconciliation — Aug 834 feed PHI · restricted

Reconcile the Bronze HSA roster against the 2 Aug 834 feed and tell me what’s pending.

Searched knowledge baseplan-operations / 834-feeds

Ask anything, or drop a file in here…

WebAuto ⌄
pending-eligibility-24.xlsxARTIFACT
member_idplanmissingsource
M-104412BRZ-HSAeff_date834:0802
M-104518BRZ-HSAeff_date834:0802
M-104603BRZ-HSAeff_date834:0802
M-104711BRZ-HSAeff_date834:0802
M-104809BRZ-HSAeff_date834:0802
M-104922BRZ-HSAeff_date834:0802
M-105044BRZ-HSAeff_date834:0802
M-105130BRZ-HSAeff_date834:0802
M-105207BRZ-HSAeff_date834:0802

24 rows · built in an isolated sandbox · never left the tenant

9

models

3

providers

1

permission model

Working surface

  • Multi-provider chat
  • Artifacts
  • Projects and knowledge bases
  • Skills, memories and workflows

Reach

  • MCP and Nango connections
  • Sandboxes
  • Outpost
  • Clients everywhere

Governance

  • Organisations and roles
  • Sensitivity controls
  • Audit and investigations
  • Limits, features and moderation
Works withMicrosoft 365Google DriveSlackGitHubMCP servers

InlineCode review agent

Review that earns its comments

An autonomous reviewer on every pull request — and the supply chain behind it.

github.com/acme/claims-api/pull/4182
src/middleware/rate-limit.ts+4−4
38 export async function rateLimit(req: Request) {
39 const bucket = windowFor(req.path)
40 if (!bucket) return next()
41
42 const key = `rl:${req.ip}:${bucket.name}`
43 const hits = await redis.incr(key)
44- if (hits === 1) redis.expire(key, bucket.window)
45+ if (hits <= 1) redis.expire(key, bucket.window)
46
47 if (hits > bucket.limit) return tooMany(bucket)
48 return next()
49 }
i

Inlinebot · confidence 0.91 · high severity

The TTL is only set on the first hit. If incr races between two workers, the key is created without an expiry and this IP is locked out permanently. Set the expiry unconditionally, or use SET NX EX.

resolved by r.mehta in 0c41f8a

$0.50–1.00

per review

10×

cheaper cached prompt block

0

comments is a valid review

The review pipeline

  • Path filters
  • Incremental diffs
  • Project detection
  • Agentic deep review
  • Validation gate

Security

  • Continuous exposure monitoring
  • Agent-verified findings
  • Threat feed and alerts

Control

  • Per-repo enablement
  • Cost accounting
  • Learnings and skills
  • Outcome tracking
Works withGitHub AppAzure SQLMicrosoft Entra IDCI checks

LakefrontDeployment platform

Your cloud. Your account. One surface.

Ship to the cloud accounts you already own, without the architecture PhD.

Lakefront · your Azure and AWS accounts

Services

Your applications and the services each one runs.

Deploy
member-portal Azure · East US
webWebmembers.acmehealth.com3 replicaslive
queue-workerWorker2 replicaslive
notificationsWebSocket2 replicaslive
nightly-syncScheduled0 2 * * *idle
claims-api Azure · Central US
apiWebclaims.acmehealth.comrev 0007building
ingestWorker4 replicaslive
adjudicationWorker6 replicaslive

BYOC

runs in your account

0

long-lived cloud secrets

Multi-region

failover within a cloud

Deploy

  • Repo to running service
  • Grouped by application
  • Live everything
  • Instant rollback

Govern

  • Org, team, project, environment
  • Brokered down to each cloud
  • No standing credentials
  • Agent scope attenuation

Operate

  • Databases and storage
  • Domains, previews and status pages
  • Costs
  • Build-and-verify AI
Works withAzureAWSGitHubCloudflareAnthropic / OpenAI

The part
everyone else
adds later.

Built under a BAA, not retrofitted for one

Every provider in the data path is covered. One point of compliance, not a subcontractor chain.

US infrastructure, AES-256 at rest

Encrypted Azure SQL in US regions. Secrets in an HSM-backed vault.

Zero training on customer data

No voice, transcript or document ever trains a model.

Compliance enforced at runtime

DNC before the dial, opt-out during the call, classification on every request.

Least privilege, brokered not stored

No long-lived cloud key to leak. Agents scoped below the human behind them.

Bring it somewhere it has to hold up.

Thirty minutes, a scoped pilot on your own data, and the architecture review our auditors get.

Request a demo Talk to sales