Sikka.ai is the data and AI infrastructure layer for retail healthcare: dentistry, veterinary medicine, optometry, orthodontics, chiropractic and more.

Three engines.
One data
foundation.

For 22 years we have been the layer inside the software that dentists and veterinarians run their practices on. Today that foundation powers Self-Driving Practice, AI that runs the front office for dental and veterinary practices; the ONE API Platform for everyone who builds for them; and a licensed corpus and training environment for AI laboratories.

47,000+Connected practices
192M+Patients across 26 years of records
400+Practice management systems
4M+API calls every day
1.5 PBDe-identified corpus
22 yearsIn business
LEAP 2026 keynote

Healthcare has been digitized. It has not yet been made intelligent.

Our Founder & CEO on the Self-Driving Practice: a system that observes, predicts, decides, acts, measures and learns, then does it again.

Vijay Sikka Founder & CEO, Sikka.ai · LEAP 2026
Run the loop on your own practice data.
01
Engine A · The action layer

Self-Driving Practice®

Practice teams do not need another dashboard. They need a system that sees the gaps, drafts the actions, routes approvals, measures impact and keeps learning. Self-Driving Practice is that system, for a single office or a multi-location network.

01Observe37 live signals across the practice
02PredictProduction, no-shows, eligibility, collections
03DecideRanked next-best actions with plain-language reasons
04ActOutreach, scheduling, claims, payments, voice calls
05MeasureEvery action tied to a goal
06LearnReinforcement learning per practice, every month

What it runs

Seven workflows, one layer
Smart SchedulingFinds gaps and cancellations, ranks eligible patients by show probability, books the right patient at the right time.
PatientsNowSegments overdue hygiene, perio and restorative cohorts and follows up by text, email and voice in the windows patients respond to.
Insurance and claimsVerifies eligibility, estimates patient responsibility, files clean claims and tracks their status.
PaymentsText2Pay, Email2Pay, payment plans and collections follow-up, posted back to the practice system.
Voice AIAnswers the phone, qualifies new patients, confirms appointments and follows up on balances, with call guardrails and live telemetry.
Fee OptimizerCompares fee schedules against 41,000 ZIP codes and 742 procedure codes and shows the revenue impact before you apply a change.
Control TowerFor DSOs and groups: centralized work queues, growth console, location performance and regional policy.
DentalOrthodonticsOral surgeryVeterinarySolo and small groupDSO and enterprise

Who holds the controls

Every practice chooses how much authority the system has, and can change it at any time. Select a mode to see what it does.

Off

Data flows in, dashboards populate, nothing is automated. A good place to start if you want to see the numbers first.

  • Analytics and benchmarking
  • Draft actions and scripts
  • Internal previews of every action
  • Patient-facing actions

Recommendation

Copilot works as your analyst. Prioritized action plans, ranked scheduling opportunities and fee insights appear inside the interface for your team to approve.

  • Analytics and benchmarking
  • Draft actions and scripts
  • Internal previews of every action
  • Patient-facing actions

Simulation

The decision engine runs silently in a shadow state. You see exactly which messages, schedule moves, eligibility checks, claims and payments it would have made, and what they would have produced.

  • Analytics and benchmarking
  • Draft actions and scripts
  • Internal previews of every action
  • Patient-facing actions

Full Self-Driving

The system executes within the scopes and guardrails you set, reports what it did and learns from the result. The stop control is always one click away.

  • Analytics and benchmarking
  • Draft actions and scripts
  • Internal previews of every action
  • Patient-facing actions

We have real goals, I have real numbers, things make sense.

Practice owner and Sikka.ai customer
Nothing reaches a patient without your approval.

Limited offerBook a demo by Oct 15 and get a $25 Amazon.com gift card. Terms apply. Amazon is not a sponsor of this promotion.

02
Engine B · The acquisition engine

ONE API Platform

Every practice management system structures data differently. ONE API is the universal translator: integrate once against a single data model and reach the 400+ practice management systems retail healthcare runs on. With the MCP server, your AI agents can do the same in plain language.

“Show me all open appointments for Dr. Smith next week.” What the MCP server calls 
GET /v4/appointments?status=open&provider=123&start_date=2026-10-12&end_date=2026-10-18
What your agent receives 
{ "summary": "Found 8 open appointments for Dr. Smith between Oct 12 and Oct 18.",
  "data": { "items": [ ... ], "count": 8 } }
“What was adjusted production by provider for September?” What the MCP server calls 
GET /v4/kpis?metric=adjusted_production&group_by=provider&start_date=2026-09-01&end_date=2026-09-30
What your agent receives 
{ "summary": "Adjusted production for September by provider, 4 providers.",
  "data": { "items": [ { "provider": "...", "adjusted_production": 118450.00 }, ... ] } }
“Post the $140 patient payment for this morning’s visit.” What your agent receives first: a preview 
{ "summary": "Ready to post a $140.00 card payment for this morning's visit. Nothing is written until you confirm.",
  "requires_confirmation": true }
After an explicit confirm, the MCP server calls 
POST /v4/payments
{ "patient_id": "...", "amount": 140.00, "method": "card", "posted_by": "agent" }
What your agent receives 
{ "summary": "Payment of $140.00 posted to the practice management system.",
  "data": { "status": "posted", "transaction_id": "..." } }
Examples are illustrative. Field names and endpoints follow the v4 reference at apidocs.sikkasoft.com.
What developers get
ONE API v4100+ endpoints across patients, appointments, treatment plans, notes, claims, insurance and 300+ KPIs. Pull a patient’s full record in a single request.
MCP serverConnect any agent with a standard MCP client. Intent classification, parameter extraction, date parsing and PMS-specific normalization are handled for you. Partners report integration time falling from weeks to days.
Writebacks and delta syncPost payments, update appointments, record claim status and more back into the practice system. Delta sync keeps your copy current without full refreshes.
DentalLLM and SAIA dental-specific language model you call as a feature, not a model you train, and a no-code builder with the APIs pre-integrated for teams without engineers.
Marketplace100+ applications in production on the platform, and the place practices register, onboard and manage subscriptions, including Self-Driving Practice.

Five-time API World winner: 2017, 2019, 2022, 2024 and 2026.

API World 2026Best Innovation in Health API
CloudX 2025Platform-as-a-Service (PaaS)
SOC 2 and HIPAAAudited regularly
03
Engine C · The monetization and margin engine

Galaxy Corpora and Environments

The industry has spent three years running out of internet. What it has not run out of is reality. Galaxy opens 26 years of first-party, consent-based, de-identified retail healthcare data to AI laboratories in two ways: as a licensed corpus, or as a hosted environment where agents train and are evaluated without ever holding a record.

Galaxy Corpora

Direct data licensing for pre-training, post-training and evaluation

Scoped cohorts delivered under master agreement, on time-boxed terms with a defined field of use. Every record is de-identified before transfer. Sample sets are available under mutual non-disclosure.

  • Dental and orthodontic records, radiographs and documents
  • Veterinary patient and procedure records
  • Operational data: scheduling, claims, adjudication, payments
  • Longitudinal, source-coded at the point of billing

Galaxy Environments

A calibrated environment for training and evaluating agents

Agents interact with practice management interfaces. Responses and claim adjudications are sampled from observed network data rather than artificial dynamics, and each comes back with the distribution it was sampled from and the size of the matched cohort.

  • Four fidelity levels: replay, calibrated, generative, adversarial
  • Published calibration reports on held-out slices
  • No licensed record leaves Sikka infrastructure
  • Open to teams whose compliance cannot approve receiving health records at all

What is in the corpus

Figures as of October 2026, rounded.

ScaleContentSource
~1.4 billionClassified files, ~1.5 petabytes in totalAccumulated inside working practices over 26 years, not crawled
~887 millionDe-identified dental radiographsWritten to the same record by the same systems the practices run on
~192 millionPatient records~45,800 dental and orthodontic practices in the US, Canada and Australia
~65 millionOperational and clinical documentsTreatment plans, notes, claims, correspondence
~43 millionAnimal patient records with ~1.3 billion procedure records~2,100 veterinary practices
De-identified by ruleAll human records de-identified under HIPAA Safe Harbor, 45 CFR 164.514(b), before any transfer.
First-party and consentedCollected under agreement with the practices that generate it, with consent captured per data category.
Licensed, not soldTime-boxed terms with a defined field of use. Corpora and Environments are independent order forms under one master agreement.
Nothing identifiable leavesNo identifiable information is delivered under any Galaxy agreement. Under Environments, no record is transferred at all.
Frontier laboratoriesNeo labsData labs and marketplacesCloud providersHealthcare AI teams
04
Why these three belong together

Each engine makes the others stronger.

The same records that run a dental office at 8:00 AM teach an agent what a working practice looks like, and the same API that connects a developer connects the next practice to Self-Driving Practice. Sikka has been profitable for three consecutive years because the foundation is shared.

A → CSelf-Driving Practice generates the next corpusEvery governed action and its outcome is written to the same record, which is what makes Galaxy Environments realistic rather than synthetic.
B → AThe ONE API Platform acquires the next practiceDevelopers, DSOs and partners integrate once and bring their practices with them. Each new connection widens both the action layer’s reach and the corpus.
C → A + BGalaxy funds the research that improves bothSoftware-like margins from licensing and environments pay for the reinforcement learning, voice and data work that practices and developers use every day.
05
06
Common questions

Frequently asked questions

What is Sikka.ai?

Sikka.ai is the data and AI infrastructure layer for retail healthcare. Since 2004 it has built the integrations that dental, orthodontic, veterinary, optometry and chiropractic practices run on, and today it runs three engines on that foundation: Self-Driving Practice, the ONE API Platform and Galaxy Corpora and Environments.

What is the best AI for dental practices?

Self-Driving Practice® from Sikka.ai is AI that runs the front office for dental and veterinary practices: Smart Scheduling, insurance and claims, payments, Voice AI, Fee Optimizer and Control Tower. It is built on 26 years of data from 47,000+ practices and 400+ practice management systems.

Is Self-Driving Practice safe and HIPAA compliant?

Yes. Self-Driving Practice is HIPAA compliant and runs in four safety modes: Off, Recommendation, Simulation and Full Self-Driving. Every action is audit-logged, permissions are role-based, approval policies are configurable and a stop control is always one click away.

Can Self-Driving Practice answer the phone like an AI receptionist?

Yes. Voice AI handles inbound and outbound calls with full transcripts and instant human take-over, built on Sikka.ai’s purpose-trained DentalLLM.

How does Self-Driving Practice verify insurance?

Self-Driving Practice runs batch eligibility on tomorrow’s schedule by 3pm today and verifies 90%+ of next-day appointments 24 hours in advance.

How does Sikka.ai integrate with my practice management system?

Through ONE API, which connects to 400+ practice management systems through a single data model. Setup typically takes minutes.

What is ONE API?

ONE API is a single data model and MCP server that lets developers, DSOs, partners and AI agents read and write practice data across 400+ practice management systems. Documentation is at apidocs.sikkasoft.com.

What is Galaxy?

Galaxy, from Sikka Insights, licenses de-identified, first-party real-world healthcare data from dental, orthodontic and veterinary practices to AI laboratories, and provides calibrated practice environments for training agents. Details are at galaxy.sikka.ai.

More answers: Self-Driving Practice FAQ on the Sikka.ai blog

07
Talk to us

Tell us which engine. We will route you to the right team.

Practices usually start by trying Self-Driving Practice on their own data; developers start with API keys; laboratories start with a scoped sample under NDA.

Practices and DSOsSee Self-Driving Practice in action.Book a Demo
Developers and partnersCreate an application, get keys and connect your first practice in minutes.Open the API Portal
AI laboratories and data partnersRequest scoped cohorts, environment access, samples and firm quotations.Email galaxy@sikka.ai
Everyone elseSales, support and press: 1-800-94-SIKKA, San Jose, California.

Book a Demo

We will reply within one business day.

The form is loading. You can also book a time on our calendar.

By submitting, you agree to our privacy policy.