Clinical assessment for
insurance pre-authorization
SIT-I is our insurance integrity platform for health insurers. It works as a control layer that comes into play before any payment is made. When a pre-authorization request arrives, coverage, clinical documents, file history and SUT compliance (Türkiye's reimbursement rules) reach the specialist on a single screen, ready for review. The system runs routine checks automatically; specialists spend their time on critical decisions and always make the final decision themselves.
The role of SIT-I
in insurance processes
SIT-I works alongside the specialist across five steps of the health insurance process, from pre-authorization to medical advisory. At each step it prepares the information the specialist needs and leaves the assessment and the decision to them.
- 01
Pre-authorization
When a pre-authorization request arrives, the policy's coverage and the clinical documents attached to the request open on the same screen, so the specialist can assess them side by side.
- 02
Policy issuance
Coverage, waiting period and exclusion details are read from the policy document and moved into the relevant fields in the system.
- 03
Claims
While a claim is reviewed, the insured's previous pre-authorizations are visible as well, and the amounts are compared against one another.
- 04
Reimbursement
Before payment, the system checks how much of the policy limit has been used and whether the claim complies with SUT.
- 05
Medical advisory
The specialist can put a clinical question about the file to the system through Doctor Assistant and Council Mode.
Files reach the specialist ready for review
By the time a file reaches the insurance specialist, the information needed for the assessment is already gathered on one page. Because routine checks are done in advance, the specialist can spend their time on complex cases.
Policy analysis
Coverage, waiting periods and exclusions are extracted from the policy and shown right next to the file.
Limit tracking
While assessing the application, the specialist sees how much of the insured's limit has been used and how much remains.
File history
The insured's previous pre-authorizations and their amounts appear in the same view as the new application.
SUT compliance
The application is checked against SUT rules before the file reaches the specialist, and any non-compliant item is flagged.
Clinical context
Clinical observations sit on the same page as policy and amount details, so the specialist decides with the clinical picture in view.
Consulting the system
Through Doctor Assistant and Council Mode, the specialist can put questions about the file directly to the system.
AI helps the specialist
in daily operations
In SIT-I, AI does not take over the decision. It makes it easier for the specialist to ask questions, read documents and see which rule was applied.
Questions are asked in everyday language
The specialist can question the system the way they would consult a colleague: “What should I not miss in this patient?”
Document and inference shown separately
Every answer shows clearly and separately what the document says and what the AI infers, so the specialist always knows where a piece of information comes from.
Decision rules run in the rule engine
The rules a decision rests on are applied by the rule engine. AI only explains the resulting outcome and its rationale.
Pharmacy claims
In pharmacy claims, 6 separate rules cover the 6 rejection types, and each rejection type is checked by its own rule.
Verified data
The data AI works with is masked in line with the requirements of KVKK (Türkiye's personal data protection law) and stays inside the institution.
MINA's model evaluation
Distinguishing fraud, reimbursement compliance and clinical trap detection are three areas directly relevant to insurance processes.
You can find the task-level evaluation results for MINA v3.7 on the MINA v3.7 page.
A shared basis for assessment
at the hospital and the insurer
Difficulties in health pre-authorization mostly arise in the workflow between the hospital and the insurer. Opinion AI reduces these difficulties with a shared evaluation protocol that both sides use. The hospital prepares the application with TIS in its own system, and the insurer evaluates the same application with SIT-I in its own system. These are two separate deployments: each side keeps its data in-house, and nothing is transferred to Opinion AI. Both deployments use the same clinical AI model, MINA, and rest on the same regulations and rules. Documents are read with the same schema, clinical summaries follow the same structure and the AI score is calculated against the same criteria.
Hospital side
At the hospital, TIS runs with the AI agent MINA and the audit simulator MINA Cockpit. While the operator prepares the pre-authorization, MINA flags missing documents in advance, so the application reaches “ready for insurer” status before the hospital submits it.
Insurer side
At the insurer, SIT-I runs with the AI agent MINA and the insurance cockpit MINA Cockpit. The specialist evaluates the application and approves it, requests missing documents or rejects it. The rationale for the decision is reported automatically, and the insurer passes this report to the hospital with its response.
Documents are not re-entered
SIT-I reads the documents sent by the hospital directly on the insurer's side, so the same information does not have to be entered by hand a second time. Phone, fax and email traffic decreases, and so does the risk of lost documents.
Both sides use the same criteria
TIS and SIT-I score clinical fit, SUT compliance and file completeness against the same criteria. The hospital therefore sees a likely rejection reason before submitting, and the insurer finds missing documents already flagged in the file.
The pre-assessment is ready
When a file reaches the insurer's system, MINA prepares the pre-assessment. The specialist starts from a ready clinical synthesis, and gaps are flagged before the file reaches them.
A shorter appeal process
When a file is rejected, the insurer sends the rationale to the hospital together with its reason code. If a single document is missing, the hospital completes only that document and the file is re-evaluated.
What changes in
insurance operations
Specialist time goes to critical decisions
Before the file reaches the specialist, MINA reads the documents, checks the ICD-10 codes and reviews SUT compliance. With routine checks automated, the specialist can focus on complex cases.
Decisions rest on shared context
MINA's rationale and its comparison with similar past cases put the same context in front of every specialist. This supports consistent decisions across specialists on the same type of case.
High costs become visible early
MINA Cockpit provides an actuarial view of the portfolio, showing how many cases are active, how many of them are high-value and how many are at risk of rejection. Unexpectedly high accruals come to light before the month-end report.
Fraud risk is screened automatically
MINA's “Fraud Risk Analysis” query systematically screens for recurring diagnosis patterns, pairings between hospitals and physicians, and atypical relationships between procedures and policies. Findings go to the specialist for review.
Policy coverage appears on one screen
The insured's policy limit, remaining amount, waiting period, contracted-provider status and past approvals appear on the same screen. Answering “Is this procedure covered?” does not require a phone call.
The hospital and the insurer use the same schema
SIT-I uses the same schema and document types as TIS on the hospital side. As a result, there is less correspondence caused by a mismatch between the documents requested and the documents sent.
MINA and MINA Cockpit
inside SIT-I
SIT-I places MINA's clinical AI directly into the insurance specialist's workflow. MINA supports decisions on individual files; at the portfolio level, MINA Cockpit provides an actuarial view and audit simulation.
File-level decision support
MINA works alongside the insurance specialist. It reads documents, records vital signs and lab results, checks SUT compliance and evaluates the status of the application. It surfaces similar cases, explains the rule engine's outcome with its rationale and source, requests additional information when needed and cites clinical guidelines. The final decision always stays with the specialist.
Portfolio-level audit
MINA Cockpit is designed for the insurance manager and looks at the portfolio as a whole rather than at individual files. Ahead of official audits and actuarial reporting, it estimates how many cases are active, how many of them are high-value and how many are at risk of rejection. When a critical indicator crosses its threshold, the manager receives an alert.
The organizations
SIT-I is designed for
Health insurance companies
SIT-I suits insurers that manage individual and group health policies and handle high pre-authorization volumes every day. In these companies it helps protect specialist time, improve decision consistency and scale fraud screening capacity.
Complementary health insurance (TSS)
Complementary health insurers provide coverage on top of SGK (the Social Security Institution) and work at high volume on thin margins. In this segment, scaling operations depends on routine checks being automated.
TPAs and self-funded organizations
Organizations that run their own health fund for employees, and the third-party administrators (TPAs) that manage such funds, can also use SIT-I. In these setups it helps make health spending both predictable and auditable.
The final decision stays with the specialist
SIT-I leaves the decision to the specialist and supports it. Pre-authorization, claims and reimbursement decisions are made by the insurer's medical specialists, and AI shows the rationale and source behind them.
Responsible AI principle
People always make the decision. AI shows its rationale and the source it relies on, and every output can be audited.
Physician Ethics and Advisory Board
The board's physicians from 15 specialties regularly review the model's clinical accuracy and ethical boundaries.
Data stays in-house
SIT-I runs in the insurer's own environment, and data is not transferred to Opinion AI. The institution itself is the data controller, and personal data is masked in line with KVKK.
Related guides and articles
Each institution runs
its own deployment
Opinion AI is neither a hospital system nor an insurance system, and it does not act as an intermediary between the two. TIS runs inside the hospital and SIT-I inside the insurer as separate deployments, each keeping its own data in-house. Because both sides assess the same case with the same clinical AI and the same rules, their decisions rest on shared criteria. To see SIT-I in your own operations, you can submit a POC request.
POC Request