Runtime and Sumsub
Runtime vs Sumsub: Sumsub verifies users and businesses with KYC, KYB, AML screening, and case management. Runtime agents work the exceptions and cross-team investigations around those checks.
TL;DR
Sumsub is a full verification and compliance platform: KYC, KYB, AML screening, transaction monitoring, and case management with its Summy AI Copilot. Runtime agents do the work that spills outside Sumsub, across your ledger, processor, tickets, and other teams, and can call Sumsub as a tool. Most teams would run both.
| Feature | ||
|---|---|---|
| What it is | ||
| Core job | ||
| KYC, KYB, AML screening | ||
| Case management | ||
| Data it reasons over | ||
| Teams covered | ||
| Where it runs | ||
| Agent access | ||
| Pricing |
Runtime is not an alternative to Sumsub; most teams that use Sumsub would run both. Sumsub verifies and screens people and businesses with KYC, KYB, AML, and its own case management. Runtime is the AI agent harness for payment and fintech teams, and its agents work the alerts and exceptions those checks create, across systems Sumsub doesn't hold and teams beyond compliance.
What each does
Sumsub is a verification and compliance platform used by fintech, crypto, iGaming, trading, and marketplace companies. It covers KYC (ID documents, liveness, proof of address, non-document checks, reusable KYC), KYB with UBO verification, AML screening and ongoing monitoring, transaction monitoring, fraud prevention, and the Travel Rule for crypto. On top sits an operations layer: case management, a workflow builder, a rule engine, risk scoring, FIU reports, and the Summy AI Copilot, which summarizes cases and surfaces risk signals inside the platform. Sumsub also offers an in-house team that manually verifies complex corporate documents, and in June 2026 it released an MCP integration and open-source agent skills so outside AI agents can configure Sumsub and help with daily compliance work. Sumsub says it serves more than 4,000 clients.
Runtime is a harness for building and running many agents across the org. Anyone on risk, compliance, onboarding, payment ops, finance, or support builds agents from their SOPs. Agents work on their own isolated computers in your cloud and reach tools through APIs, databases, CLIs, MCP servers, and a browser for portals with no API. They start read-only, stop for approval before anything moves money, and every run is recorded end to end.
Put simply: Sumsub checks who someone is and whether they are risky. Runtime agents work out what to do about it across the business.
Where they overlap
Sumsub does more than verification, so the overlap is worth naming.
- Case management and AI review. Sumsub has its own case management, and Summy helps investigators summarize cases and spot risk signals. Sumsub also says AI agents analyze AML screening hits and filter false positives. If your reviews start and end in Sumsub, that may be enough.
- Agent access. Sumsub's MCP integration lets general-purpose AI agents review applicants, run analytics, and propose configuration changes, with configuration changes approved by a person. That is a door for agents, not a place to run them. Runtime is one place to run them.
- Workflows and rules. Sumsub's workflow builder and rule engine already route applicants. Runtime doesn't need to replace that.
The line is the edge of Sumsub's data. A compliance review on a payments company often needs the customer's transaction history, held funds at the processor, past support conversations, chargebacks, or notes from another team. Sumsub doesn't hold those. Runtime agents can reach them.
How they work together
A typical flow for a flagged customer or an ongoing-monitoring alert:
- Sumsub flags. An applicant fails a check, an AML monitoring hit fires, or a transaction monitoring rule trips. The case opens in Sumsub.
- A Runtime agent picks it up. Through Sumsub's API, a webhook, MCP, or a Slack alert, the agent pulls the applicant, the check results, and Summy's case summary.
- It investigates beyond Sumsub. The agent checks your ledger for the customer's activity, the processor for related payouts or disputes, your support tool for recent conversations, and earlier cases on the same person or business.
- It drafts the decision with evidence. Following your SOP, the agent writes what triggered the case, what it checked, what it found, and a recommendation: clear, request documents, restrict, or escalate for a SAR decision.
- A person approves in Slack or Teams. Freezing an account, releasing funds, or replying to a sponsor bank always waits for a person.
- The agent writes back and closes the case. It updates the case in Sumsub, notes the ticket, and keeps one record of the whole run: trigger, every query and tool call, approvals, cost, and result, exportable for your sponsor bank or examiner.
The same pattern covers periodic re-reviews. The agent pulls fresh Sumsub results, compares them to what the business is actually doing in your systems, and prepares the review for sign-off.
Where Runtime enters the picture
Sumsub creates a queue: failed checks, AML monitoring hits, and transaction monitoring alerts that a person has to clear. Runtime is the layer that works that queue, and the work that spreads from it. Say an ongoing AML monitoring hit fires on a customer with an outgoing wire pending. Compliance reviews the hit, payment ops holds the wire, support fields the customer's question, and finance sees an unreconciled item. Each team usually investigates in its own tool. On Runtime it is one investigation with one record, and Sumsub's case is one input to it.
Runtime is the AI agent harness for payment and fintech teams. It doesn't replace Sumsub's checks or case management; it gives compliance, payment ops, finance, and support agents that use Sumsub as a tool next to everything else they touch.
- No single vendor to depend on. Agents run in your AWS, GCP, or Azure account or fully self-hosted, on your choice of sandbox provider with fallbacks, and on any model (Claude, GPT, Gemini, or open-weight models in your cloud for PII), with harness routing across Claude Code, Codex, and OpenCode.
- Guardrails you set. Agents start read-only, freezing an account or releasing funds waits for approval, access follows roles, credentials are masked, SSNs and card numbers are stripped, and every run is recorded.
- Built for teams. The agent clearing Sumsub alerts shares templates and one memory with the agents in support, payment ops, and finance, so each resolved case helps the next one.
- An engineer, not an account executive. A forward-deployed AI engineer who built payment infrastructure at Hulu's payments team at Disney, Finix, Modern Treasury, and BlackRock maps your compliance SOPs and builds the first agents with your team.
Create a banking ops agent that handles RFIs, hold-harmless requests, and returns of funds from our partner bank, verifies against our ledger, and drafts replies for my approval.
bank-ops-agent
Tools picked from your prompt
When you need both
- Your Sumsub alerts often need data from your ledger, processor, or support tools
- Compliance flags turn into work for payment ops, finance, or support
- You want one record per case, including what happened after the flag
- You want agents on other queues too, not only compliance
When Sumsub alone is enough
- Your reviews can be resolved with Sumsub's own data and case management
- Summy and Sumsub's AI screening keep your backlog under control
- You want verification and its case queue from one vendor and nothing more yet
- Other teams aren't ready for agents
Frequently asked questions
Is Runtime an alternative to Sumsub?
No. Sumsub verifies people and businesses: ID documents, liveness, AML screening, KYB, and transaction monitoring. Runtime does not do those checks. Runtime agents work the exceptions and investigations around them, and can use Sumsub as a tool.
Does Sumsub have AI agents?
Yes. Summy AI Copilot, launched across the Sumsub platform in February 2026, summarizes cases, answers product and compliance questions, and builds analytics, with decisions left to people. Sumsub also uses AI agents in KYB and AML screening, and in June 2026 released an MCP integration so outside agents like Claude or ChatGPT can work with Sumsub.
How much does Sumsub cost?
Sumsub publishes pricing. Basic is $1.35 per verification with a $149 monthly minimum. Compliance is $1.85 per verification with a $299 monthly minimum. KYB, transaction monitoring, and fraud prevention are on a custom plan through sales.
Can Runtime agents use Sumsub's MCP server?
Runtime agents can reach tools through APIs and MCP servers, so they can use Sumsub's API or MCP integration with credentials you scope. Anything that changes a decision or moves money waits for a person to approve.
Why use Runtime if Sumsub already has case management?
Sumsub's case management works on Sumsub's data. Many reviews also need your ledger, processor, support tickets, and other teams. Runtime agents work across those systems and keep one record of the whole investigation.
Where does Runtime fit next to Sumsub?
Runtime works the queue Sumsub creates. Sumsub verifies and screens people and businesses and opens cases when something is flagged. Runtime agents pick up those cases, investigate across your ledger, processor, and support tools, draft the decision for a person to approve, and handle the follow-up for payment ops, finance, and support.