Denis Stark
Co-founder & CEO
28 years in payments. Issuing and acquiring at Shift4, Network International, Worldline. Built and led card-ops, implementation, and product teams across three continents.
LinkedIn →AI co-pilot for payment operations
For fintechs in issuing, acquiring, and PSP. Pre-trained on Visa/Mastercard rules, scheme dispute mechanics, and the regulations that touch your stack. Calibrated to your runbooks in days — not quarters.
Payment operations don't scale with headcount
Card operations sit between compliance, ops, and IT — and no single team owns the fix. Volume grows; the expert who knows Reg E timelines, scheme dispute codes, and your own runbooks doesn't.
- One or two SMEs become single points of failure
- Vendor and scheme documentation is fragmented across PDFs, portals, and tribal knowledge
- Horizontal AI tools (ChatGPT, Copilot) don't understand chargeback reason codes, BSA typologies, or your specific BIN configuration
- Hiring a payments expert takes 6–9 months — if you can find one
A payments-aware assistant, embedded where your team already works
Bulka.AI plugs into Slack, Teams, and your ticketing system. It answers questions and drafts responses using your runbooks, scheme manuals, regional regulations, and ticket history — every answer cited to source.
Compliance Q&A
Instant cited answers on scheme rules, regional regulations (Reg E/Z, BSA, PSD2, local equivalents), your internal procedures, and your team's historical interpretations. No more "let me get back to you tomorrow."
Operations co-pilot
Reject resolution, ticket triage, and routing. The AI drafts the response; your analyst approves and sends.
Change-scope analysis
Visa, Mastercard, and regulator bulletins triaged against your card products and BIN configuration — what applies to you, what doesn't, what action is required.
Audit-ready trail
Every question, every answer, every source — searchable, exportable, formatted for scheme audits or regulator review.
Built for the people who actually do the work
Card Operations Analyst
Reject resolution and ticket triage. AI drafts the response to Visa/Mastercard rejects, citing the relevant scheme rule.
Compliance / Disputes Analyst
Reg E and scheme-rules Q&A. Dispute timelines, BSA typologies, chargeback reason codes — answered with citation, not opinion.
Compliance Officer
Change-scope analysis. Scheme releases mapped to your products and BIN ranges.
Internal Audit
Searchable, exportable audit trail of every AI-assisted decision.
Business Analyst
Requirement analysis and effort estimation against existing platform capabilities.
Fraud Analyst
Triage suspicious transactions against scheme rules and known typologies. Surface precedent cases from your own ticket history.
An example query
A worked example using only publicly documented Mastercard chargeback rules — what an analyst would see in Slack or Teams.
bulka.ai
live
Analyst
What should I do when a cardholder reports an unauthorized transaction under Mastercard reason code 4837 — No Cardholder Authorization?
bulka.ai
For Mastercard reason code 4837 (No Cardholder Authorization), here is what the issuer needs to process the chargeback:
Required documentation (one of)
- Cardholder letter, email, message, or completed Dispute Resolution Form stating the transaction was not authorized
- For corporate cards: a written complaint from a company or government-agency representative when the authorized cardholder is no longer employed there
Time frame
- Chargeback must be filed within 120 calendar days of the Central Site Business Date of the transaction
If the merchant submits a second presentment
- A pre-arbitration case may be filed if the original chargeback was valid, Identity Check (3-D Secure) was not used, and either the second-presentment documentation is not compelling or the cardholder reasserts fraud
- Pre-arbitration must be filed within 45 calendar days of the second presentment, and at least 30 calendar days before escalating to arbitration
Important: all supporting documentation must be legible and linked to the case filing. Any documentation required but not provided in the chargeback will not be considered if the dispute escalates to arbitration.
Source: Mastercard Chargeback Guide — Reason Code 4837
Three pilots, three segments
PSP in the US
Acquiring processor in EU
Issuing / Acquiring in ME
Hours are redirected to judgment work, exception handling, and higher-value tasks — same headcount, more done. Figures are from early pilots with small teams; outcomes scale with team size and use-case mix. We'll size yours during the discovery call.
We fit into your existing vendor stack
For fintechs, we contract directly — standard MSA, no surprises.
For banks and larger processors, new-vendor onboarding can take months: TPRM review, SIG questionnaires, MSA negotiation, security review. If you already have an MSA with a firm that can prime AI engagements, we deliver as their subcontractor — no new MSA, no net-new TPRM scope.
- Cloud or on-premises deployment — your choice
- Pilots operate on documentation and anonymized samples — no production cardholder data in scope
- No data-science team required on your side
- Outcome-anchored fees — we carry the build risk; you pay against measured productivity gains
60 days. One team. One use case. Measurable results.
A 60-day pilot with weekly progress readouts. Pricing for small teams starts at $5K/month. We need 2–3 SMEs for ~4 hours of calibration, one operational team as the test group, sandbox or read-only access to relevant systems, and your scheme manuals and runbooks.
What you have at day 60
Working software
A live AI co-pilot deployed in your environment, in daily use by your operational team. Integrated into Slack/Teams and your ticketing system, with a full audit log of every query, answer, and source citation.
Measured results
A verified productivity delta against your own baseline — resolution time per ticket, escalation-to-SME rate, hours freed per week, end-user accuracy.
Decision-grade documentation
A final ROI report with verified numbers, a scoped scaling plan, and a security, integration, and TPRM posture summary — everything you need to decide what's next.
Operational findings
A current-state workflow map for the pilot team, a pain-point inventory ranked by impact, a use-case scoring matrix (effort vs. value), and a stakeholder readout with named owners.
Pilot outputs become inputs for what's next — no rework, no re-discovery. A decision, not a leap of faith.
Built by people who've been in your seat
Four co-founders, 70+ years combined in payments — across issuing, acquiring, PSP, and every major card management platform.
Konstantin Ryadov
Co-founder & CTO
Leads core AI development. 19 years in payments, with a focus on agentic AI and cybersecurity — building AI systems that hold up under regulated workloads.
LinkedIn →
Ivan Goncharov
Co-founder & Chief Platform Officer
Leads platform engineering, DevOps, and integrations — including the applied-AI components that connect Bulka.AI to client systems. 12 years at OpenWay and BPC building and operating the platforms your teams use today. 3 years at IBM.
LinkedIn →
Aleksandr Sovetov
Co-founder & Chief Delivery Officer
Issuing/acquiring implementations and platform migrations. 15 years in payments at Network International and Nexi. 30+ CMS implementations and 40+ portfolio migrations across the team.
LinkedIn →We know payments. We know your platform.
- 30+ CMS implementations for banks and payment processors
- 40+ portfolio migrations across teams
- Visa, Mastercard, and UnionPay certifications — issuing and acquiring
- Deep working knowledge of scheme rules and dispute mechanics
- Hands-on across the major card management platforms — OpenWay (Way4 — our deepest specialty), BPC SmartVista, VisionPlus, Base24
- AI/ML deployment experience in regulated environments
Let's run a pilot
60 days. One team. One use case. Measurable results.