Incident management procedure
For banks, the scope of work is defined with information security and compliance. Security, backups, logs and notifications are built into the procedure with owners and deadlines.
Support for systems and AI models: SLA-based response, 24/7 monitoring, backups and updates. Subscription scope, on-call coverage and development are agreed after an audit.
The bank is accountable for incidents
Kazakhstan. From July 2026, the bank’s CEO is personally responsible for information security, and any system outage longer than an hour must be reported to the regulator — an on-call shift and a recovery procedure are now a requirement, not an internal choice. Uzbekistan. By law, a bank compensates customers for cyber fraud losses if it failed to meet security requirements, and is added to a public list; banks are already restricting access for millions of devices every six months.
Post-launch support: SLA-based support, monitoring, backups and updates
Taking systems into support, fixes, infrastructure and model maintenance. Hosting, access and each party’s responsibilities are defined before the start; a monthly report covers work done and incidents.
For banks, the scope of work is defined with information security and compliance. Security, backups, logs and notifications are built into the procedure with owners and deadlines.
Team composition and on-call schedule depend on load, system criticality and SLA. A subscription lets you bring in an external support team.
Response tiers, target availability and fix times are set separately. Monthly report: SLA, incidents, releases and plan.
Single point of contact; response within 1, 4 or 24 hours depending on severity. 24/7 on-call for critical incidents as per the contract.
Fixes within agreed timelines; the volume of minor enhancements in the subscription and the task queue are agreed with the client.
Features and integrations in two-week sprints, based on an agreed plan and estimate. Delivery
Test, pre-production and production environments. Deployment checks and a rollback plan tailored to the system and infrastructure.
Backups of databases and configurations, two sites and recovery tests. Frequency, acceptable data loss and recovery time are set in the contract.
Availability, errors, load and queues; alerts to the on-call engineer and the client, incident logging and review.
Updates, access control and logs on an agreed schedule. Regulator notifications follow the established procedure, with designated owners on the client side.
Monitoring of quality, drift, costs and logs. Updates or fine-tuning when needed and only with approved data. AI Control Gate
SLA, incident root causes, availability, releases and plan; monthly review with the client.
Response: critical — 1 hour, high — 4, normal — 24. Availability and resolution are agreed separately. Pricing depends on the systems and services covered.
Client infrastructure or Alibaba Cloud under contract; hosting and access eligibility is verified before any data is migrated.
Quality and cost control, updates and, if needed, fine-tuning — a separate subscription scope. AI products
Development in sprints or engineers joining the client’s team. Team augmentation
“Launch is the middle of a project, not the end. A system lives for years: the regulator changes requirements, load grows, models drift. That is why we stay on support with an SLA and a monthly report — the bank sees what works and what we have improved.”
Oleg SolabayevCTO
Yes, after an audit, documentation and a check of monitoring and backups. The scope of work and the SLA start date are set in the contract.
Response by severity: 1, 4 or 24 hours; critical incidents are covered by 24/7 on-call duty. Response time, how it is counted and exclusions are defined in the SLA; resolution time is set separately.
On the client’s premises or in Alibaba Cloud under contract. Hosting, transfer and access eligibility is verified against the data and applicable requirements.
Support, backups, monitoring, updates and reporting within the agreed scope. New features and integrations are estimated separately; the threshold for minor changes is set in the contract.
Quality, drift and cost control; logs according to agreed requirements. Updates and fine-tuning depend on the model, data and task; the scope is defined separately.
The ticket queue and statuses are available to the client. Monthly: a report and review of SLA, incidents, availability and releases.
Your request goes to our business development team: we will get in touch, work out what needs to go live at your bank and propose a solution.
Thank you, your request has been sent
A business development manager will contact you using the details you provided.
