Enterprise
Transformation and delivery across 2 major banks, 1 oil company and 3 global IT integrators
Founder and Managing Partner at Amalgama
Kirill Kiryushin makes AI useful
Twenty years across consulting, enterprise technology and business ownership shape every practical AI decision.
14+Years in software and AI
50+Clients across industries
$35M+Value delivered
98%Client retention
AI needs experience before it needs another model
Models produce plausible output. Businesses need something stricter. The result must fit existing tools, permissions, handoffs, review rules, budgets, and the way people actually work.
Delivery experience changes the order of the project. Start with the business problem. Map the data, people, approvals, failure modes, and economics. Then decide where AI belongs.
A demo can ignore the awkward cases. A working system cannot. Those cases decide whether AI saves time, creates risk, or becomes another tool nobody trusts.
The work behind the point of view
Four parts of the background matter when AI meets a live organisation.
Consulting
Multiple projects with McKinsey, BCG, PwC, EY, Accenture and other global consulting teams
Operator
3 businesses built and 2 exits, with direct responsibility for the commercial outcome
Education
Master's degree in Management. Strategic is the leading CliftonStrengths theme
37.8M+assessments completed worldwide
Strategic, tested at global scale
CliftonStrengths is Gallup's 30-minute development assessment. It measures 34 research-validated talent themes built through decades of research. More than 37.8 million people have completed it worldwide.
Kirill's leading theme is Strategic. Gallup describes it as spotting relevant patterns, creating alternative routes and choosing a practical way forward when the situation is complex. It does not grade intelligence or seniority. It identifies recurring ways of thinking that can be developed into strengths.
What experience changes in an AI project
Choose the work
Find a workflow with enough volume, cost, delay, or risk to justify intervention.
Map the constraints
Account for systems, permissions, data quality, handoffs, exceptions, and the people who own them.
Keep control
Put review, confidence thresholds, logs, and escalation where the business consequence requires them.
Prove the result
Measure time, cost, accuracy, adoption, and business impact after the system meets real users.
Selected training and certifications
AI platforms, cloud, enterprise technology, and project delivery.
Read the working argument
AI in Business Is Becoming a Workflow Problem
Why production value depends on exceptions, permissions, data, and ownership after the demo.
Read articleAI Consulting Is Not Traditional Consulting with Faster Slides
A practical comparison of advice, implementation, evidence, and responsibility after launch.
Read articleEnterprise AI Implementation from Pilot to Scale
How to move a useful pilot into production without losing controls, adoption, or economics.
Read articleResearch and review standards are documented in the Editorial Policy.
Start with one workflow worth fixing.
A focused assessment identifies the strongest use case, the constraints, and the practical next step.