Process & AI Automation
Process first, then AI
Automation only helps if the process is clear and the data is reliable. In consumer lending I start with how the work actually runs, agree with stakeholders what we will measure, and then build: rule changes, data pipelines, dashboards, and AI where it fits.
How I work
From use case to adopted solution
Find the opportunity
I look at lending processes end to end and find where the written policy, the data and what operations actually do stop matching. Those gaps are usually where the useful automation and AI cases are.
Simplify before automating
Before tooling, I specify the process change: what step is redundant, which control actually fires, which handover creates rework. Coordinating the Latvian Customer Service unit, my resource planning, report analysis, and process improvements helped it finish best of Luminor's three Baltic units on efficiency, quality, and sales.
Translate business to build
I turn stakeholder questions into documented requirements, assumptions, and success criteria, then work with technical teams on the solution. I documented source-to-KPI logic for a new consumer finance product so it could be monitored on the same portfolio views as the rest of the book; later risk and performance reporting reused that layer.
Define value and KPIs
I maintain portfolio KPIs for credit risk reviews and measure the impact of lending rule changes before and after they go live. Automation gets the same treatment: agree the measure up front, then show the change moved it.
Automate the recurring work
I replace one-off extracts with reusable workflows: Alteryx workflows turning ERP tables into recurring extracts that feed Tableau portfolio KPIs, the first portfolio snapshot view for a POS phone financing product with no reporting layer, and a migration of recurring Excel Customer Service reporting to automated Power BI.
Get it used
I am my team's go-to person for AI tools and help colleagues use them for analysis and documentation. Whether people use a tool after go-live decides whether it was worth building.
Toolset
Technical & Delivery Toolset
AI
Cursor · Claude · Microsoft Copilot (daily AI-assisted analysis, documentation, and agent workflows)
Automation & data
Alteryx · KNIME · Power Automate · Power Query · SQL (DBeaver, DbGate, MySQL Workbench) · Python · R · Excel/VBA
Reporting
Tableau · Power BI
Operations & delivery
Jira · Confluence · SAP CRM / SAP CC · Agile delivery with digital tools
Where I'm heading
Where I want to grow
Next, I want a role between the business and the technical teams: finding where AI and automation pay off in lending, building the business case, and seeing it through to a tool people use. My MSc in Digital Strategy and AI Management at RSU, the RTU course in data analysis and process automation with AI, and five years in banking and lending are the base for that.