Technical Context
I was immediately drawn not to the poster but to the lineup. This isn’t another talk about “the future of AI” but a solid mix of people digging from different angles: AI agents, AI automation, finance, and management. For those thinking about AI implementation in a company, this format is usually more useful than a dozen calls with software salespeople.
Factually: Vitaliy Kotivskyi is listed as an AI Architect and founding engineer, Dmytro Havrosh with a focus on automation and business processes, Maria as a fractional CFO with hands-on AI integration in finance teams, and Yarema Hrytsyshyn with the management angle. The panel topic itself is also no-fluff: how AI agents actually work, what’s happening with agentic coding, where AI delivers savings, and where it just eats budget.
I particularly like that they’re not trying to lump copilots and agents into one mush. If the conversation is honest, the most valuable part will be the delineation: where a regular workflow is enough, where you need an agentic scenario with checks and exceptions, and where it already makes sense to build a full-fledged agent with access to data and tools.
Another plus: the finance section. In my projects I constantly see teams wanting to “let AI into accounting” without sorting out access rights, risks, verification, and boundaries of responsibility. If they discuss that without magic on the panel, that’s already a win.
Business and Automation Impact
I would go there not for inspiration but for filters. The first filter is simple: which processes are even worth automating right now, and which ones are too early to touch. The second: when AI integration pays off, and when it’s cheaper to stick with strict regulations and conventional automation.
Business owners, product folks, and operations teams already suffering from manual processes, approvals, reporting, and first-line analysis will win. Those still buying “AI magic” without architecture, data access, and a proper process owner will lose.
I constantly break down these forks with clients: where you need an agent, where a scenario is enough, where everything breaks on access rights and data quality. If after this panel you realize it’s time to stop talking and start building a working scheme, at Nahornyi AI LAB we can calmly dissect your process and build AI automation exactly where it truly lifts the burden from the team rather than adding a new toy to the stack.