When we launched the AI Product Ideas Contest inside Sigma Software, we weren’t sure what to expect. Would people engage? Would the ideas be serious? Would engineers who spend their days in code and client deliverables take time to think about this?
30+ ideas in one month answered that question.
The most powerful ideas were already inside
In April 2026, we asked the people of Sigma Software a simple question: if you had an AI product idea, what would it be? In one month, we got 30+ answers from engineers, designers, recruiters, accountants, and salespeople. And several of those ideas are already being built.
Most companies, in their chase for AI integration, look outside, scout startups, hire consultants, buy vendor tools. We were sure we didn’t have to. We had something better: our own people.
More than 2,000 specialists across 20+ countries, working inside real businesses every day in aviation, finance, healthcare, and more domains. These are people who see exactly what’s broken, because they live with it. If anyone knows where AI could actually help, it’s them.
So instead of looking for ideas elsewhere, we launched the AI Product Ideas Contest and asked our own specialists to bring theirs. And as the venture builder arm, we could promise something real in return: the best ideas would get backed, developed, and built.
“The best product ideas come from people who understand the problem deeply. Our specialists – engineers, designers, recruiters, accountants – work inside real businesses every day, with deep domain knowledge across industries from aviation and automotive to finance, healthcare, and defense. They see what’s broken before anyone else does. This contest proved that the potential to build something meaningful was already here. That’s exactly what our Venture Builder is for: turning that potential into real products and companies.”
– Valery Krasovsky, CEO & Co-founder, Sigma Software Group
Everyone had something to say
Engineers, designers, recruiters, accountants – ideas came from every corner of the company. Thirty-plus of them in a single month, from people across roles, time zones, and business units, many of whom had never thought of themselves as “product people.”
That tells you something. When you give people a real channel and a real reason to use it, they show up. And they come with ideas grounded in something real: the daily friction they actually feel.
Eugene Park, AI Support Transformation Department Lead at Sigma Software, put it well:
“The ideas presented in the contest show a strong understanding of real internal needs and a genuine drive to improve existing processes. What’s especially valuable is that participants didn’t just propose solutions – they thought through practical implementation and impact. AI makes it significantly easier and faster to develop initiatives like these. The strongest ideas are always rooted in real problems, and that’s clearly visible in this contest. As AI continues to evolve, the ability to act on those ideas and deliver tangible results is growing faster than ever.”
– Eugene Park, AI Support Transformation Department Lead
From spreadsheet to stage
Thirty-plus ideas went through review. Ten made it to Demo Day.
The format was intentional – not a formal pitch competition, but a conversation. Finalists presented their ideas to a jury that included board-level stakeholders, and the feedback went both ways. It wasn’t about who could sell the room. It was about which ideas had real legs.
Two winners emerged. Yevhenii Lymar, Senior Product Manager, won with an AI Candidate Pre-screening Tool – a system designed to take the repetitive part of technical interviews off senior engineers’ plates. Juliia Krivitskaya won with an AI B2B Negotiation Simulator, a tool to help sales teams practice high-stakes conversations before they happen for real. Each walked away – and something more useful than that: a path to actually building it.
Yevhenii said it better than we could:
“What I’m most excited about isn’t the win itself – it’s that we’re now moving toward a pilot with real roles and real candidates.”
That’s the point. The prize was never the finish line.
What comes next
Winning wasn’t the end of the process – it was the beginning of it.
Three ideas are already moving into proof-of-concept solutions: AI Time Tracking Copilot, AI Pre-Sales & Proposal Copilot, and AI Candidate Pre-screening Tool. Ideas with internal application go straight into implementation. Ideas with broader market potential get mentorship and support from Labs.
This is what our venture builder actually means in practice. Not a fund that writes checks from a distance, but a structure that helps people inside the company turn real problems into real products.
And we’re not stopping here. Season 2 launches on August 1 – and the goal is 50+ ideas. Not because the number itself matters, but because every idea beyond the first season’s tally of 30 represents someone who didn’t think they had anything to contribute and then changed their mind.
