How These Founders are Redefining Analytics with Multi-Agent Systems

Alexis Clarfield-Henry

For years, Ibrahim Ashqar had a front-row seat to one of the most frustrating inefficiencies in modern business: companies drowning in data but struggling to extract meaningful insights. As a data scientist for over ten years across both large enterprises and startups, he saw firsthand how traditional business intelligence (BI) tools, often synonymous with endless dashboards, failed to deliver real value. Instead of empowering teams with actionable insights, they created more confusion, more manual work, and, ultimately, more bottlenecks.

At every company he worked at, the same pattern emerged: data teams were overwhelmed with ad-hoc requests, while business leaders waited days, sometimes weeks, for answers that could make or break key decisions. “More dashboards doesn’t mean more insights,” Ibrahim realized. “The most valuable decisions still required deep manual analysis—something most teams just don’t have the resources for.”

Ibrahim wasn’t the only one frustrated by this reality. His co-founder, Tudor Boiangiu, P. Eng., MBA, spent his career bridging the gap between data analytics and business strategy, leading initiatives that helped Fortune 500s turn complex data into actionable insights. With a deep understanding of both the technical and operational challenges enterprises face, Tudor is the perfect partner to help reimagine how businesses interact with their data––in natural language and without technical barriers.

But the technology just wasn’t there yet.

That all shifted with the rapid advancement of large language models (LLMs). Suddenly, it became possible. And they knew if they didn’t take advantage of this opportunity, someone else would.

The Birth of Lumi AI

With their combined experience in data and analytics, Ibrahim and Tudor saw an opportunity to break the cycle. In 2023, they founded Lumi AI, a conversational analytics platform that acts as an AI-powered data analyst, unlocking insights buried deep in enterprise data systems. Their vision was clear: remove the dependency on overwhelmed data teams, eliminate the need for complex SQL queries, and give business teams the power to explore their own data—instantly.

But building Lumi AI wasn’t just about integrating LLMs into analytics. “You can’t just connect an LLM to structured databases and expect magic,” Tudor explains. “The reality is, enterprise data is messy. It’s wide, vague, inconsistent. A simple question can require significant effort to answer due to unclear relationships, missing business context, and tribal knowledge locked in people’s heads.”

Most companies trying to integrate LLMs into analytics rely on text-to-SQL models, which pull only basic metadata, like table and field names, using simple retrieval techniques. But this approach is fundamentally flawed: Vague table names make queries unreliable, poor field descriptions lead to misinterpretation, missing relationships break joins, and lack of business context means key metrics aren’t understood.

Even if LLMs could generate perfect SQL, that wouldn’t be enough. A competent human analyst does much more—they ask clarifying questions, debug errors, summarize insights, explain assumptions, and suggest actions based on business context. Unlocking LLMs’ full potential in analytics requires a structured, agentic approach, where multiple specialized AI agents collaborate—just like a skilled data team would.

This is what sets Lumi AI apart. Instead of treating AI as a simple query generator, Ibrahim and Tudor built a multi-agent system, where different AI components work together to refine outputs, validate results, and provide meaningful insights with the necessary context.

The result? An AI-powered analyst that doesn’t just retrieve data—it understands the problem, explores nuances, and delivers insights business teams can trust.

Moving Forward With Conviction

Ibrahim’s path to building Lumi AI wasn’t just shaped by technical expertise—it was shaped by his deep understanding of the frustrations businesses faced. At Stord, where he previously led analytics, he experienced firsthand how even the most sophisticated dashboards failed to meet the needs of business teams. Despite a highly capable 12-person data team, they were constantly bombarded with requests around supply chain optimizations, sales forecasting, investor reporting, and more.

That experience cemented his conviction that analytics needed a new paradigm. And with Tudor, they were the right team to build it.

But, like any startup at the forefront of new technology, the road isn’t easy. Building agentic workflows that mimicked a human analyst was technically complex, with endless edge cases to solve. On top of that, navigating enterprise B2B sales—with long cycles, security clearances, and AI governance councils—was a challenge in and of itself.

Despite the challenges, one thing kept them going: market reactions.

“Seeing the sheer joy on someone’s face when Lumi instantly answers a question they’ve been struggling with for months—something they couldn’t get from their existing BI tools––is what kept us going,” Ibrahim says. That has been the driving force behind our journey.”

And the market reaction has been undeniable. “Almost every time we get in front of someone, the problem resonates immediately, and they love our approach,” Tudor shares. The demand spans industries like retail, grocery, and manufacturing. Anywhere complex supply chains exist, companies are eager for a better way to extract and act on data.

Beyond product validation, Lumi AI’s early customers have played a crucial role in its success. “We’re lucky to have incredible customers who’ve not only shaped our product but even helped us secure seed funding by vouching for us to investors,” Ibrahim says. “Forever grateful to those people—you know who you are.”

The Founder Journey: Relentless, Resilience, and Focus

For Ibrahim, the journey of building Lumi AI has been one of relentless problem-solving. Building a startup, he admits, is tough. “Most of the journey is terrifying, with moments of euphoria,” he says. “But those highs—they keep us pushing forward.”

If there’s one lesson he’s learned, it’s the importance of staying laser-focused on the problem and maintaining tight feedback loops with customers. “Listen to the market,” he advises. “They’ll tell you what’s working—and what’s not.”

He also emphasizes the importance of having a strong support system. The pressures of fundraising, product development, and customer demands can be overwhelming. Without the right people around you, burnout is inevitable.

Looking ahead, Ibrahim and Tudor’s vision for Lumi AI goes far beyond conversational analytics. “We’re building towards a future where Lumi doesn’t just surface insights—it acts on them,” he explains. “Imagine an AI that not only identifies opportunities but also executes strategic business actions based on what it finds. That’s the real promise of agentic workflows, and we’re leading the charge.”

Why Forum Ventures?

As first-time founders, Ibrahim and Tudor knew they needed a VC partner who understood the early-stage journey. Forum Ventures was an early believer in Lumi AI, providing not only capital but also crucial introductions, pitch guidance, and go-to-market resources.

“They were one of our first checks, and they’ve been incredibly founder-friendly,” Ibrahim says. “They believed in us early on, and that made a huge difference.”

The Next Chapter for Lumi AI

With $3.7 million in seed funding from AgFunder, Lumi AI is scaling rapidly, expanding its engineering team, and continuing to refine its agentic AI approach. But for Ibrahim and Tudor, this is just the beginning.

“We’re proud of what we’ve built—not just as pioneers in this space, but as thought leaders setting a new standard for AI-powered analytics,” Ibrahim says. “Our journey has just begun, and we’re continuously pushing the boundaries of what’s possible.”