Studio

Finding Product–Market Fit: The Lapis Approach to Rethinking Market Focus

Studio

Finding Product–Market Fit: The Lapis Approach to Rethinking Market Focus

Words Founders Factory

October 20th 2025 / 8 min read


When a startup is built around a powerful piece of technology, deciding to focus on a specific industry can feel like a big decision.

In Lapis’s case, building a Knowledge Bank could be applied to any industry. Institutional knowledge drains are everywhere in businesses and Lapis’s knowledge infrastructure solves this succinctly by crawling databases, selecting the right information, and highlighting what’s important to those who need it. But finding a sector where the problem is most acute, and where organisations are ready to adopt a solution, is what determines whether a technology becomes viable in-market.

This is the story of Lapis, backed by Founders Factory and Mediobanca through our MBSpeedUp Accelerator, reimagining how knowledge is stored, retrieved, and applied inside financial institutions - and how its founders deliberately narrowed their focus to a precise, high-value product for investment professionals.

The Mediobanca relationship became the catalyst for that shift. More than just a source of capital, it gave Lapis access to real users, real data, and real workflows inside a complex financial institution. This partnership allowed the founders to test and refine their product in an environment where information retrieval isn’t just a technical challenge but core to how deals are made, portfolios are managed, and ultimately how institutional memory is preserved.

Interested in receiving more insights like this ? Subscribe to our newsletter and join 20K founders, investors & innovators.

Subscribe here

From AI research to knowledge infrastructure

Lapis is a deep-tech company with an academic background. The startup was founded by Pietro, a medical doctor and anesthetist and Davide, a machine learning researcher; and was the brainchild of research collaborations with the AI Institute at the University of Birmingham, King’s College London, and UCL. 

Initially, they explored opportunities in healthcare where vast amounts of data are stored, but they quickly encountered slow procurement cycles, fragmented IT systems, and regulatory barriers that made early commercial traction challenging.

Rather than forcing a fit or creating horizontal enterprise software, the team paused and asked: where is this problem sharper, and who is ready to solve it now? 

Only after working with finance advisor Matteo and eventually Mediobanca, did they realise the opportunity in large-scale banks.

Building the Knowledge Bank for investment banks

Inside investment banks, critical information is spread across decades of PDFs, PowerPoint decks, emails, shared drives, and data subscriptions. Key deal history often lives inside a partner’s inbox, or their mind, and disappears entirely when they change roles. Analysts spend hours searching for documents that may or may not exist, with no single source of truth to rely on.

The vision was not “AI for finance” built and applied to any business, but to create a single point of access for an organisation’s private and public knowledge.

Lapis ingests multiple data sources: PDFs, shared drives, deal databases and builds a structured, searchable layer on top. Users can query these sources, retrieve buried information, extract tables and data points, and assemble financial models and memos, all within one interface.

Unlike many AI tools that rely solely on chat interfaces, Lapis is workflow-aware. Analysts don’t always need a chatbot; sometimes they need a precise table extraction, sometimes a document search, sometimes a memo assembly. Lapis tailors the interaction mode to the task, making it far more useful in daily professional settings.

Equally important is accuracy and traceability. Every data point surfaced by Lapis links back to its original source. If information doesn’t exist, the system says so. This combination of deterministic retrieval and selective AI application avoids the hallucinations that can undermine trust in enterprise settings.

According to the founders, Lapis is the institutional “librarian” that never forgets. A knowledge bank that captures, structures, and retrieves decades of information reliably.

Testing and scaling inside real institutions

The Mediobanca partnership gave Lapis more than funding: it gave them a live environment to test and refine their product.

They began with a single team, tackling specific portfolio management workflows inside a mid-sized bank. As results spread internally, other teams adopted the platform. This bottom-up expansion, starting with one team, demonstrating clear value, then scaling department by department, became central to Lapis’s go-to-market strategy.

This strategy echoes a pattern seen in some of the most successful B2B SaaS companies: embed deeply, solve a real pain point, then scale horizontally within the organisation.

Rethinking verticals to find fit

Lapis’s journey illustrates how rethinking your vertical can unlock product–market fit. Amazon famously expanded from books into broader ecommerce. Slack emerged from a gaming venture into enterprise communication. Lapis similarly evolved, not by abandoning their technology, but by placing it where it could solve a sharper problem more effectively.

Their ambition is larger than finance. As Pietro and Davide put it, they’re building the infrastructure to reshape how knowledge is managed in professional environments - starting in finance, but with potential applications far beyond.

Their story is a reminder that for deep tech startups, the right industry fit is often discovered, not predetermined. By deliberately shifting from hospitals to financial hubs and embracing the concept of the knowledge bank, Lapis has carved out a distinctive and scalable place in the market.

We're looking for ambitious founders in fintech.

Learn more about our MBSpeedUp Accelerator.

Learn more
Words by
Share article

Latest Articles

factory news

Investing in Qi4M’s OrcheXtra: The Agentic Ecosystem to Manage Investment Cycles with Precision

Learn more about Qi4M, and how they’re transforming investment products through an agentic ecosystem that connects data, processes, and people across the investment value chain.

trends

Venture Building with AI Agents

The future isn't AGI replacing industries. It's thousands of specialized agents, each excellent at one task, orchestrated together. Hear from our Product team on their experience of building AI agents over the last 14 months.

trends

Understanding the US-Australia Mining Framework

This collaboration between two of the world’s most resource-rich and innovation-driven nations sets the stage for an industrial era built on shared investment, faster deployment, and technology-led sustainability.