Qubera: The Sword Startup Challenge finalist processing alternative investments

Meet Tanvi Singh, co-founder of Qubera. The fintech startup makes alternative investment decisions auditable and fast. Tanvi is one of the 5 finalists of the Sword Startup Challenge, competing for venture clienting opportunities, expert support, and industry exposure.

Sword Switzerland, a global technology transformation company, has launched the second edition of the Sword Startup Challenge, its Swiss innovation competition dedicated to startups developing solutions for the wealth management sector. Organized in collaboration with Venturelab, the program began in January and will run through June 2026.

In which situations does your solution create the most impact, and who is usually involved on the customer side?
The highest-impact moment is the weeks before an investment committee meeting, when a team decides whether to commit capital to a fund or a direct deal. They are buried in PPMs, LPAs, track records, and side letters under tight deadlines, with no structured way to cross-check findings. We compress this from weeks to hours, with every result traceable back to the exact clause or table it comes from. The main users are investment analysts and portfolio managers running diligence, the CIO approving decisions, and increasingly compliance and risk teams who need the process to withstand scrutiny. In wealth management, client advisors also benefit, as they can base recommendations on a defensible analysis rather than instinct.

What kind of validation have you seen so far that your solution creates value?
We have one design partner and two trial clients processing live fund documents in real workflows. We were also invited to BNP’s Global Family Office Conference to present Qubera. This is active diligence rather than sandbox testing. Feedback shows tasks that once took days of extraction and reconciliation now take a fraction of the time, with improved traceability and confidence. On the research side, we published FinSheet-Bench with the University of Zurich, testing leading AI models on complex financial spreadsheets. The best standalone model reached 82.4%, while performance on the hardest multi-fund files dropped to 48.6%, confirming that general-purpose AI is insufficient for this use case. We also placed Top 10 at the Swiss FinTech Awards 2025 and completed an Innosuisse-funded study with UZH, both independently evaluated for technical and commercial merit.

If you started a pilot with one of Sword's clients, what would success look like in the early stages?
Concretely, three things in the first four to six weeks. First, we process a set of the client's actual documents, PPMs, quarterly reports, capital call notices, and deliver structured, source-traced outputs that they can compare against their own manual work. That side-by-side is where it clicks. Second, compliance or risk confirms that the audit trail holds, that every data point links to a specific page, clause, or table. In regulated environments, this is often what makes or breaks adoption. Third, and this one matters most to me personally: two or three people on the team actually use it in a real decision, not as a test. If it becomes part of how they work rather than something they evaluate, the pilot worked.

From your perspective, what makes a collaboration between a startup and a large corporate successful?
Being honest about what each side actually needs. The corporate environments bring complexity that no startup can simulate, real regulatory constraints, messy data, and internal politics. We bring speed and the willingness to rebuild something if the first version misses the mark. That exchange only works when both sides commit to a real use case, not a sandbox exercise, and when there is someone on the corporate side who owns the outcome and can actually make decisions. From our experience, the pilots who work are the ones where the client puts real documents on the table from day one. That forces everyone to deal with actual complexity. Polished demos tell you very little.

What specific opportunity do you see in the Sword Startup Challenge, and how would you leverage it to create value with one of Sword's corporate clients?
Sword's clients manage significant allocations to private markets. The diligence behind those allocations is still manual, fragmented, and hard to audit. Most of them know this is a problem; they just have not had a clear path to fix it. What we bring is not a pitch deck; it is a working product. We can take a client's fund documents, process them within hours, and deliver structured, traceable outputs they can benchmark against what they do today. The Sword Challenge gives us the right setting for that, an introduction with shared context rather than a cold email. If we can show one wealth management team that their three-week diligence cycle runs in hours with full traceability, that becomes a reference for the rest of the network. That is what we are here to do.