Qumra Capital: “AI has essentially become our day-to-day assistant”

“AI has essentially become our day-to-day assistant,” said Ofer Vishkin, Associate at Qumra Capital. “It’s embedded across our workflows - from summarizing company materials for due diligence to generating initial competitive landscapes, building TAM estimates, and validating data assumptions.”

For Qumra Capital, artificial intelligence isn’t a distant investment thesis, it’s part of the firm’s daily routine. It runs quietly in the background of the firm’s operations, but Vishkin is quick to note that people remain essential, especially in a market where benchmarks are shifting and competition is fierce.

Ofer Vishkin

“While AI tools are valuable, they’re far from perfect,” he explained. “Even when cross-referencing across multiple platforms, the outputs require refinement and human judgment.”

You can learn more in the interview below.

Fund ID
Name and Title: Ofer Vishkin, Associate
Fund Name: Qumra Capital
Founding Team: Boaz Dinte, Erez Shachar
Founding Year: 2014
Investment Stage: Growth Stage
Investment Sectors: Israeli deal flow (Generalist)

On a scale of 1 to 10, how has AI impacted your fund’s operations over the past year - specifically in terms of the day-to-day work of the fund's partners and team members?

8 - AI has essentially become our day-to-day assistant. It’s embedded across our workflows - from summarizing company materials for due diligence to generating initial competitive landscapes, building TAM estimates, and validating data assumptions. That said, while AI tools are valuable, they’re far from perfect. Even when cross-referencing across multiple platforms, the outputs require refinement and human judgment. Lastly, we’ve also implemented dedicated sources for smarter, more efficient sourcing that look into headcount growth, time since last round, team backgrounds, etc, and provide us with very interesting dealflow opportunities.

Have you already had any significant exits from AI companies? If so, what were the key characteristics of those companies?

Not yet. The AI sector is still in its early stages, and most relatively mature companies are in the growth and scale-up phase. While there were some impressive acquisitions in the industry, the vast majority of exits are likely still ahead.

Is identifying promising AI startups different from evaluating companies in your more traditional investment domains? If so, how does that difference manifest?

From a growth lens, our evaluation framework remains consistent. However, AI investing requires a sharper focus on certain elements - and in many cases, a higher bar. Top AI performers today are scaling faster than any SaaS companies we've seen, so benchmarks are shifting. The competitive landscape is also more intense, with users often trialing multiple tools across categories like image generation or developer copilots.

What specific financial performance indicators (KPIs) do you examine when assessing a potential AI company? Are there any AI-specific metrics you consider particularly important?

When evaluating AI startups, especially in the early stages, we often find that traditional SaaS metrics - like ARR, CAC, and retention - don’t tell the whole story or aren’t yet mature enough. That’s why we take a much more blended approach. We focus on metrics like revenue growth, pilot-to-paid conversion, gross margins, and usage-based engagement, while recognizing that these often require context given how early and volatile the space still is. Then, given how early the AI space is, qualitative diligence plays a bigger role. We spend a lot more time understanding customer workflows, technical differentiation, and the sustainability of performance.

How do you approach the valuation of early-stage AI startups, which often lack significant revenues but possess strong technological potential?

This is something we, as a growth stage fund, do not come across often. However, we would normally try to assess the commercial pipeline as thoroughly as possible, build out forward-looking projections - even if those carry high uncertainty - and then base our valuation on expected future revenues.

What financial risks do you associate with investing in AI companies, beyond the usual technological risks?

Beyond the typical early-stage risks, AI companies often carry high cloud infrastructure costs, particularly for model training and inference. Many are also highly dependent on access to proprietary or third-party data - creating both legal and operational exposure. Additionally, emerging regulations around data privacy, AI transparency, and model accountability could materially impact monetization pathways.

Do you focus on particular subdomains within AI?

We're fairly agnostic in terms of AI subdomains. So far, we've engaged with companies in natural language processing and computer vision - our portfolio includes players like Nexar and Lightricks in those areas. Qumra has also invested in Blocks, an early-stage no-code low-code tool. Recently, we’ve researched numerous subdomains within cybersecurity that incorporate strong elements of AI.

How do you view AI’s impact on traditional industries? Are there specific AI technologies you believe will be especially transformative in certain sectors?

We believe that AI will fundamentally reshape multiple traditional sectors, including healthcare, logistics, finance, and manufacturing. Technologies like generative AI, autonomous decision-making, and predictive analytics are already driving major efficiency gains. In particular, industries with large, unstructured data pools are seeing the most immediate impact.

What specific AI trends in Israel do you see as having strong exit potential in the next five years? Are there niches where you believe Israeli startups particularly excel?

Israel has deep domain expertise in several verticals, and AI is acting as a catalyst in accelerating outcomes. Cybersecurity, for instance, has already experienced acceleration, with areas like data security, AI-driven SOCs, and browser security gaining strong momentum. Obviously, AI security itself is gaining an exceptional momentum with exits such as Prompt, Apex, and Robust Intelligence in a very short amount of time. Another standout is vertical AI: startups like Sensi, Buildots, Aidoc, Eleos, Exodigo, and PhaseV have seen impressive traction. Hence, very high expectations.

Are there gaps or missing segments in the Israeli AI landscape that you’ve identified? What types of AI founders are you especially looking to back right now in Israel?

Yes, we’ve identified some notable gaps. While Israel has produced a strong wave of application-layer AI companies - particularly in cybersecurity, vertical SaaS, and productivity - there’s still a relative lack of foundational AI infrastructure plays. For example, we saw fewer startups working on proprietary LLMs, orchestration layers, inference optimization, or novel model architectures. There’s also limited activity in areas like vector databases, synthetic data generation, or GPU-efficient training frameworks compared to the U.S. or parts of Europe.

That said, the pace is accelerating and we’re starting to see more technically ambitious teams emerge. The most promising startups are often led by highly technical founders with elite academic backgrounds and strong access to top talent. That’s our north star when evaluating early teams: can they tackle deeply complex problems, and do they have access to the right people to execute? I think this gets even more significance in AI vs traditional SaaS startups.