“What’s innovative today might be commoditized tomorrow”: IL Ventures on the new rules for spotting AI winners

“AI is causing a disruption in every market, including how you evaluate an AI startup. Product-market fit and ability to execute still take center stage, but with AI, you need to assess whether the team is actually innovating in a sustainable way,” said Yoni Heilbronn, Managing Partner, and Elad Ziklik, Strategic Advisor on AI to IL Ventures and Former Global Head of AI at Oracle. “Things that used to take startups dozens of people and multiple years, can now be replicated by a small team and a few months, if not weeks.”

The duo joined CTech for its VC AI Survey to share how the new rules for spotting AI winners are changing the investment game. “In AI, timing matters differently. The velocity of change in this space means that what’s innovative today might be commoditized tomorrow,” they added. “So, we’re looking not just for strong products but for teams that can adapt fast, compound learning, and build defensibility beyond the model itself.”

Yoni Heilbronn and Elad Frenkel, IL Ventures

You can read more below.

Fund ID
Name and Title: Yoni Heilbronn, Founder & Managing Partner
Fund Name: IL Ventures
Founding Team: Yoni Heilbronn, Elad Frenkel
Founding Year: 2021
Investment Stage: Pre-Seed, Seed
Investment Sectors: AI, Logistics, Supply Chain, Robotics, Energy, Smart Manufacturing.

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 tools help us be more efficient in our day-to-day, especially when dealing with analysis tasks - from screening to due diligence and market intelligence, as they allow us to crunch a lot of data quickly and get bottom lines. At this point in time, AI has become a decision support tool that is very effective.

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

Not yet.

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

AI is causing a disruption in every market, including how you evaluate an AI startup. Product-market fit and ability to execute still take center stage, but with AI, you need to assess whether the team is actually innovating in a sustainable way. Things that used to take startups dozens of people and multiple years, can now be replicated by a small team and a few months, if not weeks.

Also, in AI, timing matters differently. The velocity of change in this space means that what’s innovative today might be commoditized tomorrow. So, we’re looking not just for strong products but for teams that can adapt fast, compound learning, and build defensibility beyond the model itself.

The other challenge is in how you evaluate success. ARR used to be a metric that you could trust, but in today's AI-powered world, innovation is happening so fast, and the cost of switching is negligible, even for enterprises. You're all in on Lovable? You might be on Bolt next week, or on Cursor a month later. You're all in on this startup that does amazing video generation? Google just released Veo 3, everybody will move there. Meeting summarization with voice? Open AI just released that, and you're done.

You always had to worry about a Microsoft or a Google releasing something that could swallow your business, but these were "slow behemoths" that took forever to innovate. Now, the next startup is maybe a few weeks away from killing your product.

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?

You should still look at the core SaaS-style metrics: gross margins, LTV/CAC, retention, MoM growth, etc. But with AI companies, there sometimes is a layer of model economics: inference cost per user, usage intensity, and scalability of their infra.

If they’re building on foundation models, we look at how that cost impacts gross margins. What is the cost of serving a user? Are they on a path to inference efficiency at scale, or perpetual dependency? If they’re training their own models, how smartly are they using capital to build something defensible?

Also, data flywheel indicators matter: are they capturing proprietary user interactions that improve the product? That feedback loop is often the only place where defensibility lies in today's AI-powered world.

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

Like any other investment, the strength of the team, the big problem being solved, the significant market, the viability of long-term growth, and having a significant technological moat.

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

Dependency risk: Many startups rely on external APIs like OpenAI or Google. If pricing or terms change, the economics can collapse overnight. See, for example, the recent hype around Deepseek and its impact on pricing.

Regulatory uncertainty: Especially in healthcare, finance, or education. AI introduces new questions around explainability, bias, and data use that can delay or derail commercialization. Just last week, Cloudflare announced a new feature where the default on their servers is to now deny AI bots from crawling through your website unless you give them explicit permission (and charge for it). Suddenly, Gemini, OpenAI, etc., could lose access to 25-50% of the internet, unless they start paying for it.

But the biggest risk in my mind is the "ARR illusion" - startups might show promising ARR early on, but switching costs are near zero if the product is just an interface over commoditized models. Customers will churn the moment a cheaper or better wrapper shows up. So ARR is a vanity metric unless the product has real stickiness - proprietary data loops, integrations, workflows, or ecosystem effects.

Do you focus on particular subdomains within AI?

No, we invest in any AI domain that disrupts traditional industries.

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

AI is already reshaping traditional industries, and in fact, every industry. If there's a CXO out there who is not thinking about how their business is getting disrupted by AI, their shareholders should literally fire them immediately. Every industry, every business, is having its Kodak / Blockbuster moment - it's just a question of when and how fast.

Industrial automation, logistics and manufacturing – vision and sensor fusion paired with AI-powered robotics will change how factories, logistics centers and warehouses operate across every domain.

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 historically shines in deep tech. While Israeli tech sort of missed the train a bit when it comes to core models (there are no Israeli startups that are at the top of the foundational LLM models), I’m particularly optimistic about:

• AI-native dev tools and ops layers (e.g., observability, fine-tuning, optimization)
• Vertical agents for sectors where Israel has domain depth - insurance, logistics, or agriculture, precision medicine
• Safety, privacy, and explainability tooling - regulation will demand it, and Israel has the talent

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?

Israeli startups are often engineering-heavy, focusing on tech. You can see many startups that scale by hiring US leaders for sales, marketing, even operations. I think there's a relative gap in product-led, UX-forward AI startups in Israel, ones that think beyond the algorithm and obsess over usability. I think this comes from the background and breeding grounds for all these founders - 8200, 81, etc. - where you focus on getting things done, and less on usability, delight, and customer experience. In the AI space, you need people who can turn cutting-edge tech into delightful, sticky products, and Israeli startups typically do not excel at this.

You should be looking to back:

• Founders who deeply understand a specific pain point or domain, not just the tech
• Teams that blend AI fluency with design and go-to-market muscle. Distribution is everything in the new world.
• Builders who care about long-term defensibility, not just novelty.
• The best ones are usually second-time founders or domain experts paired with top-tier AI talent.