Investor Fabian Westerheide: Can AI also be used for VC? Yes, but.

Our author Fabian Westerheide was already investing in AI long before everyone else was. And he also uses AI to invest. What founders need to know.

Fabian Westerheide is a founding partner of the AI-focused venture capital firm AI.FUND and has been investing privately in AI companies through Asgard Capital since 2014. Westerheide provides strategic advice to public and private institutions in the field of AI and hosts the annual AI conference Rise of AI in Berlin. In this article, he takes a closer look at the scene – revealing who truly understands AI and who just pretends.

Artificial intelligence (AI) is currently permeating every industry – including the world of venture capital (VC). But how much is it really changing our business? And in what areas does VC remain remarkably analog?

I've been an investor in the AI ​​space for over a decade – first at Point Nine Capital, later with Asgard Capital, and now at AI.FUND. We took AI seriously early on, not just as a "trend," but as a core technology. Accordingly, we are deeply embedded in the ecosystem. This gives us a good perspective on the question: What is AI really changing in the VC business—and what isn't it changing?

Let's start at the beginning: fundraising . Anyone who believes that funds can be sold like products in the future—with AI-driven LP scoring, automated pitches, and chatbot follow-ups—underestimates the nature of the business.

Fundraising isn't a funnel, it's a relationship. It's about trust, credibility, and shared experiences – and that takes time. Winning an institutional investor takes years, not weeks. AI can help with data rooms, reporting, or CRM – but the closing happens on a human level. And it will remain that way.

The situation is quite different when it comes to deal flow. "AI" and "KI" are now present in almost every pitch deck – which is no coincidence in our case: As a specialized AI investor, we naturally attract precisely these companies. The volume is enormous. To handle this effectively, we rely on a broad internal tool infrastructure: automation, scoring systems, and our own GPTs for analysis and research.

Our database contains over 20,000 AI startups – with metadata, market observations, investor structure, and team data. And new ones are added every week. With the help of GPTs, our analysts can now automatically cluster, compare, test hypotheses, and identify themes. This allows one person to work with six partners simultaneously – an efficiency gain that would be unthinkable in the traditional VC structure. Deal flow today is data-driven – not as a replacement for, but as an accelerator of human decision-making.

And then comes the portfolio – and the return to reality. Reporting is still manual. Every company delivers its figures differently: different metrics, different formats, different cycles. Standardized dashboards? Hardly. AI can provide selective support here – for example, in the preparation of KPIs, the identification of outliers, or the automated creation of reports. But without prior standardization, it remains a fragmented puzzle.

Especially in early-stage investments , where fully developed systems are not yet in place, portfolio management remains an operational, often manual business. AI can help—but it can't bring order to the chaos without data order.

VCI isn't a process business in the traditional sense. It's a blend of intuition, networking, analysis, and timing. AI helps us structure the flood of information—especially in deal flow. But it doesn't replace relationships, decisions, or trust.

Anyone who wants to be successful as a VC today needs both: technological excellence and human judgment. AI can accelerate, focus, and reveal patterns – but it remains a tool. It's not a substitute for what ultimately defines this profession: the right instinct for the exceptional.