How AI is reshaping startup fundraising: Insights from two investors

On April 9, we brought together Ashok Kamal from NuFund Venture Group and Sigvards Krongorns ex-Verge HealthTech Fund to talk about something everyone's discussing but few are doing well: using AI in fundraising. We talked about real workflows, real tools, highlighting how investors are already integrating AI into sourcing, evaluation, and decision-making.

TL;DR: AI is already embedded in investment workflows, from research and deal intake to screening and expert matching, but it's augmenting decisions, not making them. The sheer volume of deal flow has made automation necessary. Key takeaway for founders: AI has raised the bar on material quality, but fit and relationships still matter most. The future isn't AI replacing investors; it's allowing investors who use AI to outperform those who don't.

The scale problem: why AI became a necessity

On the webinar, both investors opened with the same basic reality: the sheer volume of deal flow has become unmanageable without AI.

Sigvards's fund reviews over 1,000 companies a year and invests in around 10. Ashok sees roughly 1,500 deals annually and invests in about 15. The challenge isn't finding opportunities; it's separating signal from noise at scale.

This volume problem is also why so many founders never hear back. Previously, it was nearly impossible to process that many applications and ensure everyone got a response in a consistent, timely way. In this context, AI isn't a luxury; it's a necessary infrastructure. Without automation and augmentation, managing such volumes would remain both inefficient and inconsistent.

Where AI fits in the investment process

The conversation made clear that AI isn't replacing any single step; it's embedded across the entire funnel, just differently depending on what the investors are trying to do.

At the research and sourcing stage, investors use tools like Gemini and Perplexity to map and deep dive into industries. In healthcare, that might mean analyzing disease prevalence and biological mechanisms. In fintech, it's the regulatory frameworks and payment infrastructure. The goal is the same: build deep mental models before evaluating individual deals.

Deal intake and screening is where things get practical. Inbound deals from emails, events, or platforms can now be automatically converted into structured applications. Tools like Dealum's AI assistant, for instance, extract key data directly from pitch decks and pre-fill application forms, eliminating manual data entry while ensuring consistency across submissions.

Then comes pre-screening. AI analysts score startups and filter out those unlikely to fit the fund's investment thesis. But this is also where limitations show up. Both investors mentioned false negatives, startups rejected by AI that later turned out to be strong opportunities. Human overlook and contextual judgment still matter.

During evaluation, AI acts as an assistant, not a decision-maker. It helps analyze market size, team composition, and business models, but final investment decisions remain human-led. As Ashok put it bluntly: AI doesn't decide which company to fund. It improves how fast and how well investors make those decisions themselves.

One of the more interesting real-life AI applications is expert matching. Instead of the human deal flow manager manually assigning reviewers, AI identifies the right subject matter experts based on deep criteria, regulatory experience, and specific technical backgrounds, making the quality of evaluation significantly better.

From automation to augmentation

A key theme throughout the discussion was the distinction between automation and augmentation.

Automation implies replacing human judgment, while augmentation enhances it. Both investors strongly believe the future is augmentation – AI processes data, identifies patterns, surfaces insights, but it can't replace intuition, context, or relationship-building.

Sigvards highlighted that early AI experiments often produced overly optimistic outputs, where every startup appeared promising. This reinforced the importance of customizing AI tools with personalized frameworks and instructions. Default settings produce generic results; meaningful outcomes require teaching AI how you think.

Trust but verify: managing AI limitations

Despite its benefits, AI introduces new risks. The panel repeatedly stressed the importance of verification.

They combine AI-generated insights with internal discussions, expert validation, and external conversations with industry stakeholders. In specialized sectors like healthcare, that means consulting doctors, insurers, and domain experts, grounding assumptions in real-world knowledge.

They also rely heavily on public data and trusted APIs rather than accepting AI outputs at face value. If an AI claims something about market size, check it against actual datasets.

What this means for founders

For founders, AI has raised the baseline.

Basic hygiene now matters more than ever. Pitch decks, data rooms, and financial models need to be consistent, error-free, and logically aligned. AI tools can instantly detect inconsistencies – the claim on slide eight must match your financial model in an Excel file. Sloppy materials are immediate red flags.

But here's the thing: materials alone don't determine outcomes. Fit matters most. Founders should focus on targeting the right investors – by stage, sector, or geography – rather than endlessly optimizing pitch decks.

Relationships still matter, too. Despite increasing automation, venture investing remains fundamentally human. Meeting investors, building trust over time, and demonstrating consistent progress significantly increase funding chances.

Start conversations early. Investors track founders over time, looking for progress and persistence. As someone said during the discussion, investors invest in lines, not dots. They value long-term trajectories, not single interactions.

The future: AI as a co-pilot

Looking ahead, both speakers agreed that AI will become deeply embedded in investment workflows.

But it won't replace investors. It'll function as a co-pilot, supporting decisions, accelerating processes, and enabling better judgments. Funds that don't adopt AI risk becoming slower and less competitive.

At the same time, the core of venture capital won't change. Human judgment, relationships, and trust remain central. The future isn't AI replacing investors. It's investors who effectively use AI outperforming those who don't.

What we learned

The webinar clarified something important: AI is already transforming fundraising, just not in the way many expected.

Instead of automating decisions, it's reshaping how investors work, making processes faster, more data-driven, and more scalable. For founders and investors alike, success depends on integrating AI into workflows while keeping the human elements that ultimately drive investment decisions.