VCs fire analysts in favor of AI - but humans needed to add value

Automate (DVC fund runs on AI agents - what’s the cheat code to impress them?),

Trust (value-creation insights from my debate with Unusual Ventures’ managing partner)

Remember (a 19-year-old raises to fix memory + feedback in AI),

Acquire (AI tools start buying human-run services to speed enterprise adoption).

Creed is an AI companion built on Christian values with fast early traction (1M+ conversations, 20k WAU in ~8 weeks).

This is a pinch-me moment: several month ago Dmitrii was a brilliant indie developer who scaled to 1M users bootstrapped, and now is scaling with a16z Speedrun!

https://speedrun.a16z.com/companies/creed

Get an intro to Dmitrii while he’s still raising. Or fill this form to be featured: https://forms.gle/ZQ2HR5BKgMTA5E4t7

The news: DVC closed $75M and removed its analyst team, shifting sourcing/diligence to AI agents plus LPs who contribute and receive carry.

Nick Davidov’s take: ‘What used to take a full day for an expensive analyst can now be done in minutes for under 30 cents’.

Who are DVC? Marina and Nick Davidovs have been investing in AI for years now, being the early investors in Perplexity, Thinking Machines and Higgsfield. They are immigrants from Russia, just like me. And they do support immigrant founders - remember that Perplexity have an Indian and a Russian on their founding team, and Mira Murati is Albanian-born. Point is: what DVC do is worth at least watching.

What they claim to have done: replaced their analyst team with AI, that can do pre-screening and write a deal memo, while saving both investor and founder time on useless calls.

The secret sauce: they watch for soft signals, like team dynamics and connections. An example could be a Sequoia partner connecting with the founder on LinkedIn.

(Now we know what to do in the era of AI-powered funnels - look for top fund connections! But you knew that, no doubt.)

Why it matters: Automation will blow up top-of-funnel volume. In reality, it already is: I use AI in my own scouting, hear other VCs do the same. Decile Group have their own Decile Hub AI assistant. And more tailored tools are developed by various startups!

What’s on the other side of the moon? Value creation.

Last week I had a fabulous chance to debate in the comments of Unusual Ventures founder and Stanford lecturer John Vrionis. So grateful for this opportunity to learn - from an experienced human!

John argues there are three ways to play VC today: raise the most money, truly differentiate for founders, or play high volume.

I’ve proposed number 4 - media. But as valuable as attention is, its contribution is mostly around the #1, plus getting you into oversubscribed deals.

However, ss per John Vrionis, very few firms actually master #2 (real value-add).

2% of VCs generate the most value add in the VC asset class, and the rest 98% mostly rely on volume/luck, and founders suffer from bad advice.

My takeaway:

While AI-powered venture funnels blow up, investors will still invest into humans. Because the founders are who drive the startup to success. And this means - caring about the personality, recognizing team dynamics, and building trust.

VCs would likely still need humans in the loop.

Though they might not be called analysts any longer.

The news: TechCrunch’s “reinforcement gap” piece explains why certain capabilities (e.g., image gen) compound quickly while reasoning lags — tight feedback loops + long-term memory are the difference.

Why it matters: The next step-change isn’t only bigger models — it’s longer horizons of memory and denser feedback. Whoever owns the memory/reinforcement layer will capture loyalty and accuracy.

Founder checklist: show how your product 1) accumulates durable memory safely, 2) owns the feedback loop, 3) gets the value out of memory to users, or enterprise clients.

The news: Prezent raised $30M and immediately began acquiring presentation-design services firms (starting with one the founder previously built). The secret sauce: services already trusted by enterprise; inject AI to scale outcomes fast.

What surprised me here was the reverse of the acquisition pattern, and likely the value creation pattern.

It used to be like: there is a company with its client and processes - a design firm, legal firm, accounting firm. And they implement AI. Now AI could be the driver of value - whereas the professional firm is the trust building layer of customer acquisition.

Trust, again!

Why it matters: This could be the new AI monetization pathway. Instead of enterprises slowly adopting AI, AI startups acquire service businesses to force adoption. Expect more AI roll-ups (agents buying agencies; tools buying consultancies).

Let’s keep this a place where great founders meet the right capital — early, warm, and real.

Let’s continue building!

And keeping humans we can trust in the loop.

Have a great week of ventures and adventures!

Jane