"The check doesn't matter. It's who's writing the check that matters most." Omead Ostadan, president and CEO of Cellanome, said that about Premji Invest. Krista Peters, head of marketing at Premji, cited it as the clearest explanation she's found for why VC marketing is different from almost every other marketing job. It's also the clearest brief she could give herself: when the product is a long-term relationship on a cap table, not a widget with a price, you don't market the check. You build the reputation of the people writing it.

That's a whole different mandate. Premji is an evergreen fund, which means it doesn't follow a standard 10-year fund cycle. The Azim Premji Foundation sends capital, Premji deploys it, and there is no forced timeline driving toward an exit or an IPO. That structure changes the pitch. The goal is to be the firm a founder calls before they're even raising, because they've watched the partners show up thoughtfully in their space for months. I sat down with Krista recently to talk through how she's building toward that. These are the sharpest takeaways. Full recording linked below.

The firms winning founder trust aren’t running campaigns; they’re building a public record.

Krista flagged something from a LinkedIn conference that reframes how to think about the platform. Two completely different audiences are reading a VC partner's LinkedIn profile at the same time, with completely different goals. Agents scan the text: the headline, the byline, the About section, looking for signals of expertise and relevance. Humans look for evidence of a real person behind the account, someone worth getting on a call with. Most VC LinkedIn accounts aren't serving either audience. They announce portfolio news, post broad trend roundups, and stop there.

The accounts that actually build relationships are doing what Krista described as educational posting: answering real questions that founders are asking, in specific enough terms that both a person and a model can extract something useful. As we've written about in our AEO content strategy work, the content that gets cited is the content that answers questions directly. That applies to VC firms as much as it applies to SaaS companies. A public record of what you know and who you show up for is what gets read, cited, and remembered by founders and AI systems alike.

At Edgar Allan, we think about this as the difference between a campaign and a record. A campaign runs and stops, but a record compounds. Every consistent mention, every answered question, every attributed insight trains AI systems and human memory toward you. That's not a content strategy. It's infrastructure.

Specificity is the only way to break through on a noisy platform

Three things came up at the LinkedIn session Krista attended: specificity, relevance, and agents and humans building trust together. Of the three, specificity is the one most VC accounts get wrong.

A post that names the exact tension a founder faces when scaling a Series B in a capital-constrained environment will outperform a post about "five things we're seeing in enterprise SaaS this quarter," every time. Not because it's more polished, but because it's addressed to someone. This is the same pattern we track in AI search visibility work: generic content is readable but rarely citable. Specific content gets extracted.

For Krista, the practical version of this is building content around managing partner Sandesh Patnam's voice specifically, not around Premji as an institution. Founders want to know who they'd be working with. The partner's voice is what tells a founder whether there's alignment, and it's harder to fake than an institutional brand statement.

Turning one partner’s voice into a content system is its own discipline

This is something Edgar Allan has been working through with Krista directly. The process starts with capturing the partner's voice in raw form: recorded calls, voice notes, off-the-cuff observations after a meeting or conference. That raw material gets transcribed and then treated the way any editorial team treats a first draft. It gets shaped, not generated.

The distinction between shaped and generated is where most AI-assisted content falls apart. Pure generation produces content that sounds like everyone else's content. Shaping starts with something real: a specific observation, a proprietary data point, a point of view that couldn't have come from a general model. That's the 10% of AI-assisted content that actually performs.

As Ethan Smith at Graphite has noted, that gap between generated and shaped content is showing up in performance data. What's working is AI-assisted content that starts with real listening. For a solo marketer without a full content team, the workflow Krista described looks like this: document the partner's voice through recorded reference material, use that reference to draft, edit until it sounds like the person talking, then push it into channels where the right audience is already paying attention. This connects directly to what we've been building toward with two-lane content strategy: one lane for volume and visibility, one lane for the trust-building content that only comes from a real point of view.

The next frontier: audience research agents that do the listening work upstream

Krista's clearest content priority right now is building a system where agents surface who is paying attention to Premji's content, what questions they're asking, and what gaps exist between what the firm is publishing and what founders are actually searching for.

Edgar Allan is currently testing an audience research agent workflow using Profound to support exactly this kind of upstream listening. The goal is to surface topic gaps and audience signals before the content brief is written, so that what gets produced is already mapped to real demand, not editorial instinct. It's the same approach we use in our AEO audit work: establish a baseline of how your brand currently appears in AI-generated answers, then identify where the story is inconsistent or missing before you produce another word of content.

The full recording of my conversation with Krista covers all of this in more detail, including her background as a certified sommelier (which, it turns out, is more relevant to the job than it sounds), how Premji's evergreen structure shapes its approach to founder relationships, and where she sees VC marketing going over the next twelve months.

‍

‍

Edgar Allan built Premji's website and helped shape their brand. The voice documentation approach Krista now uses to draft content in Sandesh's voice came directly out of that working relationship. If you're a solo marketer in a relationship-driven business, the system she's building is worth studying closely.