645 Thought Leadership

The Value-Add Playbook, Rebuilt

A fund's edge used to be who you knew. It's becoming what you can turn into a tool a founder can actually run — and the discipline to know when not to build one.

I've spent the last five years on the platform side of venture, the team responsible for the value-add a fund brings beyond the check. For a long time, the biggest edge a VC fund could offer its founders was its relationships. Knowing the right candidate before a search even started. Connecting you with an advisor who has experienced your exact problem firsthand. Opening the door to your dream customer. Value-add was powered by a fund's network, one relationship at a time.

That is changing faster than most of us are ready for.

Some of that will never change: a warm introduction and good advice at the right moment will always help shape outcomes. But there is a limit to how far we can get involved. We can get a founder in the room with the enterprise buyer, but we can't close the deal for them. We can get the meeting with the candidate everyone is chasing, but we can't sell them on the mission or make the hire. Execution is the founder's job, and our job is to build the muscle.

With the rise of AI, we are at the start of a real shift in how firms like ours create value. A new kind of value-add is emerging, and that is where firm-built founder AI tools come in. They don't replace the introduction or the advice; they help founders do more with both.

The work and expertise that used to require a person are starting to fit inside a tool. Not a document or a template, but an agent a founder can actually run. Drop in your messaging and get it graded. Hand over your data room and get back a punch list of what is missing. Describe your ICP and get a qualified account list with outreach drafted. The static playbook we used to hand over as a PDF becomes something a founder can put to work at 11 pm, without waiting on anyone's calendar.

Many of the leading funds are already moving on this shift: some have open-sourced skill packs for founders, others run live build sessions or hand out model credits by the thousands. But the question is no longer whether portfolio support becomes AI-led. It is who does it well.

And here is where I think a lot of funds are about to get it wrong.

The reflex, when a new capability shows up, is to ship as much of it as possible. Publish a big skill pack. Rack up stars on GitHub. Announce that you are AI-forward. It looks like progress, and it photographs well in an LP update. But shipping volume is the wrong scoreboard, and it can quietly make a founder's life worse.

A founder's scarcest resource is attention. The tooling landscape is already deafening, with new AI and GTM tools launching every single day. Handing a founder one more generic tool, only now with your logo on it, is not a gift. It is another tab they have to evaluate and abandon, and time they could have better spent.

So the real question is how do you build a valuable founder AI tool that is also differentiated?

It's tougher than it seems, since anything generic gets copied for free the moment a model provider or an open-source project ships something like it. The only layer a fund can defend is the one built on something it alone has: proprietary data, a genuine point of view, or the codified judgment of an expert who has actually done the work.

The codified expert matters more than people realize. There is a difference between collaborating with an expert on a static playbook and productizing what that expert knows. A playbook is a fish. A founder reads it once, it goes out of date, and everyone gets the same frozen version. An AI tool built on an expert's frameworks is a fishing rod. It looks at the founder's actual work, tells them where it is weak, and coaches them over time. It scales an expert's knowledge across the whole portfolio without their calendar ever becoming the bottleneck.

Turning expertise into something founders can run, not just read, is what we are rebuilding 645's value-add playbook around. We have always been a firm that builds, as a product and engineering culture has powered and differentiated how we invest from the start. Pointing that same muscle at how we support our founders is a natural next step, purpose-built for the age of AI. The goal is simple: put sharper tools in a founder's toolkit to give them real leverage as they scale.

An example is a founder thought leadership skill that our partner, Parul Singh, has built and tested internally with our own team. She has spent over a decade as an early-stage investor and has seen firsthand the power of storytelling for founders, having built an advisory practice helping founders tell their own stories publicly. The skill helps founders find their voice, develop a strategy, and publish with consistency instead of going quiet whenever a busy week hits, which happens often for the founders we work with.

Another is a customer-centric AI advisor developed by Jeanette Mellinger, who already works with our founders on positioning, ICP clarity, and conversation readiness. She built research teams at Uber Eats and BetterUp, holds a Stanford Master's in psychology, and teaches and advises founders through early-stage funds and Harvard Business School. Founders can pressure-test their ICP, positioning, and customer conversations against the advisor, and get back sharp questions and real critique instead of a generic template, so they can better understand, reach, and build for their target customers.

The expertise is the moat. The tool is just how it travels.

The flip side of this discipline is knowing when not to build at all. If there is already a leader in the space, the honest move is to send founders there and add value on top, rather than ship a worse copy. We noticed that many of our founders use Clay, as it's one of the best GTM tools out there for finding the right customers and running sharper outreach. Rather than building a copycat version, we partnered with Clay to provide our portfolio companies with access to free credits and Clay's GTM accelerator, Wedge. We're also co-hosting a workshop with them to teach founders how to implement the platform into their GTM workflows. The tool is Clay's. Providing access to credits and tailored training is our advantage.

The rebuilt playbook is not a bigger toolbox. It is sharper judgment about which few tools deserve your name on them and can really move the needle. That is the bar we are holding ourselves to as we roll these out over the coming months. If we clear that bar, we build it. If we don't, we partner with whoever already has since our job is to equip our founders with the best, not to pad our own scoreboard.

For years, a fund's value was measured by opening doors and making intros for a founder. Our rebuilt playbook now adds an additional layer that measures what a founder can learn and execute on their own. Get the tool right and it almost fades into the background; what is left is a founder who is sharper than when they started. That was always the goal, and now we can finally build for it.