AI Unlocks New Super Powers for Founders

It's time to reconsider much of early stage VC conventional wisdom and playbooks

Over the last several months, I have been talking extensively with founders, GPs, and LPs about the seismic and mostly exciting changes that AI represents for the next generation of software products, founders, and investors who support them.

These sessions have varied widely:

- a 4-hour detailed AI product roadmap discussion with an existing $75M portfolio company around how they can potentially unlock a $1B opportunity with the right future AI product moves

- a series of deep heart-to-hearts with a native-AI portfolio company that has grown from $0 to $6M in 12 months in a market segment that has at least $1B of capital invested in other AI competitors

- a powerful push to a founder to spend all of his marketing dollars as fast as possible now that his company is seeing explosive and efficient growth since recasting their product with an AI assistant just 9 months ago - i.e., "why can't you grow 5 to 10x next year?"

- a series of in-depth mutual discovery sessions with an early AI vertical software company on how we all need to rethink the traditional product development process as the new winning standard seems to be starting first with discussing a detailed pain / use case with a prospect/customer and then focusing like crazy to see if you can solve for that … in a few weeks or so.

- an endless number of prospective AI-centric founder meetings every week where, for the most part, I see solutions that are far better than the drek on the market today. I wrestle through understanding "Does this one have legs?" especially in an over-inflated valuation market for AI companies when every new software investment in a year will be an AI company

- a 30-minute session with our existing and prospective LPs on the impact of AI on the overall industry and our perspective on why that's exciting for application software companies (because I also wrote on this same sub-stack that 75% of customer-facing app companies go away)

I still believe this is indeed a do-or-die situation—i.e., to be (AI) or not to be. But instead of doom-casting, let me offer up why I think this is the most exciting time to be an early-stage software start-up founder (if you are great), despite the opportunity for legacy players with deep meta-data to recast a likely clunky UI experience with AI and extend their shelf life.

Aside from not having to manage much of the brain-killing minutia of a SaaS company directly tied to the fact that your product interface sucks (i.e., onboarding and adoption, obsessing over the suitable screens that take months to research, test, and deliver, etc.), here are four significant new opportunities for AI-centric companies that just didn't exist for 99.9% of pre-AI SaaS companies and, as such, how investors need to reconsider their traditional playbooks for investing and supporting their portfolio companies:

Superpower 1: The Ability to Penetrate Brand New Markets / Tap New Spend
AI is suddenly making software viable for buyers who previously lacked the time, technical knowledge, or budget to consider it. We're seeing users within industries adopt software for the first time, pulling from payroll budgets rather than IT spending. AI isn't just expanding existing markets - it's creating entirely new ones. Intelligent founders can also effectively target previously unprofitable segments to serve or too complex to reach, especially in the VSB market. The most exciting part? These new customers often have a higher willingness to pay since the ROI is immediately apparent in headcount savings. We're seeing deals sized at multiples of traditional software spending because the value proposition is tied directly to FTE costs or realized FTE productivity gains versus promised ones.

VC Playbook adjustment # 1 - The size of the TAM for an investment is not the size of the existing software spend today with projected growth forward. It's a good floor but misses the point that AI unlocks massive potential additional spending for a software product in three general ways: 1) more users within a target company will use the software, 2) new segments of target customers now may use software for the first time and 3) the premiums customers are willing to pay can be much higher, especially if they are funded from the payroll budget rather than tools one.

VC Playbook adjustment #2 - Investors must revisit the traditional go-to-market economics of serving specific customer segments, especially when targeting unsophisticated users and VSB segments. Historically, these are incredibly hard customers to efficiently target as the cost to acquire and serve is not worth the squeeze vis a vis their average dollar spend & retention. With products that effectively abstract away the "software" hurdle for these segments, we see companies effectively target these customers in an outbound manner with <300/month price points delivering 3x growth with magic numbers >1.

Superpower 2: The Ability to Land New Customers Easier
AI solutions can provide innovation for specific use cases that immediately blow customers away with a >10x improvement over their current approach.   It is not traditionally easy to either a) rip and replace a product from a customer or b) have a customer buy software for the first time to solve a given pain. But, if you have a product that now creates an "oh holy $$$t, that is too good to be true" moment with your prospects, we're seeing founders driving much faster initial ARR growth curves out of the gate with just founder selling and doubling down on acquisition efforts earlier than usual. The market rewards those who can aggressively grab market share when they have a shocking product. The good news is that it's much easier to scale a first principles GTM team when your AI-centric application 1) is incredibly easy to use and 2) is incredibly easy to understand the specific jobs to be done you solve for customers.

VC Playbook adjustment # 3 - Investors and founders need to think about products differently - you can not start with a wedge, sell on vision, and deliver a terrific product after series A when you finally have a well-staffed R&D team and the time to deliver greatness. Your product as a founder solves very well with high precision and accuracy for "enough jobs to be done" for a customer or it does not - it's a pass/fail value proposition.. As such, your software product's quality/value bar has just gone up 5-10x from customers - even at the seed stage. We're pushing founders to think bigger about product scope from day one. The old wisdom of nailing a narrow use case before expanding doesn't hold up when customers gravitate toward fewer, more capable AI assistants. Intelligent founders are exploring their second and third use cases before fully optimizing their first. The technical lift for expanding solution breadth is far lower with AI interfaces, making it possible to capture more customer workflows quickly. As such, that may mean it's best to advise founders to make early land grabs in adjacent use cases rather than perfecting a narrow wedge.

VC Playbook adjustment # 4 - When a founder has a "holy $#!t" product, we encourage more aggressive, bursty customer acquisition spending earlier than normal. The goal is to capture market share rapidly when you have a clear breakthrough product and to prevent competitors from planting their "first good enough agentic product into your customer". These don't tend to be as reckless investments as prior generation spends, as the feedback loop on pay-offs is much faster and can be adjusted up or down as needed. It also likely means companies must think through repeatable approaches earlier to ensure additional FTE capacity adds become productive faster.

Superpower 3: The Ability to Consolidate Your Customer’s Tech Stack Faster
In my prior predictions, I underestimated how fast customers would embrace consolidating their technical stack. Customers are gravitating to fewer vendors, particularly seeking one primary AI assistant or set of agents for each functional role. The technical lift to expand that solution breadth is dramatically lower than ever if a founder understands the "jobs to be done" and can quickly build the context such that they solve for it within their existing agentic interface. This consolidation trend creates a "winner take more" dynamic in many categories. The old notion of best-of-breed point solutions is giving way to integrated AI assistants and agents that can handle a broader scope of work - because, as I have shouted before, that is what each user within a function wants.

VC Playbook adjustment # 5 - Historically, patient investors would guide founders to focus religiously on problem one / product one. By focusing on that use case, a founder could devote their limited resources to make that product as great as possible and not distract the rest of their GTM team from having to focus on too much at the same time. The company could focus on product/use case 2 after they were well north of $10M in ARR. That is no longer the case - founders should work closely with their customers to understand what additional "jobs to be done" customers care about and look to deliver them - with the caveat of ensuring you have solved properly for the prior JTBD you signed up for.

VC Playbook adjustment # 6 - For a horizontal software offering, early-stage start-ups will often start by targeting that offering to an ICP within 1-2 target industry verticals to drive much more effective and aligned results from a small GTM team. This approach makes complete sense - unless, by the time you pursue industry 3, 4, n, a competitor has sold to them with a pretty good AI solution. Your product may be better, but given how transformational your competitor's offering was to the crap that customers had before, it could be much harder to displace than you think. As such, founders and investors likely need to "noodle longer" over when it makes sense to spread your gaze to additional segments, especially given the high likelihood that vertical-specific vendors will look to leverage the AI advantage to include your horizontal capability as fast as possible.

Superpower 4: The Ability to Build More Capital Efficient Businesses
AI is rewriting the rules for what it takes to build and scale. Valuable applications (metadata, context, etc) with agentic interfaces require far less R&D resources and time to develop and maintain. Internal AI use is turbocharging everything from marketing to product feedback loops. We're seeing companies achieve magic numbers above one while serving traditionally tricky segments. The proportional spend across functions looks completely different - customer success teams can support more accounts, marketing can generate more qualified leads with smaller teams, product development cycles are much shorter, etc. For founders who get this right, the yield on capital can be multiples higher than traditional SaaS models. Most importantly, it enables companies to move faster and respond to market opportunities more quickly than ever before.

VC Playbook adjustment # 7 - There is a relatively tried and true recipe for how much founders should spend within each of their organizations at each phase of their growth. Similarly, within each functional organization, there are general rules of thumb on which roles founders need and how many of them in relation to each other. It is time to rethink these guidelines for two primary reasons: 1) Agentic products with far less screen real-estate change the dynamic of what's required to build valuable products as well as all of the upstream and downstream resources required to sell and support them and 2) The use of AI internally for your employees dramatically increases their productivity and removes the need for many support / junior roles. Finally, I suspect that compensation ranges will change for top-flight employees. Given that founders will need fewer employees for a given function, they may pay far more for a 5-10X employee than before.

VC Playbook adjustment # 8 - Given that great founders who leverage AI properly and flex the superpowers mentioned above can likely achieve much higher ARR results per amount of capital invested, investors would be wise to work with their founders to ignore the "what I need to do to raise at A, then at B, then at C, etc" exercise and start first with how fast realistically can we grow this business within 10 years (aligned of course to the founder's rate of learning and growth) and how much venture capital do they need to raise. As part of that, identify the "business model hacks" where the founder spends far less money in certain areas than pre-AI traditional SaaS companies and what best to do with those savings. There are three new emerging conversations I see here - 1) we can skip a round - i.e., go from Seed to Series B, and/or 2) we don't need late-stage capital - i.e., beyond a Series C, there is no reason to raise beyond that and/or 3) inversely, we have such crazy market fit and it's an absolute land grab and, as such, I want to raise far more capital than the traditional norm because I know exactly how to put it to use and I want to suck the air out of the room from my competitors.

---

From an investor perspective, we, like many early-stage venture investors, are significantly tightening our screening for new investments, especially regarding the founders we back.

To exercise the superpowers above, the founders need to be, well, super.

More than anything, in this new era, LACK OF SPEED KILLS.  

Founders and their teams must move incredibly fast and leverage the power of AI and customer intimacy to gather, process, direct, and act on an almost non-stop continuous process. The opportunity to create massive value has never been greater, but it requires rethinking conventional wisdom about go-to-market, product development, and company building. As an example, here is how we have further adjusted our investment screen - hopefully helpful to those pre-seed founders out there who wonder how we and other firms evaluate them differently today.

Products that Make You Jump out of Your Seat
We focus on finding true 5-10x solution leaps - anything less isn't worth our time in this market. We need to see solutions that fundamentally upend how target customers get their "jobs to be done" done, delivering massive existing FTE productivity gains and potentially long-term FTE cost reductions immediately apparent to customers. The product must create genuine "holy $#!t" moments where prospects can't believe how much better this is than their current approach. It is a pass if we do not see that level of breakthrough impact in initial customer conversations. Yes, even at the seed stage.

The Metadata Moat Really Matters
We're instantly passing on pure LLM wrapper plays and most horizontal solutions - the value can't evaporate with the next OpenAI release or something that can be built just as well internally with an Assistant or Agent building platform. Winners must demonstrate they're creating real defensible "context" positions with a combination of metadata, ideally supplemented with unique algorithms or proprietary datasets. We're particularly bullish on vertical software plays since AI performs best when solving a specific user's complete needs with deep contextual understanding—the more industry-specific the data and workflow knowledge, the stronger the defensive position.

Somewhat Maniacally Obsessed Founders
The ZIRP-era work-life balance mindset is a complete non-starter. We need to see founders and their teams moving at unprecedented speed to match deep customer intimacy with their underlying AI infrastructure to deliver contextual fit better than others. They must be ready to make tough calls about aggressive land grabs and demonstrate almost religious zeal on building shocking, not just viable, products. Great founders and the early team that they have assembled must #$%ing crank. At speed, they must out-understand their customers' workflows, out-ship meaningful product improvements, and out-attract and onboard assemble world-class talent. For the first 20-30 core team members, this simply cannot be done in a remote-first world. The iteration loops on the business and the product needs to be so fast to emerge as a winner.

Ultimately, the bar for what constitutes a compelling investment has risen dramatically. We say "no" far more quickly to companies that would have been somewhat interesting 18 months ago. But when we see that rare combination of shocking product innovation, defensible metadata advantages, and maniacally focused founders, we're ready to move aggressively with conviction. LFG.