In the rapidly evolving landscape of enterprise technology, the “Office of the CFO” remains a paradoxical battleground. While the market appears saturated with hundreds of solutions, a significant portion of the global financial infrastructure still relies on aging, fragmented systems that struggle to meet the demands of a modern, data-driven economy. Today, we delve into the nuances of enterprise SaaS with a veteran specialist to discuss how the next generation of AI-native applications is finally unseating entrenched incumbents. This conversation explores the strategic maneuvers necessary to compete with industry giants, the limitations of artificial intelligence in high-stakes accounting, and why a founder’s ability to sell is just as critical as the code they write. We also touch upon the shifting dynamics of the investment landscape, from the importance of domain expertise to the resilience required to navigate the current exit environment.
Finance software is often viewed as a saturated market, yet many legacy systems still dominate the office of the CFO. Where do you see the most significant openings for new entrants today?
The reality is that while the market seems crowded, it is also incredibly fragmented and burdened by technical debt. When we mapped out the Office of the CFO, we identified more than 500 companies, but a staggering three-quarters of those are actually legacy players that are struggling to adapt to the modern cloud environment. What makes this sector particularly fertile for disruption is the unique position of the CFO; they aren’t just the gatekeepers of software budgets for the entire organization, they are also the primary buyers for their own internal tools. This effectively removes a layer of friction in the sales cycle because they are the direct beneficiaries of the efficiency gains. I see immense potential in both horizontal applications that serve a broad range of industries and highly specialized vertical solutions tailored for sectors like construction, manufacturing, and logistics. By focusing on these specific workflows, startups can offer a level of depth and integration that the generalist legacy systems simply cannot match.
Many legacy platforms are attempting to modernize by shifting to the cloud, but you’ve distinguished between these “SaaS 1.0” systems and true AI-native solutions. How does this distinction impact the replacement cycle?
We are seeing a profound shift where even companies that are only five to ten years old are being viewed as “legacy” because they were built for a pre-AI world. These SaaS 1.0 platforms were essentially digital filing cabinets that automated basic workflows, but they lack the underlying architecture to truly leverage generative intelligence. Modern buyers are now demanding AI-native products that don’t just store data, but actually understand, predict, and execute tasks on behalf of the user. This creates a massive opening for disruptors to target categories like sales tax, treasury management, and procurement, where ancient systems still require significant manual intervention. While some incumbents are trying to reinvent themselves, many are finding it nearly impossible to strip out their old code bases and replace them with something truly autonomous. As a result, the “new generation” is gaining ground by offering a level of speed and insight that makes the old cloud-native systems look as obsolete as on-premise servers.
In an environment where accuracy is non-negotiable, particularly in accounting and audits, how much autonomy should companies realistically grant to AI agents?
This is perhaps the most sensitive area of the entire financial stack because finance professionals are inherently risk-averse and require absolute consistency in their data. We have to be incredibly surgical about how we infuse AI into these products; it is excellent for summarizing contracts or predicting cash flow trends, but it is notoriously unreliable when it comes to raw mathematics. You absolutely cannot have an AI agent performing core accounting calculations because large language models are not designed for mathematical precision, and even a minor hallucination could lead to a catastrophic audit failure. Instead, the focus should be on using AI to manage complex workflows that don’t require precise, single-number outputs, while leaving the heavy-duty calculation to traditional, deterministic software engines. The goal is to build a “smart” solution that acts as a co-pilot, handling the tedious administrative work and allowing the human expert to focus on high-value strategic decision-making.
Disrupting core HR platforms like Workday or ADP is notoriously difficult due to the high cost of switching. What is the most effective strategy for a startup trying to gain a foothold in that space?
Trying to unseat the core payroll or benefits engine of a global giant is a Herculean task that often ends in frustration because those systems are so deeply integrated into the corporate “plumbing.” A much smarter approach is to attack the secondary products—the features that these large platforms offer but don’t necessarily treat as their “core” business. Companies like Workday or ADP have massive distribution networks, which often allows them to win deals even if their product isn’t the best in every single category. However, a specialist player that focuses exclusively on a niche like workforce management or specialized employee benefits can often build a product that is lightyears ahead of a suite-based alternative. For example, by being an AI-native specialist in a field where the incumbent is merely an on-premise solution that migrated to the cloud, a startup can demonstrate enough value to justify the “surgical” replacement of that specific module. Once you are inside the organization and proving your worth, you can expand your footprint over time.
You have a unique requirement for founders, often referred to as the “Sales Test,” which you perform before making an investment. Why is sales proficiency such a non-negotiable trait for you, even for highly technical CEOs?
It is a common misconception in the tech world that a great product will eventually sell itself, but in my experience, the CEO must be the primary evangelist for the vision. Nearly every founder I back comes from a deep engineering or product background, but if they cannot effectively communicate their value proposition to customers, partners, and even potential employees, the company will struggle to scale. I make it a point to personally attend sales calls that I’ve set up for the CEO, just to watch how they handle objections and build rapport in real-time. If I see a consistent pattern where prospective clients lose interest or aren’t eager for a second meeting, it’s a massive red flag that the founder hasn’t mastered the art of the “sell.” You need that “founder-led sales” energy to navigate the early days of a startup, because no one else will ever care as much about the mission as the person who started it.
Given that the IPO market has been somewhat volatile recently, how does the current exit environment influence your decision to fund early-stage startups?
Our focus is primarily on Seed and Series A investments, which means we are looking at a time horizon that is often five to ten years out, so we can’t afford to be overly reactive to the quarterly fluctuations of the public markets. Our firm currently manages about $15.5 billion and we are actively investing out of a $3 billion vehicle, so we have the capital and the patience to wait for the right window. While every investor dreams of their portfolio companies going public, the reality is that the vast majority of successful outcomes happen through M&A. We prefer to enter early so that even an acquisition for less than $1 billion represents a fantastic return for everyone involved. If you enter too late at a multi-billion dollar valuation, your exit options narrow significantly because very few strategic buyers have the budget for a deal of that magnitude. We back people with deep domain expertise who are building durable businesses, and we trust that if the execution is there, the exit will follow.
You’ve curated a “Failure Museum” with over 1,500 items from defunct companies. What is the most profound lesson you’ve learned from studying these corporate collapses?
The museum is actually a source of great optimism for me because it serves as a physical reminder that failure is not the end of the road, but a necessary part of the innovation cycle. When you look at 1,500 different products and companies that didn’t make it, you start to see that the most common cause of death isn’t a lack of talent or even a lack of funding—it’s often a lack of market timing or a failure to adapt to a shifting landscape. People shouldn’t be afraid to take massive risks, because even the most spectacular failures provide the “scar tissue” and lessons that lead to the next great success. I study these items to understand where things went wrong, but also to appreciate the courage it took for those entrepreneurs to try in the first place. It keeps me grounded and reminds me that building a sustainable, long-term software company is one of the hardest things in the world to do.
What is your forecast for the enterprise software sector over the next few years?
I believe we are entering an era of “intelligent orchestration” where the silos between different business functions will finally begin to dissolve. Over the next two to three years, we will see the rise of vertical AI applications that don’t just sit on top of existing data but proactively manage entire supply chains and financial audits with minimal human oversight. Companies that rely on the old “SaaS 1.0” model will face an existential crisis as AI-native competitors offer 10x improvements in both speed and cost. We will also see a resurgence in the importance of specialized domain expertise; the “generalist” software era is ending, and the winners will be those who can solve highly specific, complex problems for specific industries. Ultimately, the winners won’t just be the ones with the best algorithms, but the ones who can build deep trust with risk-averse executives by delivering consistent, accurate, and transformative results.
