skip to Main Content
Back to Insights

For the last two decades, vertical software has sat in the shadow of horizontal software. Despite some fantastic success stories like Veeva, Procore, and Toast, the SaaS and cloud evolution has been written by leaders in marketing, HR, and finance. Vertical software has been largely dismissed by the venture community as a difficult play due to smaller market sizes and a more difficult go-to-market motion, credited to heavily entrenched incumbents and a resistance to new technology. I think the reason is much simpler: solutions weren’t providing enough value, so they didn’t succeed. 

It’s hard to estimate what percentage of vertical software went cloud, but numbers range as low as 25%. We have all looked over the shoulder of our doctor and saw a glimpse of software reminiscent of DOS or early Windows. The layman argument has been that vertical companies are not technology adopters. The hard truth is that they were early adopters to technology, and the modern iterations did not provide enough value to upgrade. 

AI is changing that. The new waves of vertical AI companies can differentiate through intelligence and automation rather than just features or usability. Many of the vertical markets are poised to skip the plain-cloud evolution and jump straight into the AI revolution. The return on investment will be compelling enough to justify the organizational cost required to make that change, so much so that we anticipate vertical customers being first-movers over general other markets. 

For startups, the “difficulties” inherent to vertical markets that impeded the cloud revolution—regulatory constraints, large volumes of domain-specific data, heavily entrenched incumbents—will become exploitable, creating moats against the threat of generalized AI tooling to which horizontal SaaS is vulnerable. That, along with the expanded market size afforded by eating labor vs. software spend, make the opportunities in vertical startups exceed those in horizontals. The next decade of software will be about vertical software hinging on AI. 

Sizing the opportunity

The most profound impact of AI may be the market expanding nature. The common critique of vertical software is that the markets are too small to build massive outcomes. This idea is deeply rooted in the pricing model of traditional SaaS. Sold on a per seat model, the software buyers calculated their spend as a percentage of an employee’s salary. The cost of Salesforce might be 1% to 5% of an employee, and aggregate spend per employee might reach 10%, but generally not much more than that. This is because traditional SaaS just optimizes workflows. Tasks that were initially on paper were moved to software, then moved to the cloud. The value captured was making humans more efficient at completing tasks and better at making decisions. 

With AI, companies can automate tasks traditionally performed by human experts, capturing a greater share of their customers’ value chains. For example, instead of merely offering software to manage claims, an insurance-focused platform could automate claim approvals or fraud detection. This change in delivery of value allows the vertical software vendor to price based on value rather than relative to employee salary, and the net result is profound. We have seen examples of where the vertical AI vendor is capturing 25% or 50% or more of an employee salary. This would suggest vertical markets will be five to ten times larger through the introduction of artificial intelligence.

Capturing the value of that labor makes markets that would have previously been “too small” venture-viable and creates new billion-dollar markets. This TAM expansion means that subfractions of markets are now in play. Legal software, for example, used to be one category. Now we’re seeing software just for accident lawyers. The market can sustain a lot more winners, and founders and VCs can have more at-bats.

Why verticals will win

One of the risks of AI investing is that a general purpose LLM will be good enough to impede adoption of more specific tooling. That risk is not present with (most) verticals, making the markets a potentially safer bet, especially for early stage startups. Verticals have a few “unfair advantages” over horizontal solutions inherent to their markets. We’ve seen evidence of this through the success of long-term incumbents who have leveraged the limitations of their markets as a moat against more generalized horizontal tools (in other words, all of the reasons cloud software struggled). 

Unstructured data: The capabilities of large language models (LLMs) are uniquely suited to vertical markets, which often operate with vast volumes of domain-specific data. In industries like healthcare, legal services, finance, and manufacturing, the ability to parse, analyze, and derive insights from unstructured or highly regulated data has historically been limited by the inflexibility of horizontal software solutions. Data management is the cornerstone of most vertical software, and data migration is one of the most significant costs of switching solutions, so an AI platform that could manage this work would be hard to unseat.

Domain-specificity: Building AI applications in vertical environments will be a very specific skillset. Broad-based foundation models may lack the context required for the tailored decision-making, or the use of these tools inherently violates the regulatory frameworks present. Using OpenAI to make a loan decision is not possible today, since it cannot provide an audit on bias for regulators, even if you can solve the data regulations that come with handling personally identifiable information. Domain knowledge can build with bias, fairness, and logic of decision-making.

Regulations: Regulatory environments are complex and constraining in vertical markets. HIPPA in healthcare, compliance in trucking, OSHA in construction, and the McCarran-Ferguson Act in insurance are just a few of the examples that have to be navigated by startups. Generic or horizontal software often lacks the specificity to meet these demands, leaving businesses exposed to risks like legal penalties, data breaches, or non-compliance. 

What makes successful vertical AI companies

I’ve always held that in vertical software innovation cannot just be on technology or product, but that it needs to be bundled with a disruptive go-to-market approach. Artificial intelligence does not change this, and if anything, provides a moment where these lessons are that much more relevant and interesting, as the technology wedge is particularly clear right now.

For the next wave of vertical giants, we look to the playbooks that have worked in the past. Founders and fans of vertical SaaS may dismiss learning about these legacy (in many cases pre-cloud) vertical giants as a wasteful look into the past when the future, driven by artificial intelligence, is right in front of our eyes. I dispute that mentality vehemently and think the answers are within these companies, their strategies, and the lessons to take away. Here are just a few examples and what AI startups can learn from them:

Get in on the ground floor: How AI will impact the future of industry is top of mind everywhere. We think there may be an opportunity to get in on the ground floor by using educational pricing to capture college students before they enter the workforce. Casetext is an interesting example in legal, where three vendors of historical case software have dominated the market for decades. Those vendors are successful because every legal student is trained on them before they graduate college (a tactic used in AEC products like Autocad and Esri as well). They also have network effects, a trait so persistent it should almost be a requirement in vertical software. We don’t ultimately know why Casetext chose to sell itself to one of these three incumbents, but I have no doubt the board understood the complexity of distribution despite a revolutionary product.

Hunt the white whale: If you can deliver a new experience that previously was not available, try to understand how Checkfree “slayed the dragon” and went after the biggest customer in the industry to make themselves the standard, and eventually the network effects eliminated any competition. Is that an approach you could take?

Find the right customer: Textura’s technology was not their ultimate innovation versus previous failed attempts, but rather their understanding that the monetizable customer in construction is the sub-contractor, and not the general contractor. Much like Checkfree, the early days were rough, but as they became embedded network effects became such that most general contractors found adopting the Textura technology most palatable, especially since they did not have to cover most of the cost.

Build toward where the winds are blowing: Almost every vertical industry is distinct because of the regulations that it operates within. We presume Epic to be an overnight success, and in some ways the acceleration of growth was over a short amount of time, but it was not the result of luck. Rather, it was a deep understanding of where regulatory trends were going and building a product that was able to meet those frameworks when the competition was stuck in the past. A day does not go by where regulation is not a topic in artificial intelligence, so which founders will understand those trends and build a product that can distribute when those frameworks become legislation? Or even today, if you understand that auditability of decisioning is a regulatory requirement in many areas of insurance you are one step ahead of any horizontal solutions.

Our attempt to study the past, to find these vertical giants and learn from them, is not a static activity. It is an ongoing and ever-expanding project as we find new companies that have stayed in the shadows. But also, it’s an exploration of what is to come, as companies in the current Scale (or other venture funds) portfolios will be those companies tomorrow. These patterns are intellectually interesting to us, but that, in itself, is not the reason for this pursuit. It is more fundamental and capitalistic, which is that we think that we can borrow from these lessons and increase chances of future success. 

We’d be remiss if we didn’t touch on one of the most important factors across all vertical success stories: founders who understand the industry but are willing to think about it differently. Shoaib Makani, founder of Motive, understood the trucking industry and resisted the common trend of building a marketplace to solve problems more directly. For the founders of Sixfold, they knew a blackbox solution wouldn’t work for insurance underwriting, both for regulatory reasons and because insurers didn’t want it. Instead, they built a copilot. These innovations might not be revolutionary ideas, but they are the right ones that skip over the “obvious answers” we often see from industry outsiders. 

We are excited for any and all founders that want to join us in this exploration and dialogue. 

Back To Top