The AI application age: Meet functional analysts, the new vertical SaaS
In January, I wrote that 2024 would be the advent of the AI application age, with a new generation of AI-native “analysts,” and companies built around them, emerging to mechanize cognitive tasks. I talked about why AI-native application software represents a tremendous market opportunity, particularly for its potential to penetrate the massive global services sector.
That revolution is starting in what many are calling “vertical AI.” Going beyond traditional vertical SaaS, vertical AI companies like Harvey in legal and Hippocratic AI in healthcare made headlines in 2023 by developing LLM-based products for industries long under-penetrated by software. In the time since, a spate of competitors have emerged in these sectors, and new categories have opened up as ripe for disruption using LLMs.
In this post, I’ll dig into this brave new world of “vertical AI.” What are the defining characteristics of these new startups? Where are they next poised to emerge?
From Vertical SaaS to Functional Analysts
Today’s vertical AI companies follow in the tradition of vertical software, which is distinguished by sector-specific product offerings and workflows. In the past ~decade, vertical SaaS has emerged as a formidable software category, with the acceleration of businesses like Shopify (ecommerce - ~$95B market cap), Toast (hospitality - ~$13B market cap), and ServiceTitan (home services - last valued ~$10B).
To date, the opportunity of vertical SaaS has been that “software is eating the world.” Every industry, no matter how analog—and perhaps especially those that are—has unique business needs that can be better addressed by software than pen and paper. Indeed, some of these enterprise needs were actually net new in the Internet era, like the importance of an online storefront and seamless checkout experience that facilitated the rise of Shopify. The trick was to make these vertical SaaS businesses big enough for “venture scale,” e.g. $100M+ in annual revenue. There are familiar adages about how vertical SaaS businesses succeed, including the importance of market leadership — lower TAMs mean market penetration is key for scaling revenue— and expanding product offerings to counteract sometimes low ACVs.
In contrast, the value proposition of vertical AI is that AI is eating software and services. The new generation of vertical AI startups are largely functional rather than vertical. That is, these enterprises automate business functions rather than providing industry-specific software. Note that they can also often be “vertical” in that their products may be particularly targeted towards or useful in certain enterprise contexts. The incumbents in these categories often aren’t legacy software, but rather outsourced service providers, either professional services firms or BPO providers.
Market opportunities
You can imagine AI-native vertical startups as analytical horsepower for all kinds of cognitive work that happens in the enterprise, including audit and accounting, compliance, legal, procurement, and other administrative workflows. See my mapping of particularly interesting categories below.
Regardless of the functional work they are automating, vertical AI startups need the following to succeed:
Data advantage. Startups leveraging proprietary, industry-specific datasets have a big leg up in providing value-add vertical AI products. Not every startup will begin with a large data corpus (or need one to start), but it will be key to improve vertical product offerings with customer data generated from product use as products scale.
Domain expertise. Don’t underestimate the importance of understanding your end market. Context-specific workflows are key to successful, sticky vertical products. As others have said: “Come for the model, stay for the workflow.”
Manual, costly processes. Enterprises are interested in AI trials right now, but stickiness will demand either topline or bottom line gains. Startups should look for areas where there is opportunity to automate manual, costly processes. My guess is that vertical AI offerings may be particularly resonant in low-margin industries like healthcare services, where automating manual and labor-intensive processes can drive highly sought-after efficiencies.
I’m particularly excited about internally-facing enterprise applications. External LLM-based applications are higher-risk and harder to implement, for a few reasons. First, there are brand safety concerns, as LLMs have the tendency to hallucinate; LLM evaluation and observability tools are nascent, amplifying concerns around #1 and making it harder to put LLMs in front of customers; and differentiated proprietary data sources may be harder to come by.
Instead, internal applications allow enterprises to leverage available internal data and experiment in a lower-stakes context, while still demonstrating alignment with top-down “GenAI mandates” from CEOs. Internal LLM applications also have the added benefit of promising efficiency gains—particularly compelling as enterprises increasingly prioritize profitability in a lower growth environment (see Meritech’s latest report documenting forward-looking revenue growth and FCF margins).
Watch-outs and open questions
It’s a brave new world ahead. For the startups tackling vertical AI, a few watch-outs and open questions to navigate:
Inference costs: A particular challenge for products built on top of expensive models like GPT-4, but inference costs will affect unit economics (and capital needs) for most app-layer startups. Experiment with pricing and with your tech stack, e.g. by incorporating fine-tuned OSS models.
Find the “hair on fire” problems: Top-down “genAI mandates” are an opportunity, but also a potential trap. To succeed over the long term, startups need to get away from enterprise experimentation budget and find “hair on fire” problems within the enterprise.
Get ready to run. Execution speed will be critical in these hotly-contested categories. Don’t underestimate the power of simply being faster and building and executing quickly.
If you’re building in vertical AI, I’m excited to chat — as always, you can find me at molly@radical.vc.