The Impact of Vertical AI on Industries

I've been thinking about vertical SaaS lately. From 2018–21, only 24 % of 80 software IPOs were vertical SaaS. Why? Smaller customer pools and limited value capture kept the upside capped. AI changes the math. Instead of putting clipboards in the cloud, AI does the work itself—and that rewrites three fundamentals: 1. Value | Outputs, not clicks - Pre-AI apps sped up human workflows. - AI-native apps ship the deliverable—draft the brief, reconcile the invoice, triage the patient. When software does the work, it earns a bigger share of the value created. 2. Pricing | Usage, not seats - Seat licenses mapped to headcount. - AI teammates meter documents, calls, or tasks. 3. TAM | Core industry spend, not IT budget - Old ceilings: field-service software ≈ $5.5 B, restaurant POS ≈ $12 B, construction management ≈ $10 B. - New horizon: legal services alone top $1 T. When software augments the lawyer’s, nurse’s, or analyst’s job, it taps the services budget—not just the software line item. Takeaway: Bigger value → usage-based pricing → 100× larger markets. Bonus for founders: Many of these opportunities are untapped.

Most biopharma providers we’ve spoken to spend hours sifting through papers, patents and clinical trials, hoping to uncover commercial opportunities. Here’s the problem I see with that: > Humans process research linearly i.e., reading each paper in full to extract insights. > AI processes research contextually i.e., analyzing thousands of papers in seconds to surface the most relevant findings. Here’s why AI is changing the game for business development teams in life sciences: 1/ AI identifies patterns across thousands of documents > Humans can read a handful of papers a day. AI can analyze millions. > It recognizes recurring keywords, experimental techniques, and funding trends across vast datasets. > This means less manual review, more actionable insights. 2/ AI understands commercial relevance, not just science > AI doesn’t just summarize, it prioritizes findings based on business impact. > It can surface research linked to clinical-stage companies, industry collaborations, and commercial applications. > Instead of scanning endless publications, BD teams get a filtered list of high-value prospects. 3/ AI tracks emerging research in real-time > Manual research is static, AI research is continuous. > AI flags newly published papers, active trials, and emerging patents relevant to your business. > This means your team sees opportunities before competitors do. 4/ AI cross-references multiple sources > A BD rep might read a single paper and miss its connection to industry movements. > AI links clinical trials, patents, and publications to map the full competitive landscape. > This is how leading biotech firms identify rising players before they make headlines. Manual research is slow and reactive. AI is fast and predictive. The teams leveraging AI-powered research aren’t replacing their scientists, they’re making them exponentially more effective.

I had lunch with a founder last week who pitched me on their "AI for operations" platform. I stopped them 3 slides in. General-purpose AI isn’t cutting it anymore. DeepSeek’s January breakthrough told us something important: efficiency & performance can coexist a lot earlier than most people thought. Startups are now excelling not by scale but by focus: they’re building vertical AI that deeply understands the messy, high-stakes workflows in sectors like healthcare, finance, and defense. Specialization is the new competitive advantage. 3 patterns I’m tracking across successful vertical AI startups: First, they pick massive but high-friction and high-value workflows. “AI for sales” or “AI for operations” is too broad. What’s effective is focusing on urgent, complex processes, like: ConverzAI streamlining high-volume recruiting for staffing agencies Tennr automating messy admin work Second, they build more than model wrappers. They create proprietary feedback loops and data assets that compound over time. This instrumentation is what turns a one-off tool into a durable, defensible product. Third, they expand from beachheads of earned trust. They wedge into multi-billion-dollar industries by solving problems in the hardest, least glamorous corners. From there they earn the right to expand and unlock bigger TAM over time. Choose one gnarly high-value workflow and go deep. Otherwise you might get stopped three slides in too.

Vertical AI is where the real money is—and where the next wave of AI unicorns will emerge. Enterprise AI adoption has been painfully slow. Why? Because large organizations are drowning in siloed, incomplete, and messy data. Their tech stacks are bloated, legacy systems don’t talk to each other, and AI adoption often requires expensive and time-consuming integrations. On top of that, most companies don’t have flexible budgets to experiment with broad AI models that promise automation but fail to deliver tangible results in the short term. This is why the future isn’t in general-purpose AI. The future is in Vertical AI—deeply specialized AI companies solving high-value problems in industries like healthcare, finance, and defense. These startups don’t just throw generic LLMs at problems; they build with industry expertise, leverage proprietary datasets, and create solutions with real-world utility. The result? Faster ROI, seamless adoption, and massive competitive advantages. The next AI giants won’t be chasing broad intelligence. They’ll be laser-focused, solving the hardest problems in trillion-dollar industries. The winners won’t just be using AI—they’ll be owning the data that powers it. 🚀

This is a solid article by Activant Capital on how vertical AI agents are impacting organizations. It came out in Nov 2024, which is already old by AI standards (sigh), but the thinking still holds up. A few things that stood out to me: ⛳ The biggest shift lies in thinking of vertical AI agents as managers (not assistants). Historically, AI just helped coordinate or surface insights. Now, agents can own workflows, make decisions, and drive outcomes, like an actual ops or project manager. ⛳ Their effectiveness hinges on two things: how capable the agent is, and how well it integrates with third-party tools. This is exactly why something like MCP becomes critical. Tool access and interoperability are key reqs for more autonomous systems ⛳ Banking, insurance, software, and capital markets are flagged as the most automation-ready. Healthcare, legal, and financial services are primed for bigger disruption, massive markets, tons of repetitive human input, and structured data. ⛳ What surprised me most is that they report that there's good progress in heavily regulated sectors too! Just a few years ago, most people thought AI adoption there would lag. But that’s changing fast, even with all the compliance, privacy, and security challenges. Definitely worth reading if you're tracking how vertical agents are evolving. Link: https://lnkd.in/eYSWqgrs

Over the last year, we've seen AI deliver some clever "slingshots" or wedges into vertical SaaS - the key thing to becoming David with his magical slingshot and effectively taking down the existing Goliaths, will be building a system of record and the workflows around it. The "slingshots" we're typically seeing look like: 1) Voice - e.g. assisting in screening inbound and generating outbound calling 2) Transcription - e.g. helping record notes 3) Workflow automation - e.g. filling out forms and reports, document processing 4) Lead generation - e.g. creating emails and other messages, marketing campaign content All of these are simple, add- on tools - easy to adopt by a business, familiar work with minimal onboarding/training needed. And often 100% accuracy isn't always required (or can be checked). They are magical to their customers as they assist or even replace human labor. We're seeing these AI companies grow quickly - it's like magic to their customers and everyone wants to try AI! These slingshots are vertical-specific - trained on vertical data (e.g. legal precedent, accounting rules) and/or customized workflows and integrations (e.g. into the TMS or EHR ecosystem). Their customers don't have the resources to take horizontal AI tools to build and maintain custom software. We've seen a number of startups begin with a horizontal approach across industries, end up verticalizing here. But they can have a big, positive impact on their customers' P&L especially for: 1) an SMB where the labor cost would have been and felt directly by the owner (e.g. a doctor's office) 2) industries where there are lots of people employed or BPOs used (e.g. freight brokerage, insurance claims processing). So what does it take for these slingshots to be well-aimed and turn into defensible "operating systems"? They are coming up against ingrained Goliath systems with broad feature sets. That said, they are collecting *ground truth data* on the business. This is the underlying data on customers, their interactions and preferences, performance of the business - the data that employees often used to manually key into an ERP or CRM, then extract to understand the business. It’s all automatically available, in real-time ideally! With the right workflows from there, the AI wedge + the data enable these new companies to compete.