AI is the Service: The Rise of Vertically Integrated AI Companies

At Forum Ventures AI Studio, we build B2B SaaS companies from the ground up. We've traditionally focused on pure software businesses, but we're particularly bullish on a new type of company emerging right now: vertically integrated AI companies.

These aren't software tools sold to service firms, they're full-stack businesses where software is the service, powering delivery from end to end. Think less "here's a tool for your team" and more "we are your team, powered by AI."

We're actively looking to launch more of these companies with founders who have ambitious visions and want to redefine how services are delivered. But to understand why this model is working now, it's worth looking at why so many companies failed the first time around.

Why the First Wave Failed

Over the past decade, some very smart and well funded teams tried to reinvent professional services by layering in software. The idea was simple - take a traditional service industry, like legal, accounting, or recruiting, be the most tech forward company in the industry and through the reduced amount of human labour needed to fulfill services you would deliver faster, better, and cheaper outcomes. The vision made sense. 

Atrium, the legal startup founded by Justin Kan, raised $75 million to do exactly that. Build internal tech to automate as much of the process as possible, then deliver services directly instead of selling software to legacy firms. If they won't adopt your tools, compete with them. Reduce the human headcount. Push margins higher. Turn a service business into something that looks and scales like SaaS.

But almost all of these companies struggled to make it work. Not because the ideas were flawed, or the teams weren't strong. They failed because the underlying technology wasn't good enough. The automation didn't deliver a 10x experience. Customers weren't getting better outcomes. And under the hood, these "tech-enabled" firms were still just teams of people doing the work manually.

This played out across vertical after vertical. Legal, accounting, HR, recruiting, anywhere the work was too complex or nuanced, the software fell short. Sure, it helped at the edges. But it didn't replace the core of the service. The bet was always that automation would eventually catch up. But that future never arrived, until now.

Why AI Works This Time

What's changed is that AI is finally at a point where it can meaningfully take on work that, just a few years ago, required a trained professional. We're no longer talking about automating scheduling or document filing, we're talking about AI models that can handle complex reasoning, analyze contracts, write code, manage communication workflows, and make high-quality decisions in real time.

We’re already seeing this shift in action.

Pilot is the clearest example. It’s not an accounting software tool, it’s a full-stack accounting firm. Pilot delivers end-to-end bookkeeping and CFO services with ~$43M in ARR and ~60% gross margins. By combining in-house automation with human oversight, they’ve redefined what modern finance operations look like for startups.

Crosby is doing the same for legal. It’s a full-stack legal services company built on proprietary LLM workflows. Instead of selling to law firms, they act as the legal partner for fast-growing startups, reviewing contracts, handling legal workflows, and doing it faster and cheaper than traditional firms. Their platform already supports companies like Cursor, and they’re backed by Sequoia.

Mercor is a full-stack recruiting firm. From sourcing to screening to payroll, Mercor automates the entire hiring stack for AI companies and startups. They’ve hit $75M in ARR, are growing 50% month over month, and serve elite clients like OpenAI. 

This changes everything. For the first time, the technology is actually good enough to take over entire workflows. And the gap between what's possible today and what will be possible three years from now is even more dramatic. We're not at the end state, we're just at the beginning of the curve.

The New Economics of AI Services

There are three big reasons this model actually works, and they have everything to do with how these businesses run.

1. Speed to Revenue.
You can get to revenue fast. These companies don’t require speculative new user behavior, you’re selling services people already buy. If you’re a lawyer, start selling legal services tomorrow. Fulfill the work manually at first, and build software in parallel to streamline delivery. You’re not guessing at product-market fit, you’re building with it from day one.

2. Higher Value Per Customer.
Traditional SaaS sells into software budgets, a narrow slice of spend. But vertically integrated services sell into labor budgets and deliver full operational outcomes. Instead of charging $99/month for a legal tech tool, these companies charge $1,000–$5,000/month to actually be the legal team. The result is higher lifetime value per customer and deeper integration into the customer’s operations.

3. Improving Margins Over Time.
As the AI improves, so does the business. What starts as human-heavy becomes software-led, less headcount, faster turnaround, better margins. In legal, accounting, compliance, and recruiting, teams are shrinking as models take over the repeatable work. The result is steady margin expansion over time, as more of the cost base shifts from people to software. These service businesses begin to look, operate, and scale more like software companies.

Forum’s Track Record

Forum isn't new to this model. Two years ago we launched Scout out of our venture studio. Scout combines AI-powered tooling with expert grant writers to help startups identify, write, and win SBIR funding. Instead of selling software to grant writers, we deliver the service ourselves with a human in the loop to ensure quality and troubleshoot when needed.

Scout is one of the fastest-growing companies Forum has seen and a prime example of what happens when you use AI to supercharge a service business.

Next Up: Where We’re Building

We’re actively looking to launch more of these companies with the next generation of founders. We're open to all vertically integrated AI opportunities and are particularly excited about legal and insurance.

AI-Native Law Firm

Context
The legal industry, especially in areas like research, intake, and document drafting, has long relied on high-cost, time-intensive workflows. But LLMs are changing that. Tasks that once required hours of manual effort can now be handled in seconds with AI, creating the foundation for a new generation of legal services. Traditional firms, constrained by billable hour models and high staffing numbers, are ill-equipped to compete with this shift particularly when it comes to serving startups and tech-forward companies that are looking for agile, AI-first legal solutions.

The Opportunity
We believe the next wave of legal firms will be built from the ground up with AI at their core. These companies won’t just support lawyers, they’ll replace and outperform them in specific workflows. From fixed-fee, fast-turnaround contract review to automated legal intake and compliance checks, there's a growing opportunity to deliver more affordable, scalable, and accessible legal services. We’re looking for founders who can bring deep domain insight and a bold vision to redefine what legal services look like in an AI-native world.

Forum Ventures is uniquely positioned to build in this space. With a portfolio of 400+ startups and a network of thousands of VCs, we have direct access to a large base of customers and partners who are already underserved by traditional firms and hungry for AI-driven solutions.

AI-Native Insurance Firm

Context
The insurance industry is bloated, outdated, and built on decades-old systems that can’t move fast or serve modern customers. Quotes take days. Claims drag on for weeks. Most of the work still runs on spreadsheets, PDFs, and human judgment.

That’s changing. Foundation models and structured data unlock a new model: AI-native insurance companies that own the full stack, pricing, underwriting, distribution, and claims, and automate it end-to-end. Think real-time quotes, instant claims approvals, and embedded APIs instead of call centers.

The Opportunity
We’re building an AI-native insurance company that automates the entire value chain, from quote to claim. No agents, no paperwork, no delays. Just structured data and foundation models powering every step: underwriting, policy generation, claims processing, and customer support. The result is a self-driving insurance system that delivers real-time decisions, instant payouts, and fully embedded APIs. We’re starting with focusing on niche products where traditional carriers lag, like gig workers or pet insurance, and using automation to unlock better margins, faster service, and scalable growth.

Why Forum Is Built for This Moment

We're seeking exceptional founders to tackle these concepts head-on. With the support of a dedicated team and $250K in initial funding, you’ll have everything you need to transform these concepts into thriving businesses.

If you’re interested in building a vertically integrated AI company you can find the job posting for Legal, here and for Insurance, here.

While we’re actively building out these ideas, Forum Ventures is always open to hearing from founders with bold, original ideas. If you're working on an early-stage B2B SaaS company, even at the idea stage, we’d love to hear your pitch through the Forum Studio Founder posting.