Since the advent of ChatGPT 3.5, AI has sparked a Cambrian explosion of innovation. Decades of advancements—Moore’s law, internet connectivity, cloud computing, and mobile technology—have converged, catalyzing a transformative shift for humanity. With each new release of foundational models, AI capabilities have increased exponentially, leaving many in awe. Progress was so rapid that even Elon Musk called for a pause in AI development to avoid the risk of machines surpassing human control.

However, while excitement reached a fever pitch in Silicon Valley, businesses began questioning whether AI could truly deliver on its promises. Initial Proof of Concepts (PoCs)—once seen as full of potential—were sidelined as errors, hallucinations, and organizational inertia tempered early optimism. Despite this "trough of disillusionment," a new wave of venture-backed companies emerged, leveraging AI to take over segments of the knowledge economy that traditionally required professional expertise. Rather than simply providing software for customers to implement, these companies use AI—often supported by humans-in-the-loop—to offer repeatable, ongoing services. Although they may operate with slimmer margins than traditional SaaS, they are capturing a larger share of economic value and demonstrating significant growth potential.

The Shift from Traditional SaaS to Service-as-a-Software

To grasp the significance of this shift, it's essential first to understand the origins of traditional SaaS.

Earlier generations of SaaS companies—like Salesforce, Procore, and Netsuite—focused on digitizing analog processes and streamlining workflows. However, they relied heavily on employees, such as salespeople and analysts, to input and manage data. The sales process was grueling, with teams pitching software as a solution and convincing potential customers of its benefits. Once a sale was made, customer success teams guided clients through software integration. Only after a lengthy adoption period could SaaS companies solidify their position as the system of record, paving the way for recurring profits through contract renewals.

This dynamic has shifted in the current AI era. Rather than offering tools to boost employee productivity, AI-enabled services are replacing entire job functions. These companies deliver finished products at a fraction of the cost that internal teams or in-house tools could achieve. This "Service-as-a-Software" model bypasses the tedious adoption cycle, providing a complete solution without requiring clients to undergo disruptive organizational changes.

The Rise of Service-as-a-Software

Unlike traditional SaaS, Service-as-a-Software companies combine AI with human oversight to produce high-quality deliverables at lower marginal costs. While this model has wide applications, its greatest value is found in industries like law, where workflows revolve around unstructured text, synthesis, and content generation. A notable example is EvenUp, which has disrupted the plaintiff's law industry by generating demand letters and case chronologies for attorneys—tasks typically handled by in-house paralegals. With AI at its core, EvenUp ensures high-quality output, monitored by injury and intake experts, allowing attorneys to achieve faster settlements—often seven weeks quicker—and at 30% higher values than doing the work in-house. For instance, EvenUp can replace an $80,000-a-year paralegal at a fraction of the cost.

While law provides a high-value use case, AI-enabled services are proliferating across various industries. Knowledge economy jobs generally fall into four categories:

1. Relationship-driven with low automation potential

2. Relationship-driven with high automation potential

3. Transactional and procedural with low automation potential

4. Transactional and procedural with high automation potential

Jobs requiring higher levels of judgment will still likely involve humans to ensure quality and meet regulatory requirements. However, for transactional, repeatable tasks, AI can increasingly replace human functions entirely.

For example, simple data entry jobs are becoming fully automated, even in regulated industries. Tennr, a healthcare company, uses AI agents to manage patient workflows related to referrals and medical histories—tasks that once relied on paper faxes due to HIPAA regulations. In contrast, companies like Athena (personal assistants) and Moonhub (HR screening and onboarding), which depend on human-to-human interactions, will likely retain some human-in-the-loop elements. However, their workloads will gradually become more autonomous. Nonetheless, In industries like law or corporate accounting, where human judgment and regulatory compliance are critical, a human-in-the-loop model will remain necessary for the long term.

The Opportunity in Service-as-a-Software

Service-as-a-Software companies bridge the gap between service margins and software margins. Traditional SaaS companies typically enjoy high margins (above 80%) but often struggle with slow adoption. Conversely, service businesses have historically operated with lower margins (below 30%). However, as AI replaces labor-intensive processes, service margins are beginning to resemble software's. This convergence presents significant opportunities for both margin expansion and market growth.

As these companies become more integrated into their customers' workflows, their competitive moat strengthens. By building software around their services, Service-as-a-Software firms create "gravity" around their solutions.

For example, EvenUp positions its platform as the system of record for its users, ensuring long-term customer engagement and value. Companies that achieve this will become generational players in the first wave of the AI era, dramatically transforming the knowledge economy. We will likely see a large number of new unicorns over the next five years, and they will grow faster than ever before because their time to value creation has significantly decreased. Eventually, they will rip out their SaaS predecessors. These companies will all be more verticalized but be considerably larger because they can extract a larger share of the economic pie, thus establishing Unicorns in a plethora of different verticals.

Market Map