Monetization, Moats, and The Rise of AI-First Service Startups
Brief thoughts on the emergence AI-First service co's
The Emergence of AI-First Service Businesses
With the rise of AI technologies like ChatGPT, there have been many discussions about the opportunities for new software applications in an AI-driven world. Equally important is the question of how defensible or truly proprietary these new ventures are, especially considering how easy it has become to develop AI-driven proofs of concept, compared to the longstanding advantages that established SaaS incumbents hold.
Traditional barriers like exclusive access to data and top talent, along with the ability to target underserved TAMs, can indeed create strong competitive advantages. Yet, we're only beginning to grasp the full spectrum of opportunities that AI presents.
Reimagining Service Delivery: Beyond Software
By focusing on specific problems or niches, startups can leverage small & large language models and specialized data to rebuild service businesses from the ground up, automating them to a degree that makes them scalable like a SaaS company. I believe to achieve success with this model, companies will have to quickly move away from reliance on the human(s) in the loop because it simply kills margins. The goal should be to morph into something that emulates a SaaS model but with improved economics and defensibility as the AI becomes smarter and smarter.
AI-First Service Co’s = Venture Scale Potential
The sheer dollar volume pouring into AI infra will need to be justified by creating significant value at the application level. While some of this value will be captured by established companies, there's a significant opportunity for net new startups to disrupt traditional business models by "selling the work" rather than just selling software.
Selling the Work: A New Opportunity
"Selling the work" is distinct from simply adding AI features to existing software. It represents a shift in the value proposition, where businesses sell services powered by AI rather than just the tools to perform those services. This approach allows new players to fully automate specific segments, which incumbents might struggle to do without disrupting their existing business models.
Two companies I’m actively following that are “selling the work” right now are:
An LA-Based AI-native MSO for software development that removes 90% of humans in the loop [currently in stealth]
Mechanized.ai - An Atlanta-based no-code platform that streamlines the entire AI and ML deployment process, from ideation to production for enterprises
A New Operational Playbook
For AI-first service businesses, the strategy is clear: focus on time/people intensive use cases, create easily quantifiable value add, and use human oversight to ensure quality until full automation is possible. I believe the companies that will see outlier success in this space will start off serving a small, well-defined & niche customer base, and tangentially expand after >95% automation with human-in the loop oversight.
A Different Monetization Model
Business models for AI-First Service companies will not be your standard B2B SaaS. Thus far I’m seeing a true hybrid mix of some platform SaaS, varying usage tiers/ credit based, and outcomes based. This approach aligns revenue directly with the value delivered to the customer, creating a more flexible and scalable model that can adapt to the unique demands and varying intensities of AI-driven services. One of our existing portfolio companies, Sailes, is doing an excellent job with this hybrid revenue model as they price based on “Digital Labor™” which is the sum of prospecting tasks conducted by their core product, the Sailebot.
Bigger Markets, Bigger Opportunities
Earlier, I mentioned niche customer bases as a potential GTM for AI-first service co’s, but I want to clarify that focusing on a niche customer or segment doesn’t mean building a small business. By automating service delivery, these businesses can tap into larger budgets typically reserved for HR or operations, capturing more significant contracts than traditional software vendors. This creates a swim-lane for AI-first service businesses to target markets with smaller TAMs while still achieving meaningful value capture of the company’s domain spend.
Data Moats: Building Defensibility
AI companies are data companies and as AI-driven services push towards full automation in specific niches, the data collected along the way arguably becomes the startups most valuable asset. While SaaS incumbents may achieve broad automation, AI-first businesses focused on specific problems like offshore managed dev can build real data moats. Specific to the AI-Native managed service org use case, here are some ways I see potential moats being built (these apply to broader use cases as well):
Dynamic Resource Allocation Models: Utilizing AI to manage and optimize the allocation of offshore resources dynamically based on real-time project data. By continuously refining these resource allocation models with data from past projects (e.g., developer performance metrics, project timelines, client satisfaction scores), you can build a system that significantly outperforms static or human-managed processes.
Ecosystem Lock-In through Integration: Developing deep integrations with clients’ existing systems and workflows that leverage AI for decision-making and process automation. By becoming an integral part of their operational ecosystem, the AI-First service becomes more indispensable over time.
Optimism Is In Abundance
I believe we are in the early innings of what can only be described as the space race of AI deployment and development. While the road ahead is riddled with potential GTM, monetization, and scale challenges, the numerous opportunities to build are equally compelling. The founders building the coming wave of AI-first service businesses will fundamentally reshape how we view the world by striking the right balance between automation and service quality, navigating the trade-offs between growth and scalability, and crafting value props that go beyond mere cost savings. If you're a founder operating in this space or simply curious about the future of AI-first services, let’s exchange notes and connect.