The next great software company may not look like a software company.
It may look like an accounting firm. Or a legal service. Or an insurance broker. Or an IT managed services provider. The difference is that underneath the surface, the work will be done by AI agents, software workflows and a much smaller layer of human experts.
That is the promise of AI-native services.
For the past two decades, the startup playbook was to take a manual workflow and turn it into SaaS. Salesforce digitised sales. QuickBooks digitised accounting. DocuSign digitised signatures. ServiceNow digitised internal workflows. The software company sold tools, dashboards and licences. The customer still had to employ people to use them.
AI changes that equation.
The next wave of companies may not simply sell software to professionals. They may become the professional service provider itself. Instead of selling a tool to a lawyer, accountant, broker, recruiter or IT manager, they sell the finished work: the contract reviewed, the books closed, the policy placed, the candidate sourced, the tax return filed, the ticket resolved.
That is a very different business model. It is not “AI as a feature.” It is AI as the operating system for a service business.
Sequoia’s Julian Bek recently described this shift as “services are the new software.” The insight is powerful. A co-pilot sells the tool. An autopilot sells the work. YC has also been pushing founders in this direction, calling for AI-native service companies in categories like accounting, tax, audit, compliance, insurance brokerage and healthcare administration.
But there is an important nuance that is sometimes missed.
A lot of the current conversation assumes the buyer is an enterprise. Sell AI to the law firm. Sell AI to the accounting firm. Sell AI to the insurance company. Sell AI to the consulting firm. That is a big opportunity, and companies like Harvey and Legora in legal show how much value can be created by giving incumbents better tools.
But the bigger pull may come from customers who do not have the professional in the seat at all. For lack of a better term, we call them SMBs.
A 25-person construction company still needs contracts, tax, insurance, payroll, compliance and bookkeeping. A dental clinic still has claims, admin and finance workflows. A small ecommerce company still needs customer support, fulfilment, tax and supplier management. These businesses have real obligations, but they usually do not have in-house legal, finance, HR, procurement or compliance teams.
For them, a co-pilot is not enough.
There is nobody to co-pilot.
They do not want another dashboard. They want the work done.
That is why AI-native services could be especially powerful for SMBs. Across the UK, EU and US, there are millions of companies that are too big to ignore legal, finance and compliance work, but too small to hire full-time experts. They are forced to outsource the work, do it badly, or let the founder handle it at night.
This is where AI-native service companies have a wedge.
They can offer the missing department.
What is an AI-native service company?
An AI-native service company is not just a traditional services firm using ChatGPT internally.
It is a company built from the ground up to deliver a business outcome through software, agents, data and expert oversight. The customer does not buy seats. The customer buys the result.
A legal AI tool helps a lawyer draft faster. An AI-native legal service delivers the contract.
An accounting co-pilot helps an accountant reconcile faster. An AI-native accounting service closes the books.
An insurance workflow tool helps a broker process forms. An AI-native broker gets the business covered.
The difference is subtle but important. In SaaS, the customer owns the workflow. In AI-native services, the company owns the workflow.
That changes everything.
It changes the product. It changes the pricing. It changes the margin structure. It changes the go-to-market. It also expands the market from software budgets to labour and services budgets.
I wrote about this broader shift recently on VC Cafe: AI is no longer only attacking enterprise software spend. It is starting to attack the much larger pool of services and labour spend. That is where the next generation of venture-scale companies may emerge.
Why this is happening now
Professional services were historically difficult venture businesses.
They scaled with people. Growth required more headcount. Margins were capped by salaries, utilisation and quality control. A great law firm, accounting firm, consultancy or brokerage could be a very profitable business, but it usually did not have the scalability of software.
AI changes the margin structure.
If software agents can perform 70% or 80% of a repeatable workflow, the economics start to look very different. Human experts are still required, especially in regulated or high-stakes categories. But their role shifts. They supervise. They handle exceptions. They review edge cases. They provide accountability and trust.
The result is a hybrid company. Part software platform. Part expert service provider. Part operations machine.
That hybrid is what makes the category interesting.
The company is not replacing professionals with a chatbot. It is rebuilding the service model around AI leverage.
The market is already forming
Legal is probably the most visible early example.
The category is text-heavy, expensive and full of repeatable workflows. On one side, companies sell AI tools to law firms and legal departments. On the other side, AI-native legal companies are beginning to deliver legal outcomes directly. The emerging map includes companies such as Lawhive, Eudia, Norm AI, Justpoint, Crosby and Garfield AI.
Accounting and finance may be next.
The pain is obvious. Bookkeeping, tax, audit, payroll, reconciliation, reporting and CFO services are essential, repetitive and often underserved. Examples in the market include Rillet (AI Native ERP), Pilot.com, finally, OSOME, Cranston and Minerva (AML monitoring).
In legal, some companies sell AI to law firms. Others become AI-native legal providers. In accounting, some sell software to finance teams. Others close the books directly. In insurance, some sell workflow tools to brokers. Others become the broker.
This is not a contradiction. It is the same thesis from two angles.
Insurance is also being rebuilt in layers.
Some startups are focused on brokerage. Others are focused on claims, underwriting or policy administration. The map includes companies like Harper, WithCoverage, Pace, Fernstone and Panta.
M&A advisory is earlier, but fascinating.
A lot of the work in boutique advisory is research, buyer mapping, outreach, process management, diligence, financial analysis and document preparation. Companies such as Eilla AI and OffDeal are exploring what it means to deliver parts of that work through an AI-native model.
Procurement is another obvious candidate.
Vendor discovery, intake, quote comparison, contract coordination and negotiation support are painful workflows that many mid-market companies manage badly. Companies like Lio (formerly askLio) and Magentic AI are examples of the opportunity here.
The same pattern is appearing in other categories too. Jack & Jill is applying the model to recruiting. TaxGPT is focused on tax advisory. Anterior is tackling healthcare workflows. Serval is building in IT managed services. In consulting for example, Accenture spent $1 billion to acquire UK-based AI consultancy startup Faculty.
In AI implementation, the category is also evolving quickly. Anthropic, Blackstone, Hellman & Friedman and Goldman Sachs just announced a new AI-native enterprise services firm to help businesses deploy Claude into core operations, a signal that even the largest AI labs see implementation as a services opportunity, not just a software distribution problem
None of these markets are identical. But they share a common theme: AI is moving from helping the worker to performing the work.
This is not a kept secrete. A recent AI-native service firms database by Nikola Lazarov mapped 211 companies across 70 industries and 23 countries, with more than $5 billion raised collectively. The most funded categories include security, accounting, legal and insurance
The founder opportunity
The best AI-native services companies will not start with a broad category.
They will start with a narrow, painful workflow where the customer already pays someone else to get the work done. That matters because there is existing budget. The customer already understands the pain. The founder is not trying to create a new habit from scratch.
Case in point, Emerge, an early stage venture fund focused on educational technologies, mapped out every industry’s spend on professional services and then scored every industry-by-professional services cell against four major criteria:
- Is the spend compulsory or discretionary? Tax needs to be filed. Books need to close. Insurance needs to be renewed. Compliance requirements need to be met.
- How high are the stakes if the AI gets it wrong? if the stakes are high, it creates willingness to pay and makes trust a real moat.
- Is the workflow repeatable or bespoke? Can the company can standardise the process across customers and improve delivery over time?
- Is the customer is underserved? This is especially true in SMB markets, where the alternative is often expensive, slow or unavailable.
This is why “boring” markets are so attractive.
AI bookkeeping for dentists. AI compliance for construction companies. AI claims handling for a specific insurance line. AI tax support for cross-border freelancers. AI procurement for care homes. AI IT support for 100-person businesses. AI contract review for agencies.
These do not sound like flashy startup ideas. That is exactly why they may be good.
Boring markets often have urgent problems, existing budgets and little love from traditional software companies. Maybe that’s the reason that Y combinator is actively looking to back AI native service companies.
The hard part
There is a risk in making AI-native services sound easy.
They are not.
These companies have to solve software problems and services problems at the same time. They need reliable AI systems, but also strong operations. They need automation, but also quality control. They need distribution, but also trust. They often need regulatory understanding, insurance, human review and clear accountability.
The last 10% of the workflow may be the hardest part.
That is also where the moat may come from.
A shallow AI wrapper can generate a document. A real AI-native service company can deliver the correct document, in the right context, with the right approvals, at the right time, with someone accountable when it matters.
That is much harder to copy.
The best teams will combine deep domain expertise with technical ambition. A pure AI team may underestimate the complexity of the service. A pure services team may fail to build enough software leverage. The winners will understand both.
The next SaaS may not look like SaaS
The first generation of SaaS replaced on-premise software. The second generation made every department more productive. The AI-native generation may do something different.
It may replace the need for the department, or make that department available to companies that could never afford it before.
That is the real opportunity.
- Not software that helps the accountant. Software that closes the books.
- Not software that helps the broker. Software that gets the customer insured.
- Not software that helps the recruiter. Software that finds and qualifies the candidate.
- Not software that helps the IT helpdesk. Software that resolves the ticket.
The question is no longer only, “What software can I sell to this industry?” The better question is, “What work can I take off the customer’s plate entirely?”
At Remagine Ventures, we are especially interested in founders building at the intersection of AI, workflow automation, trust and vertical expertise. Some of the most exciting companies of the next decade may not describe themselves as SaaS companies at all. They may look like accounting firms, legal providers, brokers, compliance teams, IT service providers or advisory firms.
But underneath the surface, they will be software companies.
They will sell outcomes, not seats. They will use agents, not just dashboards. They will combine automation with expert oversight. They will start narrow, build trust and expand across the customer’s workflow.
The next great SaaS company may not sell software.
It may sell the work.
