# What are AI-native services? Canonical page: https://ai-nativeservices.com/guides/what-are-ai-native-services Reviewed: 2026-09-29 This directory covers companies that deliver services themselves using AI for core work. Examples include completed books, reviewed contracts and campaign production. The provider runs delivery and is responsible for its work. Software that helps your team or another firm provide a service falls outside this directory. ## How this directory defines AI-native An AI label tells you little about how a firm operates. We use three tests to decide whether a business belongs in this directory. All three normally need support from public evidence. Any user-confirmed exception is labeled and recorded separately. - The company itself delivers a defined service or completed output, with responsibility for quality and service failures. - AI executes core tasks. The provider describes concrete work such as categorizing transactions, preparing legal documents or producing campaign content. - AI is central to the provider's standard service operation. The delivery model is built or fundamentally redesigned around that execution, with human oversight. ## AI software, AI-assisted firms and AI-native services With AI software, your team or another service firm runs the tool and remains responsible for turning its output into completed work. A contract application might identify clauses, while your lawyer provides the legal service. Software-only suppliers are outside this directory, even when their products perform sophisticated tasks. An AI-assisted firm uses AI within an existing service model. That can improve its work, but using a chatbot or adding an AI practice does not by itself demonstrate a different operating model. This is a classification distinction, not a judgment about service quality. An AI-native service operator builds delivery around AI execution and takes responsibility for the service. It may also provide a dashboard or software interface. The practical question is who runs the process, resolves uncertainty and delivers the agreed result. ## People still have a role AI-native does not mean fully autonomous. Human review, professional judgment, customer communication and physical work can all be part of the service. We do not require a particular automation percentage or founding year. A useful test is to ask what would happen if the provider removed AI. Would its advertised delivery model fundamentally change, or would an otherwise conventional firm simply lose a productivity tool? Public workflow descriptions can help answer that question, though they cannot reveal everything about internal operations. ## What to ask before hiring a provider Start with the work you need completed. A clear scope makes it easier to compare a managed service with software, an employee or an existing outsourcing arrangement. Ask the provider to walk through a representative task from intake to delivery, including an example that needs human intervention. - What deliverable will I receive, when is it due, and what makes it acceptable? - Which tasks does AI perform, what do people review, and who fixes an error? - What information and system access must I provide, and how is that data used? - What is included in the fee, and what happens when scope or volume changes? - Can I export my records and continue the work with another provider? ## What inclusion does and does not tell you Our directory assesses the service model using public company evidence. It does not independently audit automation levels or rank service quality. An AI-native model alone does not establish that a provider is cheaper, faster or a better fit for your work. An older service firm can qualify after a demonstrated redesign. A software supplier does not qualify merely because its product is AI-native. Companies with both models need evidence of their own service delivery, and the listing covers that service only. We keep unclear cases unpublished. ## Examples from the directory These examples illustrate delivery models. They are drawn from the published directory and are not a ranking or a service-quality endorsement. ### [Acorn 9](https://acorn9.com/) Acorn 9 describes continuous AI transaction categorization followed by an accountant's review of the monthly close. The customer receives completed financial statements. This illustrates the distinction between running accounting software and hiring a firm to deliver books. Source: [Acorn 9 bookkeeping workflow](https://acorn9.com/services/bookkeeping). ### [General Legal](https://general.legal/) General Legal sells contract review, drafting and negotiation performed by attorneys working with AI. The service includes legal work and a responsible provider, rather than only access to a document tool. Source: [General Legal services and FAQ](https://general.legal/). ### [Headcount](https://www.aiheadcount.com/) Headcount describes a marketing service that combines agent production with human strategy, approvals and reporting. Its recurring agency work provides an example outside accounting and legal services. Source: [Headcount delivery model](https://www.aiheadcount.com/). [Browse all 201 firms](https://ai-nativeservices.com/) ## Sources and corrections Based on public provider descriptions and the directory's [selection criteria](https://ai-nativeservices.com/#selection-criteria). These guides explain service models, not independently tested performance. Report a correction: contact@toi.company. ## Keep reading - [AI-native accounting firms and bookkeeping services](https://ai-nativeservices.com/guides/ai-native-accounting/index.md) - [AI-native legal services and law firms](https://ai-nativeservices.com/guides/ai-native-legal-services/index.md) - [AI-native marketing agencies and services](https://ai-nativeservices.com/guides/ai-native-marketing/index.md)