As artificial intelligence evolves from conversational assistants into autonomous agents capable of performing business-critical tasks, U.S. companies are confronting a new question: how can businesses adopt increasingly powerful AI systems without losing control over their data, operational history, and digital infrastructure?
Leverage AI Inc., a U.S.-based artificial intelligence company founded by technology entrepreneurs Wei Xu and Wenjie Tan, is developing an answer aimed particularly at America’s small and medium-sized businesses.
The company has announced that Locus, its AI Agent infrastructure platform, has completed development of its core architecture and is entering commercialization.
Rather than focusing solely on making AI models more capable, Locus addresses a different layer of the AI adoption challenge: trust, ownership, security, auditability, and portability.
The platform is designed so that critical AI Agent state—including conversations, memory, permissions, identities, configuration, and audit records—can remain visible and controllable by the business using the system.
For Leverage AI, that architecture reflects a broader thesis: as AI agents gain authority to act on behalf of companies, businesses will need stronger mechanisms to understand what those agents know, what they are permitted to access, what actions they have taken, and how their information can be moved between systems.
“AI is moving from answering questions to actually doing work,” said Wei Xu, co-founder and CEO of Leverage AI. “Once an AI system can send an email, modify a business record, access sensitive information, or initiate a workflow, transparency and control become fundamental infrastructure requirements.”
“Our goal with Locus is to make advanced AI automation practical for businesses that may not have the engineering, security, or compliance resources of a Fortune 500 company.”
Addressing a Growing Challenge in U.S. AI Adoption
Small businesses play a central role in the U.S. economy, accounting for tens of millions of enterprises and a substantial portion of private-sector employment.
At the same time, smaller organizations often operate without dedicated artificial intelligence, cybersecurity, infrastructure, and compliance teams.
That creates a potential adoption gap as increasingly sophisticated AI Agents enter the workplace.
Large corporations can devote significant resources to evaluating AI security, establishing access-control policies, building audit systems, integrating internal databases, and negotiating enterprise data-governance agreements.
A small accounting firm, medical practice, retailer, professional-services company, e-commerce business, or technology startup may not have those capabilities.
Yet these businesses stand to benefit substantially from AI automation.
Leverage AI believes closing this gap requires more than providing smaller companies with access to powerful AI models.
It requires infrastructure that makes those systems understandable, controllable, portable, and auditable by default.
“The question isn’t whether small businesses want AI,” Xu said. “The harder question is whether a business owner is comfortable giving an AI system meaningful access to the company.”
“If businesses cannot clearly understand what an agent remembers, what it can access, and what it has done, adoption becomes much more difficult.”
A Different Architecture for AI Agents
Locus approaches the problem by separating the intelligence of an AI model from the ownership of the operational state surrounding that intelligence.
Instead of requiring a company’s AI memory and operational history to exist exclusively inside a proprietary cloud database, Locus is designed around a concept Leverage AI calls File-Is-State.
Under the architecture, an Agent’s workspace can be represented as a human-readable file structure.
Conversations, memories, permissions, configurations, and other important state can therefore be inspected using conventional computing tools rather than being accessible only through a vendor-controlled interface.
For technical users, ordinary commands and tools such as cat, grep, version control, backups, and file comparisons can be used to inspect or manage portions of the environment.
For businesses, the larger implication is portability.
If an organization changes infrastructure providers or AI models, its accumulated operational state is designed to remain an asset controlled by the organization rather than something that must be reconstructed from a proprietary service.
Leverage AI describes this principle as customer-owned AI state.
Five Technologies Underpinning Locus
The Locus architecture combines five primary technical components.
File-Is-State
The first is its file-based state architecture.
By expressing AI Agent state through readable files, Locus is designed to make information easier to inspect, version, back up, migrate, and audit.
This approach is intended to reduce dependence on proprietary data representations and provide organizations with a clearer record of how their AI systems operate.
Command-Line Agent Orchestration
The second component is a command-line control plane built around familiar shell concepts.
Commands including locus tx, locus q, and locus run allow operations to be recorded and reproduced in ways familiar to software engineers and infrastructure teams.
Rather than requiring every interaction to pass through a proprietary software interface, the system is designed so that important Agent operations can be inspected, logged, replayed, and reviewed.
Context-Based Access Control
The third component is Context-Based Access Control, or CBAC, an authorization model Leverage AI is developing specifically for AI Agent environments.
Traditional enterprise permissions commonly rely on roles and attributes: who a user is, what department the user belongs to, or what predefined permissions have been assigned.
AI Agents introduce an additional challenge because the appropriateness of accessing information may depend on the context in which a request is made.
Locus is designed to incorporate elements of that context into authorization decisions.
The company believes this approach could become increasingly important as AI Agents interact with financial information, customer records, internal communications, healthcare-related information, and other sensitive business data.
Distributed Human-Readable Synchronization
The fourth component is a context-aware synchronization system drawing on Conflict-Free Replicated Data Type, or CRDT, principles.
It is designed to synchronize state across multiple devices and environments while supporting offline operation and conflict resolution.
A business could therefore operate AI workflows across employee computers, mobile devices, point-of-sale systems, servers, or other endpoints while maintaining synchronized state.
Critically, Leverage AI has designed the synchronization layer so that the underlying state remains interpretable by both humans and AI systems.
Edge-First Infrastructure
The fifth component is an edge-first architecture built in part on globally distributed cloud infrastructure.
Rather than depending entirely on centralized application servers, Locus is designed to coordinate selected stateful operations closer to the devices and locations where businesses operate.
The objective is to combine the responsiveness and scalability associated with modern cloud infrastructure with greater customer control over business data and AI state.
Together, these technologies are intended to provide a foundation for AI Agents that can be deployed across real-world business environments while maintaining stronger transparency and portability.
Security and Auditability Become More Important as AI Gains Authority
The significance of these technologies extends beyond convenience.
AI Agents are increasingly being designed to perform actions rather than merely provide information.
An agent may eventually be authorized to update a customer database, communicate with vendors, prepare financial documents, schedule employees, process internal requests, or interact with other software systems.
Each additional capability increases the importance of access control and auditability.
Businesses need to know not only what an employee did, but potentially what an AI Agent did, what information it accessed, under whose authority it acted, and what context led to the action.
Leverage AI is designing Locus with those requirements in mind.
The company believes transparent state management and detailed activity records could be particularly valuable for organizations operating under security, privacy, and compliance requirements.
While the use of Locus does not by itself establish regulatory compliance, its architecture is being designed to support the auditability and data-control practices required by businesses working toward frameworks such as SOC 2 or operating in environments subject to privacy and security obligations.
Reducing Vendor Lock-In in the Emerging AI Economy
Portability is another central component of the company’s strategy.
Today’s businesses increasingly depend on cloud software, but moving years of operational information from one platform to another can be costly and technically difficult.
Leverage AI argues that this problem could become more significant with AI Agents.
An AI Agent may accumulate not only conventional records but also memory, learned business context, workflow history, authorization information, and relationships between employees, customers, and systems.
If that information becomes inseparable from a single AI provider, switching platforms could become increasingly difficult.
Locus is therefore being developed around what the company calls a zero-lock-in principle: businesses should be able to inspect, export, back up, and migrate their AI state.
“AI memory will become business infrastructure,” Xu said. “If years of institutional knowledge accumulate inside an Agent, that information should remain an asset of the company—not an asset that exists only because the company continues paying one software vendor.”
Targeting a Large U.S. Small-Business Market
Leverage AI is initially focusing Locus on digitally active U.S. small and medium-sized businesses operating across multiple applications and devices.
The company estimates an initial target segment of approximately 330,000 digitally sophisticated U.S. SMBs that are comparatively well positioned to adopt AI-driven workflows.
Locus is expected to be offered through subscriptions starting at approximately $50 to $100 per business per month, significantly below the cost of building comparable internal AI infrastructure.
The pricing strategy reflects the company’s objective of making advanced Agent infrastructure accessible beyond large enterprises.
Leverage AI also expects its local-first architecture to produce different economics from conventional centralized SaaS platforms.
Because portions of computation and state can remain on infrastructure controlled by customers, growth in customer data does not necessarily require proportional increases in centralized cloud storage.
If successfully deployed at scale, the architecture could make sophisticated AI infrastructure economically accessible to a broader segment of American businesses.
Potential Impact Beyond a Single Product
Leverage AI’s ambitions for Locus extend beyond selling an individual software application.
The company views data ownership, contextual permissions, portability, and Agent auditability as infrastructure problems likely to affect a wide range of industries as AI adoption accelerates.
Professional services, healthcare-adjacent businesses, retail, e-commerce, financial operations, logistics, hospitality, and technology companies are among the sectors where autonomous software could increasingly interact with sensitive operational information.
Establishing technical mechanisms that allow organizations to retain control of that information could therefore have implications beyond any single customer or industry.
The company believes that wider adoption of transparent and portable AI infrastructure could lower barriers to AI adoption for smaller organizations, strengthen accountability for autonomous systems, and reduce the technical dependency created when business-critical AI state resides exclusively within proprietary platforms.
That broader challenge is what attracted Xu to the problem.
“America has millions of businesses that could benefit from AI but cannot build an internal AI infrastructure team,” he said. “If advanced Agent technology is only practical for the largest corporations, we will leave a significant part of the economy behind.”
“We want to make secure AI infrastructure something a small company can actually deploy.”
Founder Wei Xu Brings a Track Record in Scaling Technology Products
Leverage AI’s development is led by a founding team with experience in consumer technology, artificial intelligence, advertising, growth, and distributed systems.
CEO Wei Xu previously held technical leadership roles at Meta, where he worked on advertising monetization and consumer growth.
According to information provided by the company, Xu led a 15-person team whose advertising products expanded from approximately $400,000 in daily revenue to more than $40 million per day, representing roughly 100-fold growth.
He also contributed to major consumer growth systems, including work related to Facebook’s People You May Know recommendation infrastructure and initiatives associated with approximately 20 million daily active users.
During his tenure, Xu led teams that reached the finals of Meta’s internal hackathons three times, where product prototypes received senior-level review, including feedback from Meta founder Mark Zuckerberg.
His subsequent entrepreneurial record provided another test of his ability to take technology from concept to large-scale adoption.
In 2021, Xu co-founded LiveIn, also known as Livehouse, and served as CEO.
According to company figures, the social application grew from zero to more than 2 million daily active users within approximately four months of its U.S. launch, without a conventional paid marketing campaign.
LiveIn subsequently reached No. 1 overall on the U.S. App Store, competing with some of the world’s largest consumer applications, and attracted coverage from major technology and business publications.
Xu later advised multiple consumer technology companies on U.S. product growth. Leverage AI says applications he supported included eight products that entered the top five of the U.S. App Store overall rankings, including LiveIn at No. 1 and Clapper at No. 2.
From 2023 through 2025, Xu worked at TikTok/ByteDance as a product manager focusing on video and e-commerce products and participated in the development and relaunch of TikTok Now.
That combination of engineering, product development, monetization, and large-scale consumer growth now informs Leverage AI’s approach to AI infrastructure.
Co-Founder Wenjie Tan Leads the Technical Architecture
Leverage AI co-founder and CTO Wenjie Tan leads engineering for Locus, including the platform’s edge synchronization, distributed state architecture, CRDT implementation, and security infrastructure.
Tan and Xu jointly developed the architecture and protocol design underlying the platform.
The founders say their objective is not simply to place an AI interface on top of existing SaaS architecture, but to reconsider how software should manage state and authorization when the software itself can act autonomously.
That distinction is increasingly important as the technology industry shifts from generative AI toward agentic systems.
Building on Leverage AI’s Earlier Commercial Experience
Locus represents the next stage in Leverage AI’s development.
The company initially applied artificial intelligence to digital advertising, developing technology designed to automate elements of video-advertisement creation, localization, and optimization.
Its early commercial activity included work with social media platform Clapper, providing Leverage AI with experience deploying AI technology for real-world business applications.
Through that work, the founders began confronting a larger infrastructure question.
Generating content was becoming easier. Giving AI systems responsibility for persistent business workflows was significantly harder.
The challenge increasingly involved memory, permissions, identity, synchronization, security, and accountability.
That realization ultimately shaped the development of Locus.
“The first generation of generative AI was primarily about creating content,” Xu said. “The next generation is about AI participating in operations.”
“That changes the engineering problem. Intelligence alone isn’t enough. Businesses need to know who controls the Agent, what it knows, what it is allowed to do, and whether they can audit it afterward.”
Expanding U.S. Product Development and Hiring
Leverage AI is advancing Locus into commercialization while expanding its product and engineering efforts in the United States.
The company is pursuing U.S. venture investment to support continued research and development, security engineering, product commercialization, and team growth.
Its hiring strategy prioritizes U.S.-based engineering and technical talent as the company develops the platform for broader commercial deployment.
Over time, Leverage AI plans to expand Locus’s integrations and capabilities so that AI Agents can operate across a broader range of business systems while preserving the platform’s core principles of customer ownership and auditability.
The company believes that successful commercialization could contribute to broader adoption of artificial intelligence among American small businesses while supporting new technical work in distributed systems, AI security, access control, and Agent infrastructure.
The Next Question in AI: Not Just What Can It Do, but Who Controls It?
The artificial intelligence industry has spent the past several years competing primarily on model capability.
Larger models can reason more effectively, generate increasingly sophisticated content, and interact with a growing number of external tools.
But as those systems become more capable, another set of questions is becoming increasingly important.
Who owns an AI Agent’s memory?
Who controls its permissions?
Can its decisions be audited?
Can a company move its accumulated AI knowledge to another platform?
And when an Agent acts autonomously, can a business reconstruct exactly what happened?
Leverage AI believes those questions will become fundamental to enterprise technology.
“When software was passive, vendor lock-in was primarily a technology and economic issue,” Xu said. “When software can act autonomously on behalf of a company, control of that software becomes a governance issue.”
“Our principle is straightforward: your AI should work for your business, its memory should belong to your business, and its actions should be accountable to your business.”
For America’s small businesses, that distinction could determine not only which AI systems they adopt—but whether they are willing to entrust autonomous AI with meaningful business operations at all.
About Leverage AI Inc.
Leverage AI Inc. is a U.S.-based Delaware C corporation developing artificial intelligence infrastructure for small and medium-sized businesses.
Its Locus platform is designed around customer-owned AI state, human-readable data, contextual access control, distributed synchronization, edge-first infrastructure, auditability, and data portability.
The company’s mission is to enable businesses to adopt increasingly capable AI Agents while maintaining visibility and control over their data, memory, permissions, and AI activity records.
Leverage AI is advancing Locus through commercialization in the United States while expanding its engineering, security, and product-development efforts.
