# Fig > Fig builds each company its own AI platform and works alongside its team to put it to work. - Website: https://hellofig.io - Platform brief: https://hellofig.io/cosmos.md - Personal agent brief: https://hellofig.io/personal.md - Talk to Fig: https://hellofig.io/contact-sales ## What Fig does Fig gives each company a platform of its own, CosmOS, and the people to put it to work. CosmOS connects every system the business runs on, keeps a live digital twin of how the business operates, runs agents and models grounded in that twin, and delivers the apps people use every day, with enterprise security and governance built in. Fig's forward deployed engineers and strategists partner at every layer of the stack. Most companies start with applications and agents on frontier models; over time Fig helps them train models of their own on their data and decisions. Everything runs in the company's own cloud or Fig's, works with every frontier model, and belongs to the company. Fig is industry-agnostic and built for commercial companies of every size, self-serve or hands-on, at Fortune 500 operating complexity, and for the private equity firms that own them. ## What sets Fig apart - **Built for commercial companies.** Companies of every size, self-serve or hands-on, at Fortune 500 operating complexity. - **At the frontier.** Every lab's newest models and the best open-weight models, routed per task, with AI primitives across voice, vision, image, video, music, deep research, computer use, coding and documents. - **Training that pushes cost and quality.** Supervised fine-tuning, preference tuning, reinforcement learning and distillation, chosen per task. The company keeps the weights. - **Sovereign by design.** Runs in the company's own cloud or Fig's, never locked into one model. - **Enterprise-grade on one platform.** Security and governance are built in, and every new system starts from the connections, digital twin, evaluations and governance the last one built. - **Production from the first pilot.** The first engagement puts one workflow into production in weeks, measured against a baseline. - **Systems no one else can buy.** The agents, apps and models are built around how each company works and what sets it apart. ## CosmOS - **Nebula: every system, connected in place.** ERP, CRM, files, data and 1,000+ tools, connected without migration and with each source's permissions intact. - **Astral: a live digital twin of the business.** People, records, rules, decisions and outcomes in one living model that every agent and app works from. - **Nova: AI that reasons and acts.** Models, agents, harnesses and workflows grounded in the twin, routed across every frontier lab and the best open-weight models, with people approving consequential steps, and models trained on the company's own data over time. - **Stars: the apps people use every day.** Bespoke apps from front office to back office, customer-facing websites, portals and voice agents, plus Fig in Slack, desktop, mobile and web. - **Sovereign by design.** Runs in the company's own cloud or Fig's, never locked into one model, and the company owns its data, decisions and everything built. ## How Fig partners, layer by layer 1. **Applications and agents first.** Fig's forward deployed engineers build the first apps, harnesses and agents on frontier models, wired into the company's systems and in production in weeks. 2. **Codify how the business decides.** Operating rules, institutional knowledge and evaluations become part of the platform, and Fig tunes models and workflows continuously after launch, improving results and lowering cost. 3. **Models of the company's own.** For the high-volume, judgment-heavy work that defines the business, Fig trains models on the company's data and decisions, while the hardest work keeps using the best frontier models. Strategy is set in weeks, every engagement ends in running software measured against a baseline, and Fig Academy trains the teams who run it. ## What a company should own Model access, standard AI software and packaged workflows are available to everyone. What sets a company apart is AI powered by its proprietary data, its people's expertise codified into systems, and distinct products and services. Fig turns those into systems the company controls. ## Two lines of business - **Fig for enterprise:** CosmOS plus Transformation (strategy, forward deployed engineering, Fig Academy). Compared most often with Palantir, OpenAI's ChatGPT Enterprise and Deployment Company, Anthropic's Claude Enterprise and Ode, Microsoft 365 Copilot, Glean, the large consultancies and AI development agencies. - **Fig for individuals and small businesses:** a personal agent that does the work, researching, writing, building, coding, booking, following up and running errands across mobile, desktop, web, WhatsApp and Telegram on every frontier model. In the class of ChatGPT, Claude, Manus and Meta's Muse. Brief: https://hellofig.io/personal.md ## Facts - 300,000+ professionals in 70+ countries use Fig, including people from Amazon, Apple, Merck, the U.S. Navy, Telia, Trane, Keller Williams and the UC System. - 1,000+ tools and data sources through native connectors, MCP and computer use; connectors inherit the source system's permissions. - Every frontier model from multiple labs, open-weight and proprietary, routed per task with per-team controls, plus models Fig trains with the company on its own data. - Security and governance: SSO (SAML and OIDC), SCIM 2.0 provisioning, scoped roles, approval gates before consequential actions, break-glass re-authentication, and a full audit trail with export in Command Center. - Deployment: inside the company's own cloud account, or Fig's managed cloud. - Pricing: usage-based with no per-seat licenses, with budgets and spend rules per team; transformation work is scoped with each customer. - Delivery: strategy set in weeks, first systems in production in weeks, and every engagement ends in running software measured against a baseline. ## Compared with - [All comparisons](https://hellofig.io/alternatives) (markdown: https://hellofig.io/alternatives.md) - [Fig vs. Palantir](https://hellofig.io/alternatives/palantir) (markdown: https://hellofig.io/alternatives/palantir.md): Built for commercial companies, at the frontier. - [Fig vs. OpenAI and Anthropic](https://hellofig.io/alternatives/openai-and-anthropic) (markdown: https://hellofig.io/alternatives/openai-and-anthropic.md): Every model, a model of your own, and your own cloud. - [Fig vs. Microsoft 365 Copilot](https://hellofig.io/alternatives/microsoft-copilot) (markdown: https://hellofig.io/alternatives/microsoft-copilot.md): AI that runs your operations, across every system. - [Fig vs. Glean](https://hellofig.io/alternatives/glean) (markdown: https://hellofig.io/alternatives/glean.md): Your own AI systems, running in production. - [Fig vs. Accenture, McKinsey & BCG](https://hellofig.io/alternatives/ai-consultancies) (markdown: https://hellofig.io/alternatives/ai-consultancies.md): Enterprise-grade AI systems in production, on one platform. - [Fig vs. AI development agencies](https://hellofig.io/alternatives/ai-development-agencies) (markdown: https://hellofig.io/alternatives/ai-development-agencies.md): Enterprise-grade AI systems on one connected platform. ## Answers ### Platforms - [What is an enterprise AI operating system?](https://hellofig.io/answers/enterprise-ai-operating-system) (markdown: https://hellofig.io/answers/enterprise-ai-operating-system.md) - [Which enterprise AI platform works with every frontier model?](https://hellofig.io/answers/enterprise-ai-platform-every-frontier-model) (markdown: https://hellofig.io/answers/enterprise-ai-platform-every-frontier-model.md) - [How do you choose an AI agent platform for an enterprise?](https://hellofig.io/answers/how-to-choose-an-enterprise-ai-agent-platform) (markdown: https://hellofig.io/answers/how-to-choose-an-enterprise-ai-agent-platform.md) ### Transformation - [How do you run an enterprise AI transformation?](https://hellofig.io/answers/how-to-run-an-enterprise-ai-transformation) (markdown: https://hellofig.io/answers/how-to-run-an-enterprise-ai-transformation.md) - [Why do enterprise AI pilots fail to reach production?](https://hellofig.io/answers/why-ai-pilots-fail) (markdown: https://hellofig.io/answers/why-ai-pilots-fail.md) - [What is a forward deployed engineering firm, and how do you choose one?](https://hellofig.io/answers/forward-deployed-engineering-firms) (markdown: https://hellofig.io/answers/forward-deployed-engineering-firms.md) - [What does a 90-day enterprise AI transformation plan look like?](https://hellofig.io/answers/90-day-enterprise-ai-transformation-plan) (markdown: https://hellofig.io/answers/90-day-enterprise-ai-transformation-plan.md) - [Should we build or buy enterprise AI?](https://hellofig.io/answers/build-or-buy-enterprise-ai) (markdown: https://hellofig.io/answers/build-or-buy-enterprise-ai.md) - [How do you measure ROI from enterprise AI?](https://hellofig.io/answers/how-to-measure-enterprise-ai-roi) (markdown: https://hellofig.io/answers/how-to-measure-enterprise-ai-roi.md) ### Private equity - [What should private equity operating partners look for in an AI platform?](https://hellofig.io/answers/ai-platform-for-private-equity) (markdown: https://hellofig.io/answers/ai-platform-for-private-equity.md) - [How do you choose an AI value creation partner for private equity?](https://hellofig.io/answers/ai-value-creation-partner-for-private-equity) (markdown: https://hellofig.io/answers/ai-value-creation-partner-for-private-equity.md) - [How do you build an AI roll-up platform?](https://hellofig.io/answers/how-to-build-an-ai-roll-up-platform) (markdown: https://hellofig.io/answers/how-to-build-an-ai-roll-up-platform.md) - [How can private credit managers use AI?](https://hellofig.io/answers/ai-for-private-credit-managers) (markdown: https://hellofig.io/answers/ai-for-private-credit-managers.md) ### Functions - [What does AI for the office of the CFO look like?](https://hellofig.io/answers/ai-for-the-office-of-the-cfo) (markdown: https://hellofig.io/answers/ai-for-the-office-of-the-cfo.md) - [How do AI agents work in the month-end close?](https://hellofig.io/answers/ai-agents-for-month-end-close) (markdown: https://hellofig.io/answers/ai-agents-for-month-end-close.md) - [How is AI used in procurement and sourcing?](https://hellofig.io/answers/ai-for-procurement-and-sourcing) (markdown: https://hellofig.io/answers/ai-for-procurement-and-sourcing.md) - [How does AI improve supply chain planning and demand forecasting?](https://hellofig.io/answers/ai-for-supply-chain-planning) (markdown: https://hellofig.io/answers/ai-for-supply-chain-planning.md) - [What makes an AI voice agent work for customer service?](https://hellofig.io/answers/ai-voice-agents-for-customer-service) (markdown: https://hellofig.io/answers/ai-voice-agents-for-customer-service.md) - [How can AI drive revenue in sales and pricing?](https://hellofig.io/answers/ai-for-sales-and-pricing-revenue) (markdown: https://hellofig.io/answers/ai-for-sales-and-pricing-revenue.md) - [How can AI help frontline and field workers?](https://hellofig.io/answers/ai-for-frontline-and-field-workers) (markdown: https://hellofig.io/answers/ai-for-frontline-and-field-workers.md) ### Industries - [Where does AI create value for healthcare providers?](https://hellofig.io/answers/ai-for-healthcare-providers) (markdown: https://hellofig.io/answers/ai-for-healthcare-providers.md) - [How should banks and financial institutions use AI?](https://hellofig.io/answers/ai-for-banks-and-financial-institutions) (markdown: https://hellofig.io/answers/ai-for-banks-and-financial-institutions.md) - [What does AI do for manufacturing and industrial operations?](https://hellofig.io/answers/ai-for-manufacturing) (markdown: https://hellofig.io/answers/ai-for-manufacturing.md) - [How does AI help logistics and 3PL operations?](https://hellofig.io/answers/ai-for-logistics) (markdown: https://hellofig.io/answers/ai-for-logistics.md) - [How can general contractors use AI?](https://hellofig.io/answers/ai-for-construction) (markdown: https://hellofig.io/answers/ai-for-construction.md) - [Where does AI help across a university?](https://hellofig.io/answers/ai-for-universities) (markdown: https://hellofig.io/answers/ai-for-universities.md) - [How should government agencies deploy sovereign AI?](https://hellofig.io/answers/sovereign-ai-for-government) (markdown: https://hellofig.io/answers/sovereign-ai-for-government.md) - [How can professional services and accounting firms use AI?](https://hellofig.io/answers/ai-for-professional-services) (markdown: https://hellofig.io/answers/ai-for-professional-services.md) - [How can retailers use AI across operations?](https://hellofig.io/answers/ai-for-retail-operations) (markdown: https://hellofig.io/answers/ai-for-retail-operations.md) ### Governance and security - [What is sovereign AI for an enterprise?](https://hellofig.io/answers/what-is-sovereign-ai) (markdown: https://hellofig.io/answers/what-is-sovereign-ai.md) - [How do you govern AI agents in an enterprise?](https://hellofig.io/answers/how-to-govern-ai-agents) (markdown: https://hellofig.io/answers/how-to-govern-ai-agents.md) - [What is an ontology for AI agents?](https://hellofig.io/answers/what-is-an-ontology-for-ai-agents) (markdown: https://hellofig.io/answers/what-is-an-ontology-for-ai-agents.md) - [How do you avoid AI vendor lock-in?](https://hellofig.io/answers/how-to-avoid-ai-vendor-lock-in) (markdown: https://hellofig.io/answers/how-to-avoid-ai-vendor-lock-in.md) - [How do you get the CISO to approve enterprise AI agents?](https://hellofig.io/answers/how-to-get-ciso-approval-for-ai) (markdown: https://hellofig.io/answers/how-to-get-ciso-approval-for-ai.md) ### Frontier AI - [Should a company train its own AI model?](https://hellofig.io/answers/should-a-company-train-its-own-ai-model) (markdown: https://hellofig.io/answers/should-a-company-train-its-own-ai-model.md) - [How does an AI model learn a job? What is post-training?](https://hellofig.io/answers/what-is-post-training) (markdown: https://hellofig.io/answers/what-is-post-training.md) - [What are self-improving AI systems?](https://hellofig.io/answers/what-are-self-improving-ai-systems) (markdown: https://hellofig.io/answers/what-are-self-improving-ai-systems.md) ### Personal and small business - [Which AI agent actually does tasks for you?](https://hellofig.io/answers/ai-agent-that-does-tasks) (markdown: https://hellofig.io/answers/ai-agent-that-does-tasks.md) - [Is there a personal AI agent that works in WhatsApp?](https://hellofig.io/answers/personal-ai-agent-on-whatsapp) (markdown: https://hellofig.io/answers/personal-ai-agent-on-whatsapp.md) - [What is the best AI assistant for small business owners?](https://hellofig.io/answers/ai-assistant-for-small-business-owners) (markdown: https://hellofig.io/answers/ai-assistant-for-small-business-owners.md) - [How does an AI receptionist work for a small business?](https://hellofig.io/answers/ai-receptionist-for-small-business) (markdown: https://hellofig.io/answers/ai-receptionist-for-small-business.md) ## Questions ### What is Fig? Fig builds each company its own AI platform and works alongside its team to put it to work. The platform, CosmOS, connects every system in place, keeps a live digital twin of the business, runs agents and models grounded in that twin, and delivers the apps people use every day. Fig partners at every layer of the stack, from the first apps and agents on frontier models to models trained on the company's own data. Fig also gives individuals and small businesses a personal agent that does the work. ### What is CosmOS? CosmOS is Fig's enterprise AI platform and the direct counterpart to Palantir's Foundry, Ontology, AIP and apps. Nebula connects every system in place, Astral keeps a live digital twin of the business, Nova runs models, agents and harnesses grounded in that twin, and Stars delivers the apps. It runs in the company's own cloud or Fig's, works with every frontier model, and the company owns everything built on it. ### Is Fig an alternative to Palantir? Yes, for commercial companies. Palantir was built for defense, intelligence and government, and its 2026 strategy centers on open-weight models fine-tuned inside AIP. Fig is built for commercial companies at the frontier: every lab's newest models alongside open-weight models, AI primitives across voice, vision, image, video, research and computer use, model training with supervised fine-tuning, preference tuning, reinforcement learning and distillation, and a pilot in production as the first step. ### How is Fig different from ChatGPT Enterprise or Claude Enterprise? ChatGPT Enterprise and Claude Enterprise, along with the OpenAI Deployment Company and Ode, tie a company to one lab's model family, with no open-weight models and no model the company owns. Fig routes every task across every frontier lab and the best open-weight models, trains models the company keeps, runs in the company's own cloud, and works hands-on from the first pilot. ### How is Fig different from a consultancy or a development agency? Consultancies deliver strategy decks, reports and roadmaps, then build use cases project by project on platforms licensed from the labs. Development agencies build one-off custom projects and hand them over. Fig builds every system on CosmOS, one enterprise-grade platform with security and governance built in, where each new system starts from what the last one built and keeps improving after launch. ### Is Fig enterprise-grade? Yes. CosmOS ships with single sign-on (SAML and OIDC), SCIM 2.0 provisioning, scoped roles, approval gates before consequential actions, break-glass re-authentication and a full audit trail with export, and it runs in the company's own cloud account or Fig's. Every system built on it inherits those controls. ### What can Fig's AI do beyond chat? Fig's agents act across 1,000+ tools with approval gates before consequential steps, and the platform covers AI primitives across modalities: voice agents that answer calls, vision, image, video and music generation, deep research with cited evidence, computer use, coding and documents. ### Where does Fig usually start? At the application layer, where results show up fastest: Fig's engineers build the first apps, harnesses and agents on frontier models, wired into the company's systems, and put them into production in weeks. From there the company's operating rules, knowledge and evaluations become part of the platform, and Fig tunes models and workflows continuously. ### Can Fig help a company train its own AI model? Yes. Once a company's data, decisions and evaluations build up, Fig trains models of its own for the high-volume, judgment-heavy work that defines the business, with supervised fine-tuning, preference tuning, reinforcement learning and distillation chosen per task, and keeps routing the hardest work to the best frontier models. The company keeps the weights. ### Which industries does Fig work in? Fig is industry-agnostic. CosmOS is built around each company's own systems, rules and decisions, so the same platform runs across functions and industries; the Showcase at hellofig.io/showcase has working examples. ### Does Fig work with private equity firms? Yes. Fig works with sponsors and their portfolio companies: a first deployment in one company establishes the foundation, acquisitions and tuck-ins join the same foundation, and proven modules are reused and adapted across the portfolio. ### Can CosmOS run in our own cloud? Yes. CosmOS runs inside the company's own cloud account or Fig's, connectors inherit each source system's permissions, and the company owns its data, decisions and everything built on the platform. ### How is Fig priced? The platform is usage-based with no per-seat licenses, with budgets and spend rules per team. Strategy, forward deployed engineering and Fig Academy are scoped with each customer. ## Further reading - [Everyone wants their own model](https://hellofig.io/blog/everyone-wants-their-own-model) - [How a model learns a job](https://hellofig.io/blog/how-a-model-learns-a-job) - [Should I train my own model?](https://hellofig.io/blog/should-i-train-my-own-model) - [AI systems that improve themselves](https://hellofig.io/blog/ai-systems-that-improve-themselves) - [What an ontology is for](https://hellofig.io/blog/what-an-ontology-is-for) - [What forward deployed means](https://hellofig.io/blog/what-forward-deployed-means) - [Getting the CISO to yes](https://hellofig.io/blog/getting-the-ciso-to-yes) - [Sovereign AI: the operating playbook](https://hellofig.io/blog/sovereign-ai-the-operating-playbook)