# Compozelabs > Official Compozelabs business agent. Compoze Labs designs and builds the software, data, and AI your business runs on, with one team that owns it from first… ## Ask it a question Compozelabs answers for itself at https://api.hailera.com/mcp/compozelabs. A client that speaks MCP can connect to that address; the tool is ask_compozelabs and the protocol is 2026-07-28. Where this file and the agent disagree, the agent is current. ## About Compozelabs Overview Compoze Labs is a Minneapolis-based team of strategists, designers, data engineers, AI specialists, and software engineers. The company designs and builds the software, data, and AI systems businesses rely on, with one team accountable from the first conversation through launch and beyond. What Compoze Values - People first: A culture of candor, collaboration, curiosity, and authenticity. - Delivery excellence: Work is measured by what ships and how well it performs in production, not by hours logged or presentations delivered. - Client success: Compoze aims to build lasting partnerships so clients can own and extend what they build together. Leadership - Paul Hilsen — Co-Founder and CEO - Andrew Larsen — Co-Founder and CTO - Eric Carr — President - Jeff Rogers — CDO Source: https://compozelabs.com ## What Compozelabs does Production AI Products Compoze Labs helps product and platform teams build AI assistants, agents, and knowledge systems intended for reliable everyday use. Its AI work spans three layers. AI Foundations - AI data strategy - Data ingestion pipelines and vector infrastructure - Embedding and retrieval architecture - Evaluation frameworks, regression tests, and observability - Prompt monitoring, audit trails, and governance AI Operating Model - AI advisory and strategy - Prioritization and build-versus-buy decisions - Architecture and operating model design - Fractional CAIO support - Organizational AI maturity assessments - Team enablement, workshops, documentation, and training AI Products - Multi-step agentic systems with orchestration, tool use, and error recovery - Knowledge and retrieval-augmented generation systems grounded in company content - AI assistants and copilots embedded in products or workflows Delivery Stages - Discovery and scoping: A 1–2 week assessment of data, workflows, and team readiness to define the smallest worthwhile system. - Build and evaluate: A 6–10 week phase producing working software instrumented and tested against evaluation sets. - Production and scale: Ongoing hardening, team training, monitoring, and governance. The website states that many engagements deliver a working version in 6–10 weeks, with production rollout in 4–6 months depending on integration complexity. Functional Applications Compoze designs AI for marketing and sales, customer service, operations, and product and engineering teams. Examples include lead research, campaign personalization, grounded support assistants, intake and routing automation, reporting, reconciliation, embedded product AI, and developer tooling. Example For Understood.org, Compoze built an AI assistant using retrieval-augmented generation against the organization's expert-reviewed content library. The assistant helps parents find guidance and gives the content team information about what families are searching for. Core Capabilities Compoze Labs provides connected capabilities across the modern technology stack: Application Development - Web platforms, customer portals, internal tools, and multi-tenant SaaS - Native and cross-platform mobile applications for iOS and Android - Cloud-native systems on AWS, Azure, and GCP - Custom software designed to scale with business growth - Managed services after launch Data Engineering Compoze builds the data layer that operations, analytics, and AI systems depend on. This includes: - Cloud-native data migrations and observability - Data pipelines and connections across systems - Semantic layers and data contracts - Governed structured and unstructured data for AI Systems Modernization Compoze modernizes legacy ERP and core systems incrementally. The approach includes keeping what works, upgrading what does not, avoiding big-bang cutovers, and building modular capabilities that can adapt over time. Application Architecture Applications can be designed for human and agent callers with: - Versioned, programmatic-first APIs - Event-driven patterns and auditable agent actions - Observability for non-deterministic behavior - Identity and access controls for people and agents - Zero-trust security practices Compliance and Security Compoze builds systems around the compliance posture required by the data and business context. The standards listed on the website include SOC 2, HIPAA, PCI-DSS, GDPR/CCPA, FERPA, and zero-trust security. Source: https://compozelabs.com ## Where Compozelabs works Lists Rogers as a location. Names Paul as a service area. Areas beyond these are not published. ## AI Engineering Helping Teams Become AI-Native Compoze Labs works with engineering teams that use AI coding tools but need shared standards, review patterns, governance, and measurement. Senior engineers co-build alongside client teams to make AI-assisted development sustainable and production-ready. Maturity Levels - Ad Hoc: Individual experimentation without shared tools, review patterns, or measurement. - Standardizing: Approved tools, basic policies, and the beginning of AI-aware code review. - Integrated: AI connected to CI/CD, tiered review, reusable prompts and skills, and measurable results. - AI-Native: Parallel AI coding sessions, repository configurations such as CLAUDE.md and AGENTS.md, workflow evaluations, autonomy scoring, and visible return on investment. Services - A two-week engineering maturity assessment and prioritized roadmap - Tooling audits and reviews of security and intellectual-property risks - AI coding standards and CI/CD quality gates - Custom skills and codebase-aware playbooks - Repository configurations that encode architecture and engineering patterns - Developer workshops and training - Senior engineers co-building with client development teams Measurement Compoze tracks measures such as cycle time, defect rate, AI-assisted code percentage, autonomy, standards coverage, and developer satisfaction to determine whether AI is improving engineering outcomes or creating technical debt. Source: https://compozelabs.com ## Delivery Approach and Client Fit How Compoze Works Compoze's engagements follow a full-lifecycle approach: - Understand: Clarify the business challenge, affected people, and definition of success through working sessions, technical due diligence, and a written brief. - Build: Design and develop solutions that fit existing operations, using short cycles and user feedback. - Ship: Deliver working software through predictable milestones, with testing, telemetry, security, evaluations, and documentation included in the delivery process. The company's approach emphasizes clear priorities, visible progress, measurable outcomes, and ownership that remains with the client. Documentation and runbooks support the handoff so the client team can continue developing and operating the system. Teams Compoze Works With - Teams launching something new: Founders and product leaders taking an idea from the whiteboard to market. - Teams outgrowing their systems: Organizations constrained by legacy tools and manual workarounds. - Teams putting AI to work: Leaders moving beyond demos toward production AI with appropriate guardrails. - Teams scaling under pressure: Companies whose data and platforms must keep pace with rapid business growth. Reported Results The services page reports: - 90% client retention rate - More than 200 successful projects delivered - 40% increase in operational efficiency - 3× faster time to market The homepage also reports more than $20 million in client e-commerce revenue on a platform Compoze built and more than $1 million in new merchandising agreements from a site it shipped. Source: https://compozelabs.com