Two layers, one platform — out-of-the-box capabilities for AI-driven teams, and an accelerator that lets Kairoi ship purpose-built solutions in days and weeks, not quarters. Configured as CompassPE for private equity; as Compass AI for portfolio companies and operators.
Why Compass AI exists
Compass lets Kairoi deliver AI solutions faster, higher quality, and at lower cost than even AI-empowered traditional development teams. Teams don’t just ask questions. They design structured workflows, build custom AI specialists, generate production-quality documents and financial models, and monitor portfolio performance across every company they manage.
The platform is structured, governed, and model-agnostic. It runs on Anthropic (Claude), OpenAI, Google (Gemini), and open-source models, with Microsoft Azure and AWS as deployment targets where customer environments require it. The right model for each task, governed centrally, upgraded as the frontier moves. No lock-in. No inherited constraints.
Compass AI is service-delivered. We deploy and customize it for and with our clients — as CompassPE for private equity firms, and as Compass AI for portfolio companies and mid-market operators. It is not a self-serve SaaS subscription.
Two layers, one platform
Layer 1
A set of out-of-the-box capabilities that work the moment a workspace is provisioned. Custom AI assistants, intelligent spaces, portfolio company intelligence, governance, multi-model management, and storage / data connectivity. The foundations.
Layer 2
A library of solution patterns Kairoi engineers build on Compass, quickly, securely, at fraction-of-traditional cost. Executive Talent Search systems, PE Command Centers, production-planning copilots with what-if conversational AI, embedded AI assistants threaded through workflows. Hard to build alone. Feasible on Compass.
Platform Capabilities
A collaborative conversation model that asks clarifying questions and adapts to your work. Custom Assistants encode AI specialists for any role. Intelligent Spaces are domain-specialized environments that organize conversations and surface the right knowledge. Together, they turn a generic LLM into a focused, governed work surface.
A structured intelligence layer around each portfolio company. Connect board decks, financial reports, and management presentations. Dynamically captures revenue, EBITDA, leadership, risks, and active initiatives. Role-based lenses shape the analysis for Investment Leads, Operating Partners, CFOs, and Analysts.
Enterprise-grade access control with role-based permissions enforced across the platform. Multi-tenant isolation. Centralized Model Management routes work across Anthropic (Claude), OpenAI, Google (Gemini), and open-source models, with Microsoft Azure and AWS as deployment targets where customer environments require it. Intelligence-tier controls and workspace-level governance throughout. Governance is the membrane the whole platform operates within.
Compass AI connects to the systems your teams already use: SharePoint, Google Drive, Snowflake, BigQuery, Salesforce, SAP, HubSpot, and external APIs. Universal connectors with dynamic credentials. No custom glue code. No integration debt.
Platform-Enabled Solutions
Examples of what Kairoi delivery teams build on top of the platform, typically inside a 30-100 day AI Implementation engagement. Anonymized examples drawn from real engagements.
Full-firm command centers built for a specific PE firm: drill-down reporting, portfolio-by-portfolio dashboards, distinct fund and sector lenses, LP-facing views. Often started by extending a firm’s existing monthly reporting platform.
A complete AI-augmented executive search system: job descriptions as living source-of-truth, scorecards and interview guides derived automatically, blind evaluation gating, candidate pipelines that compound across searches.
A planning copilot built directly over production data, sales, inventory, orders, and external signals. Conversational what-if scenario planning replaces the spreadsheet-and-gut process with a system that reasons over operational reality.
Purpose-built assistants threaded through specific operational workflows: close-process financial reporting, deal-screening research, daily inventory and ordering. Each one encodes domain knowledge and operates within the team’s data boundaries.
Why these aren’t generic widgets
Each solution composes Compass capabilities (assistants, spaces, governance, model routing, data connectivity) with bespoke business logic and a fast-paced AI-augmented build process. From scratch, any of these would be a multi-quarter engineering project. On Compass, they are 4-12 weeks of focused delivery, and they keep evolving as the platform evolves.
CompassPE: Compass AI configured for private equity
CompassPE is Compass AI configured for private equity firms — a service-delivered intelligence platform, customized to your firm and bundled with the advisory work that makes it productive. It connects to your SharePoint, indexes your institutional knowledge, and executes complex multi-step workflows across the investment lifecycle.
It is not a chatbot. It is not a search tool. It operates like a specialized analytical team embedded inside the firm.
A generic LLM like ChatGPT or Claude is powerful. It also has no idea your firm exists. It has never read your deal memos. It doesn’t know your investment thesis, your portfolio companies, your historical underwriting decisions, or the lessons you’ve learned over twenty years.
CompassPE is built behind your firewall and connected to your SharePoint. When a partner asks whether a portfolio company has outstanding debt, the system doesn’t search the internet. It reads the loan agreement, checks the latest balance sheet, cross-references the debt schedule, and answers from your own files.
That capability (secure, firm-specific, document-grounded) can’t be replicated by asking a generic LLM to help draft something. The grounding isn’t a feature on top of the model. It’s the entire architecture around it.
Each one is shaped by what actually happens inside an investment lifecycle, not by what a generalist assistant might guess at.
The foundational models (GPT and Claude and Gemini) are increasingly commoditized. The competitive moat lives in what wraps around them: your firm’s historical deal notes, your specific investment thesis, your portfolio company data, and the PE-specific workflows that turn raw model capability into something actionable.
That context takes time to build correctly. It can’t be bought off the shelf. The firm that builds it now will compound that advantage over every subsequent fund cycle.
We are not asking you to throw away what you’ve already built. If your team has developed custom GPTs or internal prompting frameworks, that work doesn’t have to be abandoned. We incorporate it into the CompassPE environment so prior investment isn’t lost.
Pilots run in days. Full deployment in weeks.
How Compass AI Works
Model-agnostic routing: the right model for the right task, with no lock-in. Runs on Anthropic (Claude family), OpenAI (GPTs), Google (Gemini, multimodal), and open-source models (Llama, Mistral), with Microsoft Azure and AWS as deployment targets where customer environments require it. An intelligence-tier system (Thinking, Normal, Fast) makes model selection intuitive for non-technical users while giving administrators governance controls.
Orchestration and cloud execution. Assistants with guided workflows and personas. Persistent context workspaces. Controlled tool loops. Reusable skills. Adaptive user intelligence. Content creators for documents, spreadsheets, and presentations. The composable surface that lets the platform-enabled solutions described above land in days and weeks rather than quarters.
MCP-native connectivity reaches into the systems clients already run. No custom glue code, no integration debt. Universal connectors with dynamic credentials. The integrations list below shows what’s wired today.
Governance wraps the entire architecture, not a feature added later, but the membrane the whole platform operates within. Data boundaries, access controls, cost structure, and prompt integrity are enforced at every layer. The result: every outcome delivered through Compass AI is reliable, scalable, and tech-debt free.
Integrations
An intelligence layer on top of systems you already run.
Document management
SharePoint, Google Drive, local file systems
Data warehouses
Snowflake, BigQuery
CRM / ERP
Salesforce, SAP, HubSpot
External APIs
Financial data providers, custom integrations
Custom SaaS
Shopify, ShipStation, proprietary platforms
Built for enterprise requirements
— Role-based access control with granular, workspace-scoped permissions
— Feature-level rollout management for gradual capability deployment
— Audit logs of every model call: provider, model, tokens, and cost
— Model governance with workspace-level tier floors and defaults
— Storage connector framework with dynamic credentials and circuit-breaker protection
Compass AI is not itself a SaaS platform. We adopt and align with the best practices of the certification frameworks our clients require, and we focus on meeting your specific requirements in the solutions we deploy for you.
See also: Kairoi’s firm-level Trust Center → for vendor inventory, architectural commitments, and the gaps we’re honest about.
Schedule a briefing with our team to see how Compass AI works inside firms like yours.
With your OK, we'll use Google Analytics cookies to see which pages people read and how they found us. We don't use them for ads. Privacy policy