The Capability Map
for Enterprise AI
Investment Intelligence.
IntakeOS is the system enterprises use to answer where they should apply AI and automation - and in what order. Every feature on this page is engineered around four principles: drive value, maximize efficiency, stay easy to use, and accelerate time to qualified decision.
Every feature exists to deliver on
four non-negotiable principles.
IntakeOS is built around four pillars that govern every product decision. If a feature doesn't sharpen investment decisions, save time, simplify work, or accelerate time to qualified opportunity - it doesn't ship.
Drive Value
Every feature is wired to an investment decision. Auto-calculated ROI, hours-recovered models, ranked opportunity alternatives, and portfolio cost-of-delay visibility make the AI investment decision the unit of work - not screens or stories.
See the business outcomes we optimize forMaximize Efficiency
Replace 8-14 days of fragmented discovery conversations with a 45-minute structured AI interview that yields a qualified, scored, and prioritized investment recommendation. One platform, zero rework.
Ease of Use
Business users chat - no training required. Administrators get one console for orgs, tech catalogs, scoring weights, and audit. Developers get OpenAPI specs and codegen. Everyone gets a guided onboarding flow.
Speed to Value
From the first organizational conversation to a qualified, defensible automation candidate - in 45 minutes. From that candidate to a deployed AI agent in production - in under a day. Decision velocity is a competitive advantage.
From organizational conversation to qualified decision - then to deployed agent - in under a day.
Traditional automation programs spend months moving from organizational conversation to a defensible investment decision. IntakeOS compresses that into a single 45-minute intake - structured, qualified, and portfolio-ready. Here is the actual sequence - every step is shipping software, not roadmap.
VARA opens an intake chat
Business user is invited via email - no training, no onboarding. Adaptive conversation begins.
Problem-First gates run
Problem statement validated. Root-cause question resolved. System-fix vs. automation path confirmed.
7-Pattern decision tree fires
Scoring engine identifies the recommended automation pattern, ranked alternatives, and developer complexity.
Executive report generated
Mermaid current + future state maps, ROI calculator, dimension scoring, AI narrative - all auto-produced.
Agent export ready
Click 'Export to Agent' - pick Claude, GPT, Gemini, UiPath, Copilot Studio, or native runner.
Agent live in production
Pre-built test harness validates, escalation routes are configured, audit logging is on. Value starts immediately.
Jump to any capability
Thirteen feature domains. Over 90 individually documented capabilities. All shipping in the platform today.
The intake conversation does the work that a six-person discovery team used to do.
VARA is IntakeOS's AI Business Analyst - the interview and discovery engine that systematically extracts operational intelligence from the people who actually understand the work. Any employee, any time, zero scheduling overhead. Every intake produces qualified, scored, portfolio-ready opportunity data that feeds directly into the platform's scoring and investment decision layer.
VARA - AI Business Analyst
IntakeOS's interview and discovery engine - a purpose-built AI analyst that conducts structured discovery conversations, capturing the workarounds, tribal knowledge, and shadow processes that system logs and documentation miss. Every finding feeds the platform's scoring, portfolio, and decision layer.
Conversational Chat UI
Business users see a chat. That's it. No forms, no fields, no training. They describe their work, VARA asks the right questions, and a structured intake assembles itself in the background.
Adaptive Question Flow
Question paths branch based on prior answers. If a process involves documents, VARA pursues the IDP sub-tree. If it spans systems, VARA pursues integration questions. No one-size-fits-all script.
Deep-Dive Modal
When the AI surfaces an ambiguity, business users get a one-click Deep Dive modal to clarify without losing context - keeping the main conversation focused while edge cases are captured.
Qualification Progress Tracker
A live progress panel shows which qualification gates have fired, which fields are filled, and what VARA still needs - turning an opaque conversation into a visible workflow.
Questionnaire Fallback
Power users and admins can switch from conversation to a structured questionnaire view to manually edit fields, override answers, or pre-fill from existing process docs.
SOP Document Attachments
Drag-and-drop existing SOPs, process docs, runbooks, or screenshots directly into the intake. Stored in object storage, linked to the intake, and surfaced to VARA as context.
Process Video Attachments
Upload screen recordings of the process as it's actually performed. VARA references them, the report links to them, and downstream agents can use them as ground truth.
Email Invitations
Admins invite business users to an intake with a single click. Branded invitation with one-click join link, sent through Resend with full deliverability tracking.
Automated Reminders
Stalled intakes auto-surface for follow-up. Send a one-click reminder with the latest progress summary embedded - no inbox archaeology required.
Multi-Persona Demo Workspace
Pre-seeded demo with Acme, Globex, XYZ Corp, and Meridian Health - each with admin and business user personas - so you can experience IntakeOS from every angle in minutes.
Guided Onboarding Flow
First-time admins walk through a structured onboarding wizard that configures their org, departments, technology catalog, and first intake invitation in under five minutes.
Run the interview where your people already work.
VARA doesn't require anyone to learn a new tool. The same adaptive intake interview - questions, qualification gates, and scoring underneath - can run natively as a Slack DM, a Microsoft Teams chat, or a threaded Outlook/email conversation, in addition to the web chat. Business users answer from whatever inbox or app they already have open; the intake, scoring, and report pipeline behind it is identical no matter which channel they used.
Native Slack Intake
VARA DMs the business user directly in Slack and conducts the full adaptive interview there - no separate tab, no login screen, replies happen right in the thread.
Microsoft Teams Chat Intake
The same intake conversation runs as a native Teams chat, so employees already living in Teams for their day-to-day work never have to leave it to complete a discovery interview.
Outlook / Email Threaded Intake
For users who prefer email, VARA runs the interview as a threaded Outlook/email conversation - each reply advances the intake exactly like a chat message would.
One Pipeline, Every Channel
Slack, Teams, email, and the web chat all feed the same qualification gates and deterministic scoring engine - the channel changes where the conversation happens, never the rigor behind the result.
A four-gate methodology that prevents automating the wrong thing.
Most automation programs skip straight to pattern matching - leading to expensive bots that paper over broken systems. IntakeOS bakes the consulting-grade Problem-First framework directly into the intake. Every opportunity is qualified through four explicit gates before any pattern is recommended.
Gate 1 - Problem Statement Validation
VARA first confirms the problem is defined, measurable, and mature enough to qualify. Vague pain points are clarified before any pattern conversation starts.
Gate 2 - Root-Cause Inquiry
Explicitly asks: 'Could this be solved by fixing the underlying system instead of automating around it?' If yes, the engine outputs SYSTEM_FIX scope - not a bot recommendation.
Gate 3 - 7-Pattern Decision Tree
If automation is appropriate, the engine walks the exact 7-pattern decision tree from the slides: System Integration, RPA, IDP, Low-Code BPM, Process Mining, Machine Learning, Generative AI.
Gate 4 - Developer Complexity Classification
After pattern selection, one closing question determines whether the solution fits a Citizen Developer scope or requires Expert Developer involvement - feeding directly into backlog refinement.
IDP Sub-Tree Qualification
When documents are in scope, a nested qualification tree fires: structured vs. unstructured, data extraction vs. classification, leading to IDP, RPA + BPM, or Generative AI depending on document characteristics.
Qualification Path Audit Trail
The exact sequence of gates that fired, with field values that triggered each gate, is captured on every report - auditable, defensible, and ready for any compliance or governance review.
System-Fix Prominence
When the root cause is a system issue, the report leads with that finding - a deliberate, amber-banded callout that prevents teams from automating their way around bad architecture.
Root Cause Type Tagging
Every intake is tagged as system_fix, automation, or unclear - letting CoEs see at a glance how much of their pipeline is actually automation work versus deferred IT modernization.
Pattern Rationale on Every Report
Selected pattern always comes with one-line rationale (why this pattern, why not the alternatives) plus ranked alternatives in case the primary choice hits an obstacle.
A deterministic rule engine the AI can explain but cannot override.
The scoring engine is the technical heart of IntakeOS. Eight hand-tuned rules run deterministically against every intake. The recommendation is the rule output - auditable, repeatable, and never hallucinated. AI is used to explain and enrich the result, never to choose it.
8 Deterministic Scoring Rules
Hand-tuned rules running against structured intake fields. Same inputs produce the same outputs, every time. No probabilistic recommendations on six-figure investments.
AI Cannot Override the Engine
AI generates the narrative, explains the choice, and surfaces context - but the recommended pattern is the rule engine's output. Auditable, defensible, deterministic.
Rules-Triggered Audit Log
Every rule that fired (or didn't) is captured with the field values that triggered it. Full transparency for compliance, governance, and CoE retrospectives.
Dimension Scoring
Each opportunity scores across multiple dimensions - Business Impact, Complexity, Volume, Data Quality, System Access - surfaced in a radar chart on the report.
Configurable Scoring Weights
Admins tune the relative weight of each scoring dimension to match their CoE's priorities - favor ROI-heavy work, complexity-light pilots, or strategic-fit opportunities.
One-Click Re-Score
Edit any intake field on the questionnaire and re-run the engine instantly. See exactly how the recommendation, scores, and ROI shift in response to new information.
Ranked Alternative Patterns
The report always shows ranked runner-up patterns with their scores - so if the primary recommendation hits a constraint, the backup path is already qualified.
AI Reasoning Narrative
After rules run, AI generates a plain-English explanation of why the chosen pattern fits - bridging the gap between deterministic logic and stakeholder communication.
Current-state and future-state diagrams, auto-generated.
Every qualified opportunity surfaces with current-state and future-state process diagrams synthesized directly from the intake conversation. These are not documentation artifacts - they are the analytical foundation for investment decisions, showing exactly where automation applies, where humans remain in the loop, and where exceptions escalate.
Auto-Generated Current State Map
Mermaid.js diagram of the process as it exists today - branching paths, exception handling, system touchpoints, decision points - synthesized from the intake conversation.
Auto-Generated Future State Map
Mermaid.js diagram of the optimized, automated process - showing where humans remain in the loop, where the agent or bot takes over, and where exceptions escalate.
Pain Point Annotations
Every step that surfaced friction during the intake gets annotated on the current-state map - making the cost of the current process visible at a glance.
Process Maps Workspace
Dedicated workspace to view, refresh, copy, and embed process maps. Export the raw Mermaid source for use in Confluence, Notion, or your own documentation.
Exception Path Mapping
Every identified exception type gets its own branch on the future-state map with clear escalation routing - making human-in-the-loop touchpoints explicit and designable.
Volume and SLA Overlays
Maps surface per-step volume and handling-time data captured during intake - instantly identifying the highest-value steps to automate first.
A board-ready report, generated in 30 seconds.
Every qualified intake produces a complete investment recommendation - qualification path, dimension scoring, ROI projection, pattern rationale, ranked alternatives, and AI-generated narrative - structured for the decision-makers who act on it. Defensible, auditable, and ready for a CFO or board review.
Auto-Generated Executive Report
Full recommendation document - qualification path, scoring dimensions, ROI projection, pattern rationale, AI narrative, process maps - generated end-to-end in under 40 seconds.
Built-in ROI Calculator
Calculates labor hours recovered, error-cost avoided, cycle-time compression, and full-year financial impact from the volume and handling-time data captured during intake.
Qualification Path Banner
Top-of-report banner makes the root cause finding impossible to miss - amber for system-fix scope, emerald for qualified automation opportunity.
Dimension Scoring Visualization
Radar-chart view of how the opportunity scores across Business Impact, Complexity, Volume, Data Quality, and System Access - at-a-glance triage for portfolio decisions.
Ranked Alternative Recommendations
Below the primary pattern, ranked alternatives with their own scores, rationale, and complexity classification - so the backup path is already qualified before it's needed.
AI Narrative Generation
Plain-English reasoning sections - business summary, pain points, recommendation justification, change management considerations - written by AI from the structured intake data.
Regenerate Narrative on Demand
If the AI provider hiccups or the narrative needs a refresh, one-click regeneration uses the latest intake data without re-running the rules engine.
Graceful Narrative Failure Mode
If AI fails, the deterministic scoring, qualification path, dimension scores, and ROI are preserved - only the narrative is missing, with a clear retry banner.
Re-Score from Report
Admin re-score button on the report header re-runs the engine against the latest stored intake fields without touching the AI narrative - instant qualification refresh.
Report Generation Timestamps
Every report shows when it was scored and by which run - full provenance for the recommendation, ready for any audit or governance review.
One intake, every format the rest of your org needs.
Reports don't live in one place. Engineers need XML. Executives need PDF. Operators need Markdown. IntakeOS exports every intake into the format each stakeholder needs - without re-keying, re-formatting, or re-explaining.
The Agent Bridge goes both ways. Agent-ready exports built for 20+ agentic tools - and VARA as a callable tool so agents qualify work before they build. Explore the Agent BridgeBranded PDF Export
Polished, branded PDF of the full executive report - cover page, ROI summary, qualification path, dimension scoring, process maps, AI narrative - ready to email to a sponsor.
Markdown Export
Clean Markdown export of the entire report - paste directly into Confluence, Notion, Linear, GitHub, or any markdown-native tool without reformatting.
ProcessDefinition XML Export
Structured XML export of the process definition - compatible with UiPath, Microsoft Power Automate, and any downstream RPA or workflow platform expecting standard process schemas.
Agent Environment Exports
One-click exports tailored to each major AI agent runtime - Claude Projects, OpenAI Assistants, Gemini Workbench, UiPath Autopilot, Copilot Studio, IntakeOS Native Runner.
Public Object URLs
Attachments referenced in reports get stable public or private object storage URLs - embedding cleanly in exported documents, agent contexts, or downstream systems.
Self-Contained Export Bundles
Every export includes everything a downstream consumer needs - process flow, business rules, exception handling, system specs, SLA parameters, escalation routing.
From qualified intake to deployed autonomous agent in hours.
The newest and most powerful capability in IntakeOS. Every qualified intake can be exported directly into the major AI agent runtimes - Claude, GPT, Gemini, UiPath, Microsoft - as a complete, executable agent context. No development cycle. No handoff document. The intake is the deployment artifact.
It starts with finding the right work. Deployment is only half the story - VARA first uncovers the processes agents can own end-to-end. See how we uncover agentic outcomesAnthropic Claude Projects Export
Structured project context that loads directly into Claude's project memory - full process definition, business rules, exception handling, escalation protocols.
OpenAI Assistants Export
JSON assistant configuration with tool definitions and instructions - deployable via the Assistants API in minutes with a pre-built test harness.
Google Gemini Workbench Export
Vertex AI agent configuration with Grounding, Function Calling, and Extensions pre-configured based on the technology catalog.
UiPath Autopilot Export
Hybrid process definition combining AI reasoning for judgment-heavy exception handling with traditional RPA execution for structured data operations.
Microsoft Copilot Studio Export
Power Platform solution package importable directly into your tenant - pre-configured Power Automate flows, Copilot Studio topics, and Dataverse audit logging.
IntakeOS Native Agent Runner
Managed runtime that executes process definitions using your configured AI provider - scheduling, logging, exception routing, and performance monitoring included.
Pre-Built Validation Test Harness
Every agent export ships with sample test cases derived from the intake - validate behavior end-to-end before promoting to production.
Agent Performance Monitoring
Dashboard tracks straight-through rate, exception volume, escalation patterns, and ROI realization against intake projections - closing the feedback loop.
Agent API Key Management
Dedicated admin console for managing API keys, rate limits, and access scopes for every connected agent runtime - one source of truth for credentials.
Intake Agent Customization
Tune VARA's tone, depth, and pursuit aggressiveness per organization - making the intake conversation match your culture and consulting style.
See your entire automation portfolio in one view.
Most CoEs operate blind - opportunities scattered across spreadsheets, prioritization happening in committee meetings, and the value of work-in-progress invisible. IntakeOS makes the entire pipeline visible, value-quantified, and continuously re-rankable.
Unified Backlog View
Every intake - qualified, in-progress, or completed - in a single ranked, filterable, sortable view. The CoE's single source of truth for the opportunity pipeline.
Smart Filters & Search
Filter by status, department, pattern, developer complexity, ROI tier, or root-cause type. Find the work that matches available capacity in seconds.
Pipeline Value Visibility
See the cumulative projected ROI of opportunities at every stage of the pipeline - quantifying the cost of delay across the entire backlog.
Continuous Re-Ranking
As scoring weights change or new intakes complete, the backlog re-ranks automatically - keeping prioritization current without quarterly committee cycles.
Developer Complexity Sorting
Sort by Citizen vs. Expert developer scope to match opportunities to the right delivery team - citizen-developer queues for Power Platform, expert queues for custom builds.
Department & Org Roll-Ups
Roll up pipeline value by department, business unit, or organization - surfacing where automation investment is concentrated and where the gaps are.
One console for every platform configuration decision.
The admin console is where your CoE configures, governs, and audits the platform. Organizations, departments, users, technology catalogs, scoring weights, audit logs, knowledge bases - every operational lever is in one place, with role-based access and full audit trail.
Multi-Organization Management
Manage multiple tenant organizations from a single admin surface - for consulting practices, enterprise CoEs serving multiple business units, or platform operators.
Departments & Hierarchy
Model your organizational hierarchy explicitly - departments roll up to orgs, users belong to departments, and intake routing follows the structure.
User & Role Management
Admin and Business User roles out of the box. Invite users, assign departments, manage roles, and revoke access - all from a single console.
Technology Catalog
Configure the organization's actual technology stack - RPA platforms, integration tools, AI services, document platforms - so pattern recommendations reflect what you can actually deploy.
Scoring Weights Configuration
Tune the engine's dimension weights to match your CoE's strategic priorities - favor business impact, complexity, ROI tier, or any combination.
Comprehensive Audit Log
Every meaningful action - intake created, scored, exported, user invited, scoring weight changed - captured with actor, timestamp, and full context. Compliance-ready, forever.
Knowledge Base Management
Curate the AI's context - process patterns, naming conventions, organizational rules - feeding directly into VARA's intake conversations and report narratives.
Intake Agent Tuning Console
Admins configure VARA's interview behavior per-org - depth, tone, pursuit thresholds, IDP sub-tree triggers, developer complexity heuristics.
Agent API Keys
Manage outbound API credentials for connected agent runtimes - issue, rotate, scope, and revoke keys from a single console with full audit.
Licensing & Plan Management
Plan tier, seat counts, and feature entitlements visible and adjustable from the admin console - billing transparency without leaving the platform.
Internal Staff Portal
IntakeOS staff portal at /internal for platform operators - tenant overview, support tooling, and operational metrics across the entire fleet.
Bring your own AI. Plug into the systems you already use.
IntakeOS is provider-agnostic by design. Use OpenAI, Anthropic, or Gemini for the conversational AI layer. Plug into Resend for email, your own object storage for attachments, and the agent runtime of your choice for deployment. No vendor lock-in.
AI Provider Abstraction
Single AI_PROVIDER env var routes to OpenAI, Anthropic Claude, or Google Gemini - switch providers without touching application code.
Graceful Provider Fallback
If no AI key is configured or the provider fails, the platform falls back to rule-based question prompts - never breaks, never stops collecting intake data.
Resend Email Integration
Out-of-the-box email delivery via Resend - invitations, reminders, and report shares all routed through a production-grade email infrastructure.
Google Cloud Object Storage
Attachments stored in GCS via presigned upload URLs - secure, scalable, and decoupled from the application server for fast, reliable file handling.
Contract-First OpenAPI
Every API endpoint defined in OpenAPI first - codegen produces React Query hooks and Zod schemas, eliminating drift between client and server.
Agent Bridge Integrations
Agent-ready spec packs built for 20+ agentic tools — Claude Cowork, Claude Code, ChatGPT Agent, Gemini, Copilot Studio, Devin, Cursor, LangGraph, and more — with VARA exposed as a callable MCP tool for connected agents. Rolling out through early access.
PostgreSQL Data Layer
Strongly-typed Drizzle ORM over PostgreSQL - your data lives in a standard, queryable, exportable database you can bring with you.
Self-Hosted or Managed
Deploy IntakeOS in your own cloud account for full data sovereignty, or use the managed service for zero-ops convenience.
Built for the way CoEs actually work together.
Automation programs are team sports. Business users describe processes, admins qualify and prioritize, developers deliver, and sponsors review. IntakeOS makes every handoff frictionless - no email threads, no version-control gymnastics, no lost context.
One-Click Intake Invitations
Admins invite business users to a specific intake - branded email, one-click join link, no account creation friction on the recipient side.
In-App Comments & Mentions
Leave comments directly on intakes and reports, mention team members, and keep discussion attached to the opportunity, not buried in Slack.
Secure Report Sharing
Generate secure, time-limited viewing links for executive reports to share with sponsors without requiring them to log into the platform.
In-App Notifications
Get alerted when an intake completes, when someone comments, or when an agent export succeeds. Centralized notification hub keeps work moving.
Every feature.
One day to value.
Spin up a fully-seeded IntakeOS workspace, run a VARA intake on a real process, and see your first qualified opportunity in under an hour. No credit card. No installation.