Unified Digital Workspace Platform
“Clarity in the chaos of modern work.”
View Source Code on GitHubExecutive Summary
Itus AI was a Charlotte, NC-based AI startup (2023–2026) that built a productivity platform designed to unify the fragmented digital workspace. The platform aggregated data across communication tools, file systems, and scheduling platforms into a single AI-powered workspace — giving employees a streamlined experience and giving leaders visibility into operational patterns.
The platform primarily served healthcare SMBs, delivering unlimited access to premium AI models with a privacy-first architecture that ensured customer data was never used to train external AI models.
Beyond its core market, the platform gained academic recognition when it was featured in the Harvard Journal of Sports and Entertainment Law (Volume 16, Issue 1, 2025), where it was used in a published experiment exploring AI's role in patent obviousness analysis — highlighting the platform's versatility beyond its original market.
The Product
Screenshots of the Itus AI platform as it appeared during its active period. Click any image to view full size.
The core conversational AI workspace. Personalized greeting, suggested prompts, conversation history in the sidebar, and quick-action buttons for common tasks like email summarization and message rewriting.
The daily briefing view. An AI-generated Daily Summary and prioritized Action Items gave users a single starting point for their workday — pulling context from email, calendar, and chat.
The organizational admin view. Activity logs, usage insights, user management, and platform settings — giving leaders full visibility into how their team used the platform. (Activity log blurred for privacy.)
The public-facing site at itus.ai that introduced the platform’s value proposition, features, and pricing to prospective customers.
Company Overview
| Company Name | Itus AI LLC |
| Founded | 2023 (closed February 2026) |
| Headquarters | Charlotte, North Carolina |
| Structure | For-Profit LLC |
| Website | itus.ai (no longer active) |
| Platform | app.itus.ai (no longer active) |
| Contact | support@itus.ai |
| Itus AI |
Team
Joshua Beron
Co-Founder & CEO
Analytics and product development background. Led business strategy, product vision, and go-to-market.
LinkedIn →
Ari Bailey
Co-Founder & CTO
Software engineering background. Built the full-stack platform, cloud infrastructure, and Microsoft integrations.
LinkedIn →The co-founders are cousins who combined complementary business and technical expertise to build Itus AI from the ground up. What started as evenings-and-weekends work in late 2023 — sparked by a dinner conversation about the AI adoption gap in SMBs — grew into a launched platform with Microsoft integrations, academic recognition, and press coverage.
Origin Story
The idea for Itus AI was born over a dinner conversation in late 2023 between cousins Joshua Beron and Ari Bailey. Both had been tracking the AI landscape for years, and they noticed a striking pattern: the rapid adoption of tools like ChatGPT — which had reached 100 million monthly users in just two months after launch — was happening largely on personal accounts for work-related purposes.
This created a dilemma for employers, particularly small-to-medium sized businesses: either permit unregulated use of AI (risking security and privacy issues) or completely prohibit AI usage (forgoing the benefits of increased efficiency). There was no middle ground.
With Joshua's background in analytics and product development and Ari's skills in software engineering, they committed their evenings and weekends to building what they initially named “Itus” — an easy-to-use platform designed for managing AI interactions within organizations.
What We Built First
The initial product focused on two core components:
Personalized Chat System
Tailored to each individual employee, taking into account their specific role and integrating elements of their organization's framework.
Administrative Portal
Providing comprehensive analytics for detailed performance insights and recommendations.
Early Validation: Rev1 Ventures
The founding team participated in Rev1 Ventures' Customer Learning Lab, a collaborative environment with fellow entrepreneurs. The experience shaped two foundational lessons:
- Build after validation, not before — The team initially developed an MVP without fully verifying market need. This taught them the critical importance of aligning solutions with genuine user needs before investing in feature development.
- Focus on a niche — Their initial approach cast too wide a net. They learned that focusing on a specific market segment was essential for creating meaningful, effective solutions.
These early lessons led the team to pivot from heads-down feature delivery to meaningful conversations with potential customers, ultimately shaping the platform into the focused, privacy-first unified workspace it became.
Timeline
Founded
Born from a dinner conversation between cousins; evenings-and-weekends development begins
Customer Discovery
Participated in Rev1 Ventures' Customer Learning Lab; pivoted to customer-first approach
Platform Launch
Core platform launched with AI chat, role-based assistants, and Microsoft integrations
Press Coverage
Featured in Charlotte Business Journal
Academic Recognition
Cited in Harvard Journal of Sports and Entertainment Law (Vol. 16, Issue 1)
Insights Platform
Analytics engine built and demonstrated, grounded in Lean Six Sigma and behavioral design
Company Closed
Platform sunset after two and a half years of building, learning, and shipping
Problem & Solution
The modern knowledge worker operates across a sprawling ecosystem of disconnected tools — email clients, messaging platforms, cloud drives, calendars, and project management systems. This fragmentation creates several critical challenges:
- Context switching erodes focus and productivity as employees jump between 5–10+ platforms daily
- Information silos prevent teams from accessing the data they need when they need it
- Operational blind spots leave leadership unable to identify inefficiencies, bottlenecks, or underutilized resources
- AI adoption barriers prevent non-technical teams from benefiting from AI advancements due to complexity and privacy concerns
- Data privacy risks arise when employees use consumer AI tools that may train on sensitive business data
How Itus AI Solved It
Itus AI addressed these challenges by building a workspace that securely connected an organization's existing tools and surfaced relevant information through AI assistants.
Unified Workspace
Aggregated data from email, messaging, file storage, and calendars into a single workspace.
AI-Powered Prioritization
Helped employees focus on what mattered most by surfacing and prioritizing tasks.
Operational Intelligence
Generated quarterly analytics reports giving leadership visibility into operational challenges.
Privacy-First Architecture
Engineered so that AI models did not train on customer data.
Unlimited AI Access
Users had no usage caps or throttling on AI model access.
Accessible to Non-Technical Users
Designed so any team member could use it, regardless of technical background.
Platform Features
AI Chat & Assistants
- Customized Chat Experience — AI conversations tailored to individual roles and challenges
- Organizational Chat Sharing — Collaborative dialogue tools for team-wide knowledge sharing
- Integrated Web Search — Real-time information retrieval within the chat interface
- Unlimited Usage — Freedom to interact without constraints or degradation
- File Upload Capabilities — Analyze documents and files without AI data training
- Multi-Model Access — Access to premium AI models including OpenAI's latest offerings
Role-Based AI Assistants
- Role-specific system prompts — Each assistant was pre-configured with the employee's job title, department, and organization context so responses were immediately relevant
- Onboarding customization — Admins defined role descriptions and organizational priorities that shaped each assistant's behavior
- Suggested prompts by role — The interface surfaced role-appropriate starting prompts (e.g., email summarization for operations staff, policy lookups for compliance teams)
- Shared organizational context — Assistants drew from company-wide context so answers aligned with internal terminology and processes
Analytics & Reporting
- Quarterly Analytics Reports — Comprehensive operational analysis delivered to leadership
- Operational Challenge Identification — Highlights bottlenecks, inefficiencies, and areas for improvement
- Actionable Recommendations — Data-driven guidance for strategic decision-making
- Performance Insights — Visibility into team productivity and collaboration patterns
Insights Platform (Analytics Engine)
The Insights Platform was the leadership-facing analytics engine — an executive dashboard that transformed internal AI chat usage data (“chat exhaust”) into actionable operational intelligence. The core premise: every question employees ask an AI reveals friction points, knowledge gaps, and process inefficiencies that leadership can act on.
The Vision
Using AI not just as a productivity tool, but as an organizational sensor. The platform demonstrated that the questions people ask reveal as much — or more — about an organization than the answers they receive. It turned the “exhaust” from AI conversations into a decision intelligence layer that helped leadership see what’s really happening on the ground: the questions people are afraid to ask, the processes that create confusion, and the opportunities hiding in everyday friction.
Key Features
AI-Powered Insight Generation
Used OpenAI to analyze chat logs and automatically categorize them by business function, identify root causes, and suggest interventions based on Lean Six Sigma principles (identifying waste types: Waiting, Defects, Overprocessing, etc.).
Behavioral Design Principles
Applied Rory Sutherland’s behavioral psychology concepts: loss aversion framing (hours/week lost, $/month impact), commitment devices (“Own This Outcome”), visual urgency indicators, and defaults as direction (pre-selected recommended actions).
Cross-Filterable Dashboard
Interactive charts and metrics updating in real-time. Filterable by topic, date range, or role. Loaded with nearly 3,000 real chat records covering almost 2 years of organizational conversations.
Deep Dive Briefings
One-click executive summaries with root cause analysis, behavioral friction indicators, affected roles, and recommended interventions with impact justification.
State Expansion Planning
Comprehensive market analysis tool with 50-state coverage, auto-generated opportunity scoring, pipeline tracking (Target → Research → Evaluation → Active), interactive KPI cards, and export/share functionality.
Insights Platform Tech Stack
| Component | Technology |
|---|---|
| Frontend | React + TypeScript |
| Backend | Express |
| Database | PostgreSQL |
| AI | OpenAI API |
| Styling | Tailwind CSS + shadcn/ui components |
| Animations | Framer Motion |
Technical Architecture
Itus AI was built on a modern, secure cloud infrastructure designed for reliability, performance, and data privacy.
Core Platform Stack
| Component | Technology |
|---|---|
| Frontend | Svelte |
| Backend | SvelteKit |
| Database / ORM | PocketBase |
| Component Library | Skeleton |
| AI | OpenAI API |
| Deployment | Dokku on Hetzner VPS |
Infrastructure & Services
| Component | Provider / Technology | Purpose |
|---|---|---|
| Cloud Hosting | Hetzner Online GmbH | VPS hosting and storage services |
| CDN & Security | Cloudflare, Inc. | TLS protection, DDoS mitigation, edge CDN |
| AI Models | OpenAI, L.L.C. | Large language model processing |
| Email Services | MailerSend | Transactional email communications |
| Error Tracking | Sentry | Error tracking and handling services |
Architecture Principles
- Privacy by Design — AI models engineered to never train on customer data
- Organic Chat Flows — Continuous, natural conversation experiences without data leakage
- Secure Data Aggregation — Encrypted connections to all integrated workplace tools
- Scalable Infrastructure — Cloud-native architecture supporting startups to enterprises
- Browser-Based Access — No software installation required
Integrations
Itus AI integrated directly with Microsoft Outlook and Microsoft Teams — the core communication tools used by its healthcare SMB customers — to create a unified data layer:
| Integration | Category | Capabilities |
|---|---|---|
| Microsoft Outlook | Email aggregation, calendar sync, contact access | |
| Microsoft Teams | Communication | Team messaging, channel data, collaboration context |
Planned / In Development
| Integration | Category | Capabilities |
|---|---|---|
| Slack | Communication | Workspace messaging, channel history, file sharing |
| Google Drive | File Storage | Document access, file management, shared drives |
Integration Philosophy
- Secure connection protocols — All integrations used encrypted, authenticated connections
- Read-focused access — Built to surface and organize information, not modify source data
- Contextual aggregation — Combined data from multiple systems into a unified view
- Minimal permissions — Requested only the access necessary for each integration's functionality
Security & Data Privacy
Data security was a foundational principle, not an afterthought.
Privacy Commitments
- Customer data was never used to train external AI models
- Encryption for data in transit and at rest
- Subprocessor transparency with a published list of all data handlers
- Commercially acceptable security measures for personal data protection
Subprocessors
All third-party data handlers were disclosed transparently. See Technical Architecture for the full infrastructure stack.
Business Model & Pricing
Itus AI operated on a transparent, subscription-based model with no hidden fees.
Monthly
No annual commitment. Full access, cancel anytime.
Annual
Annual commitment. 17% savings over monthly plan.
Free Trial
Full platform access for evaluation. No commitment.
Every Plan Included
- Unlimited access to premium AI models
- Workspace integrations (Outlook and Teams; Slack and Google Drive were planned)
- Quarterly analytics reports with operational insights
- Comprehensive onboarding and training
- Dedicated support
- Role-based AI assistant customization
Target Market & Industry Applications
Primary market: healthcare small-to-medium businesses (SMBs) seeking to improve workplace productivity without the complexity and cost of enterprise-grade solutions.
| Industry | Focus | Use Cases |
|---|---|---|
| Healthcare | Primary | Staff coordination, operational efficiency, compliance workflows |
| Education | Secondary | Administrative productivity, cross-department communication |
| IT Staffing | Secondary | Candidate pipeline management, team coordination |
| Product Development | Secondary | Sprint management, cross-functional collaboration |
Market Context
The following industry data points informed the product thesis and go-to-market strategy. These are third-party research findings, not Itus AI metrics:
Generative AI boosts skilled workers' output by 40%.
MIT Sloan Management Review, 2023Expected generative AI to play a prominent role in their organizations.
Salesforce, 2023Believed AI would help increase their productivity.
Forbes Advisor, 2023Were already using leading AI services independently.
Exploding Topics, 2024Product Roadmap (Planned)
The following features were in development or planned prior to shutdown:
Dedicated File Repositories
Team-specific document management and collaboration tools with tailored knowledge repositories.
AI Education Experience
Personalized learning modules to elevate team AI fluency and analyze usage patterns.
Multi-Model AI Capabilities
Expanded model options beyond OpenAI for flexibility across different use cases.
Recognition
Harvard Journal of Sports and Entertainment Law
Itus AI was featured in the Harvard Journal of Sports and Entertainment Law (Volume 16, Issue 1, 2025) in a peer-reviewed article by Professor Max Stul Oppenheimer titled “The Artificial Intelligence Solution to the Patent Obviousness Problem.”
The Research
The article addresses one of patent law's most persistent challenges: determining whether an invention is “obvious” under 35 U.S.C. § 103. Obviousness is the most common reason patent claims are rejected, yet the standard remains highly subjective — prone to hindsight bias and inconsistency across examiners and courts.
Professor Oppenheimer proposed using AI tools to introduce objectivity and predictability into the obviousness analysis. The article outlines the history and challenges of obviousness determinations, proposes AI as a solution to reduce subjectivity, and demonstrates the feasibility of the approach through an experiment.
The Experiment
The Itus AI platform was put to the test in an experiment using the landmark Graham v. John Deere Co. case — one of the foundational Supreme Court decisions establishing the modern framework for patent obviousness. The results showed how AI can bring more consistency and clarity to patent analysis, demonstrating the practical feasibility of integrating AI tools into the patent prosecution process.
Significance
- The platform's AI capabilities were validated beyond workplace productivity, demonstrating versatility in specialized domains like legal analysis
- Being featured in a Harvard-published journal provides peer-reviewed academic recognition of the platform's technology
- The research contributes to the broader conversation about AI's role in improving objectivity within complex professional decision-making
Full Citation: Oppenheimer, Max Stul. “The Artificial Intelligence Solution to the Patent Obviousness Problem.” Harvard Journal of Sports and Entertainment Law, Vol. 16, Issue 1 (2025), pp. 151–182.
Read the full article →
Press
- Charlotte Business Journal (December 2024) — Featured article: “AI startup aims to transform businesses in Charlotte”
Lessons & Reflections
What Itus AI Demonstrated
- Privacy-first AI is viable — It's possible to deliver powerful AI capabilities while maintaining strict data privacy, proving that businesses don't have to choose between intelligence and security.
- SMBs need AI, not complexity — Small and medium businesses are eager to adopt AI but need solutions that abstract away technical complexity and integrate with existing workflows.
- Operational transparency drives better decisions — Quarterly analytics reports that surface operational challenges gave leaders visibility they previously lacked.
- Unlimited access builds trust — Offering AI model access without usage caps or throttling encouraged deeper platform adoption.
- Unified workspaces reduce friction — Aggregating data across email, messaging, file storage, and calendars meaningfully reduced context switching.
- AI platforms have cross-industry potential — The Harvard Journal citation demonstrated that a platform built for workplace productivity could deliver meaningful value in specialized domains like patent law.
Key Achievements
- Built and launched a fully functional AI productivity platform from concept to production
- Secured integrations with Microsoft Outlook and Teams (Slack and Google Drive planned)
- Developed a privacy-first architecture ensuring zero AI training on customer data
- Designed and implemented role-based AI assistants with organizational chat sharing
- Created a quarterly analytics reporting system delivering actionable operational insights
- Established transparent pricing with unlimited AI model access
- Cited in the Harvard Journal of Sports and Entertainment Law for demonstrating AI's feasibility in patent analysis
- Featured in Charlotte Business Journal as a notable local tech startup
- Deployed a scalable, cloud-native infrastructure on Hetzner with Cloudflare security
Why We Shut Down
There was no single failure — it was the compounding weight of several realities:
- Market timing — While SMBs recognized the value of AI, many weren’t ready to commit budget to a new platform when free tools like ChatGPT were “good enough” for their immediate needs.
- Sales cycle length — Healthcare organizations move slowly. The compliance reviews, procurement processes, and stakeholder buy-in required to close deals stretched timelines far beyond what a bootstrapped startup could sustain.
- Resource constraints — As a two-person founding team working evenings and weekends (and eventually full-time), we were building product, selling, supporting customers, and handling operations simultaneously. That’s a pace that’s hard to maintain indefinitely.
- Competitive landscape — The AI workspace category evolved rapidly. By 2025, larger players with deeper pockets were entering the space with similar value propositions.
What We’d Do Differently
- Validate with paying customers before building — Our Rev1 Ventures experience taught us this lesson early, but it bears repeating. Letters of intent and verbal interest aren’t the same as signed contracts.
- Start even narrower — We niched into healthcare SMBs, but in retrospect, we could have gone even tighter — a single workflow in a single type of practice — and expanded from there.
- Prioritize revenue over features — We built a sophisticated product. But sophistication doesn’t pay the bills — solving one painful problem for one customer does.
What We’re Proud Of
Despite the outcome, we built something real. A fully functional, privacy-first AI platform. Microsoft integrations. An analytics engine grounded in Lean Six Sigma and behavioral psychology. A product that caught the attention of a Harvard researcher. And most importantly, we went from dinner-table idea to launched product to academic citation — as two cousins working out of our apartments. That’s a story worth telling.