Praveen Manchi

Cortex AI - Deloitte

I led the design of Deloitte’s CortexAI suite, a cutting-edge GenAI-powered platform that revolutionizes how 120K+ users discover, purchase/request digital assets such as algorithms, prompts, datasets, and databases—across cloud ecosystems. This solution drove ~$2M in weekly revenue while enhancing usability and adoption across diverse business units.
Cortex AI interface collage

Overview

Cortex AI is an essential tool for enterprise AI adoption, offering a marketplace for AI/ML assets with GenAI-driven matching and integration. Our focus was on simplifying complex workflows for developers, data scientists, and business users while ensuring security and scalability. The platform supports multiple cloud providers including AWS, Azure, and Google Cloud, with features like real-time asset previews, collaborative workspaces, and analytics dashboards.
Team: Product Designers (Me🙋🏻‍♂️), 20+ Developers, Product Managers, Data Scientists, leads and stakeholders.
Timeline: January 2023 - June 2024 (18 months, with quarterly releases)
Scope: Web application with API integrations, Cloud platforms (AWS, Azure, Google Cloud), Design System, AI-powered search, one-click deployments, and customizable dashboards for monitoring asset performance.

The Challenge

  • Enterprises faced fragmented assets like datasets, prompts, code & algorithm, frameworks, etc. Current marketplaces leading to inefficient discovery and hard to integration. Complex compliance requirements and varying cloud environments created barriers, resulting in low adoption and missed revenue opportunities.
  • Dense technical jargon confused non-technical users, while lengthy onboarding processes caused high drop-off rates. Additionally, lack of personalized recommendations made it hard to find relevant assets quickly.
  • Inconsistent user experiences across devices, poor search accuracy for technical terms, and manual integration processes that could take days instead of minutes.

    • Impact & Results

    revenueBoost Revenue

    Drive ~$2M weekly through streamlined purchases and integrations, with upsell opportunities via bundle recommendations.

    adoptionEnhance Adoption

    Increase user engagement across 120K+ enterprise users by 40% through intuitive interfaces and tutorials.

    timeReduce Onboarding Time

    Cut new client onboarding by over 70% with guided wizards and automated setups.

    securityEnsure Scalability & Security

    Build a robust platform compliant with GDPR, HIPAA, and enterprise standards, handling 100K+ daily queries.

    collaborationFoster Collaboration

    Enable team-based asset sharing and reviews to improve knowledge transfer within organizations.

    The Solution

    We created a unified GenAI-powered marketplace that intelligently matches users with AI/ML assets, simplifies purchases, and automates integrations across clouds. By leveraging scalable design systems and user-centric flows, we transformed complexity into seamless experiences.

    Core features include AI-powered search with natural language processing, one-click deployments, and customizable dashboards for monitoring asset performance.
    Solution Interface Highlights
    (NDA-restricted – Envision clean dashboards with AI recommendations, one-click integrations, and progress trackers)

    Project Highlights

    • GenAI-driven asset discovery with personalized recommendations based on user history and role.
    • Seamless cross-cloud integration reducing setup time by 60%, with support for hybrid environments.
    • Intuitive purchase workflows with compliance checks built-in, including audit trails.
    • Scalable design system enabling rapid iterations and consistent branding across modules.
    • Advanced analytics for users to track asset usage and ROI, with exportable reports.
    • Accessibility features ensuring WCAG 2.1 compliance for inclusive enterprise use.
    Feature Showcases
    (Withheld – Picture interactive prototypes demonstrating matching algorithms and user journeys)

    How Research Informed Our Design Solutions

    We began with extensive user research, including interviews and surveys with enterprise users. This guided our iterative design process, validated through 100+ usability tests and heuristic evaluations.

    Key insights: 65% of users prioritized quick search results, 40% needed better mobile support, and 75% valued integration automation.
    Fragmented Asset Discovery

    Users struggled with scattered catalogs; we unified them with smart search and faceted filtering.

    Complex Integration

    Multi-cloud setups were tedious; automated matching and templates simplified this.

    Compliance Confusion

    Regulatory jargon overwhelmed; built-in guides and tooltips clarified processes.

    Lengthy Onboarding

    High drop-off rates; streamlined flows with progressive disclosure reduced time by 70%.

    Lack of Personalization

    Generic recommendations; implemented ML-based suggestions tailored to user profiles.

    Performance Monitoring

    No visibility into asset efficacy; added dashboards with real-time metrics.

    Design Process Flow
    (NDA-restricted – Visualize wireframes evolving into polished interfaces through iterations)