Case study

Zaya (Internal) - Web App + Desktop Agent + Chrome Extension

An internal recruiting productivity platform built as a three-part system: Web App (command center), Desktop Agent (orchestration + co-pilot), and Chrome Extension (in-context workflows).

AICopilotWorkflowB2BElectronChrome Extension
← Back
Zaya (Internal) - Web App + Desktop Agent + Chrome Extension screenshot

TL;DR

  • What it is: An internal recruiting productivity platform built as a three-part system: Web App (command center), Desktop Agent (orchestration + co-pilot), and Chrome Extension (in-context workflows).
  • Primary objective: Improve both speed and quality of recruiter workflows.
  • Core use cases: Better profile discovery from job boards (e.g., Naukri), AI-assisted shortlisting, and workflow improvements across sourcing → screening → outreach. (Internal-only; metrics omitted; currently in development.)

Context

Recruiters lose time and quality when workflows are fragmented: sourcing and screening happen across multiple tabs/tools, shortlisting depends heavily on manual judgment at speed, and outreach and follow-ups are repetitive and inconsistent.

Zaya was built to make recruiting workflows faster, more consistent, and easier to operate, while adding AI assistance in a controlled way.

Scope: Zaya is internal-only and currently in development. This page focuses on the system design and workflow intent, without exposing sensitive data, internal performance numbers, or proprietary implementation details.

What the product does

Zaya Web App - "Command Center"
  • Central place to manage workflows, views, and operational dashboards
  • Configuration surface for templates, rules, and reporting views
  • Provides a single hub for recruiter productivity signals
Zaya Desktop Agent - "Co-pilot + Orchestrator" (Electron)
  • Built using Electron
  • Hosts the primary "co-pilot" intelligence layer and orchestrates actions across components
  • Coordinates tasks that need reliable execution across the recruiter's environment
Zaya Chrome Extension - "In-Context Workflow"
  • Enables actions inside the browser where recruiters already work
  • Captures context and reduces manual copy/paste
  • Triggers co-pilot and workflow steps without forcing a tool switch
Better profiles from job boards
  • Streamlines discovery and capture of better-fit profiles from job boards (e.g., Naukri)
  • Makes profile intake more structured and consistent
AI-assisted shortlisting
  • Uses AI assistance to improve shortlisting and reduce noise
  • Supports faster "qualify vs. disqualify" decisions with recruiter-in-control design
Workflow automation and consistency
  • Reduces repetitive steps across sourcing, screening, and outreach
  • Uses templates and structured capture to make execution consistent across recruiters
Email support
  • Integrates with email workflows for outreach and follow-ups (vendor-agnostic)

Feature highlights

Core (daily workflow)
  • Web hub to manage workflow views and recruiter activity
  • Browser extension to capture profiles and trigger actions in-context
  • Desktop co-pilot to assist with screening/shortlisting workflows
  • Structured notes and templates to standardize evaluation
Advanced (quality + governance)
  • Human-in-the-loop co-pilot (suggestions support decisions; recruiters remain accountable)
  • Role-based boundaries and controlled access to internal workflows
  • Audit-friendly workflow traces (what happened, when, by whom)
  • Instrumentation for adoption and workflow bottlenecks (internal reporting)

Key decisions

Build a 3-part architecture instead of "just a web app"

Tradeoff: simpler build vs deeper workflow integration. Chose Web App + Electron Desktop Agent + Chrome Extension to meet recruiters inside their daily tools while keeping a central control plane.

Outcome: Lower friction, stronger adoption potential, and a foundation for reliable co-pilot workflows.
Improve quality at the source (job board intake + profile signals)

Tradeoff: optimize only shortlisting vs improve upstream inputs. Focused on getting better profiles from job boards and standardizing how profiles are captured and evaluated.

Outcome: Higher-quality pipeline and less wasted screening effort.
AI assistance with guardrails (not autopilot)

Tradeoff: maximum automation vs trust and control. Chose co-pilot design that supports shortlisting decisions while keeping the recruiter in control and enabling traceability.

Outcome: Better consistency and quality without compromising governance.

Impact

Status
In development (internal rollout plan in progress)

Technical depth

  • Defined responsibilities across web app, extension, and Electron desktop agent
  • Designed workflows for reliability (coordination, fallbacks, clear handoffs)
  • Built a co-pilot architecture that supports human-in-the-loop decisioning
  • Integrated with job board and email workflows in a vendor-agnostic way (no sensitive details)

Collaboration

  • Engineering (architecture, reliability, integration boundaries)
  • Design (workflow ergonomics, low-friction UI, cognitive load reduction)
  • Recruiting teams (shadowing, feedback loops, iteration)
  • IT/Security stakeholders (internal constraints and deployment expectations)

Testimonial

The co-pilot suggestions helped our recruiters find qualified candidates faster while keeping control and context.
Head of Talent, Zaya, Recruiting
See resume →