Project Scope (SoW)

Scope of Work for Agentic AI – Telemetry, organized with accordions.

Accordions Charts (SVG) Icons Enhanced images

Scope of Work (SoW)

Accordion view

Project Scope — Scope of Work (SoW)

Agentic AI – Telemetry program scope (as provided).

🎯1. Project BackgroundOverview

The Agentic AI – Telemetry Program is an internal QBA strategic initiative designed to capture, contextualize, and analyse user activity data from key business functions (HR Admin, Finance, Operations, etc.) to uncover process knowledge, operational intent, automation opportunities, and efficiency improvements.

A lightweight telemetry script is deployed on selected user systems to capture detailed activity signals (keystrokes, application usage, window context, timestamps). These signals are transformed into meaningful sessions, intents, and business processes through advanced contextualization and rule-based intelligence.

The output will power decision-making for process optimization, automation roadmap creation, skill gap analysis, and productivity enhancement.

2. ObjectivesOutcomes
  • Capture fine-grained user activity data across business functions.
  • Convert raw telemetry data into structured sessions and business intents.
  • Identify process patterns, inefficiencies, rework, context switching, and automation potential.
  • Create a scalable process intelligence framework for continuous improvement.
  • Enable leadership with actionable dashboards and KPIs for workforce and process optimization.
🧩3. Scope of WorkIn/Out

A. Telemetry Data Collection

In Scope

  • Deployment of telemetry agent on approved user machines.
  • Capture of keystrokes, active application, window titles/context, timestamps & idle duration, user & system identifiers.
  • Secure storage and handling of telemetry data.

Out of Scope

  • Screen recording
  • Personal data capture beyond approved metadata
  • Any content violating privacy policies

B. Sessionization & Contextualization Engine

  • Session break logic: idle thresholds, app switching behavior, context switching signals, discontinuity ranking, relationship mapping between applications.
  • Assignments: Session IDs, session types, business intent tags, process categories and subcategories.

C. Process & Intent Intelligence

  • Map telemetry sessions to business processes, functional workflows, user intent and task purpose.
  • Identify repetitive tasks, friction activities, rework indicators, training gaps, automation candidates.

D. KPI Framework & Analytics

User process throughput

Throughput per process & user segment.

Context switching index

Frequency + cost of app/context changes.

Top time-consuming screens

Time distribution by application/window.

Automation potential score

Opportunity scoring for candidate tasks.

Skill gap index

Signals suggesting training needs.

Process complexity index

Variability, branching, and rework.

Idle vs active work ratio

Engagement + interruption insights.

Session efficiency score

Cycle time + friction normalization.

E. Dashboard & Visualization

  • Session & intent by user, date, function
  • Process maps and heatmaps
  • KPI trends & comparative analysis
  • Automation opportunity rankings

F. Governance, Privacy & Compliance

  • Data anonymization & role-based access
  • Compliance with internal IT & data governance standards
  • User consent and usage transparency
📦4. Key DeliverablesOutputs
DeliverableDescription
Telemetry AgentDeployed and stable data capture mechanism
Session & Intent EngineRule-based classification system
Process TaxonomyComplete task & process classification model
KPI FrameworkDefined metrics & formulas
Analytics DashboardExecutive and operational views
Automation Opportunity ReportRanked improvement opportunities
Final Project ReportConsolidated insights & roadmap
🧷5–6. Assumptions, Dependencies & Success CriteriaDefinition of done

Assumptions & Dependencies

  • Continued availability of telemetry data capture enhancements
  • User participation and consent
  • Timely engineering support for agent & data pipeline
  • Secure infrastructure provisioning

Success Criteria

  • ≥ 95% accuracy in session & intent classification
  • Identification of minimum 20–30% automation potential across target functions
  • Measurable reduction in context switching and process friction
  • Adoption of dashboards by leadership for decision-making