OnsiteIQ Agent for Intelligence Reports
Designed an AI-powered reporting agent that transforms fragmented construction data into instant, actionable insights.
Role
Senior Product Designer (Individual Contributor)
Timeline
October 2025 – current
Type
B2B / Web Application
Impacts
Reduced time to extract the insights of the project
Improved the project report generating process rate by 2–3x
Enhancing the credibility to the digitized project progress data
Contributed to generate $1.3M MRR (Monthly Revenue Rate) and Year-to-Year growth by 42%
Overview
Problem
Project stakeholders struggled to extract meaningful insights from fragmented field data. Generating reports was manual, time-consuming, and often required revisiting multiple sources—slowing down decision-making.
Contribution
Led the design of an AI reporting agent from 0→1—defining key use cases, shaping the interaction model, and partnering with engineering to bring LLM capabilities into a production workflow.
Insights
Users needed instant, decision-ready summaries, not raw data exploration
Report generation had to integrate into existing workflows with zero friction
Need to build a trust in AI depended on clear data sources and verifiable outputs
Design Process
Understanding Reporting Needs & Prompts
Identified high-value reporting scenarios and structured prompts around real user workflows
Rapid Prototyping with AI Tools
Enabled fast iteration on interaction patterns and validated feasibility with engineering early by using Claude + Figma
Seamless Integration
Embedded the agent into existing workflows to reduce adoption friction
Solution (MVP)
Designed to reduce cognitive load and eliminate manual data synthesis, the OnsiteIQ agentic AI delivers structured insights directly within the project workflow.
Providing a curated prompts for each users to use to create any type of reports
Summarises daily risks and key signals at a glance without understanding
Generates progress summaries and trend analysis instantly
Connects insights to source data (images, reports, 360 captures) for validation
Before
Field Data
Photos, Notes, Reports, Emails
Manual Review
Approximately 30–60 min
Compile Report
Share / Validate
- Fragmented data sources
- Time-consuming synthesis
- High cognitive load
After
Field Data
Auto-collected in Platform
OnsiteIQ Agent
Instant Report
Takes seconds
Linked Sources
Validation
- Instant summaries
- Context-aware insights
- Source-linked trust
Video Demo
Impact
2–3x faster report generation
Reduced manual effort in extracting project insights
Increased adoption of AI-powered reporting across projects
Contributed to $1.3M+ MRR and 42% YoY growth
Why It Matters
A project manager or other project stakeholder can now generate a weekly report in seconds instead of spending 30+ minutes compiling data across sources.
This shifts reporting from a manual, reactive task into an automated, proactive workflow with agentic AI—enabling faster, more informed decisions.
Future Directions
AI image search and contextual linking
Smarter cross-project insights aggregation
Deeper integration with downstream workflows
I have been working on to add a image search that enables user to quickly find the point of interest based on the context. This also allows users to quickly extract and correlate with other data such as progress, gantt chart in order to extract the precise project progress.
