Megha Arora Portfolio - Senior UX Designer

scout - siterecon's
AI Agent

SiteRecon sells property measurements to landscaping companies and contractors across the US and Canada. Scout is their AI agent that lets users measure any property themselves in minutes.One conversational flow. Four stages.

VISIT scout ai

my role

Senior UX Designer

AI Saas

timeline

6 weeks

Overview

SiteRecon sells accurate property measurements to landscaping companies and contractors across the US and Canada. Traditionally, users would place an order and a team of cartographers would manually draw measurements from satellite imagery — a process that worked, but didn't scale. Scout is SiteRecon's AI agent that puts that capability directly in the user's hands. A 4-step guided chatbot experience that takes any user — regardless of technical ability — from property search to accurate measurements in minutes, with estimation and proposal built into the same flow.

The homepage was designed to be a gateway to all content on the platform. We focused on clarity, accessibility, and visual appeal:

Segmented Blog Categories: Organized by type (news, guides, reviews) for easy navigation.
Trending Headlines Ticker: Instantly informs users about the hottest topics in tech.
Product Grids: Featured cover images, blog titles, post dates, and read times for user clarity.
Search Functionality: A robust search bar allows users to quickly find topics of interest.

We used a vibrant color palette and bold typography to create a modern, high-energy vibe that appeals to tech enthusiasts. The homepage is both inviting and functional, encouraging users to explore the platform.

THE USER

Landscaping companies, contractors, and account managers who visit properties, do site audits, and build proposals for clients. These are not tech-savvy users. They're outdoors people running physical businesses, often on the field, with no tolerance for software that makes them think. That single constraint defined the design direction from day one.

The page was structured to maximize readability and encourage users to explore further. Every element was designed to support the client’s affiliate marketing goals.

"In This Blog" Section: Featured at the very start, it includes a table of contents with affiliate links to the products mentioned in the blog. This allows users to decide whether the content is relevant to their needs and provides direct access to products.
CMS Integration: A dynamic CMS was built to handle blog content efficiently, ensuring easy updates and a seamless connection between the blog database and the front end.

THE DESIGN DECISION —
CHATBOT OVER STEP FLOW

The first instinct was a 4-step flow with a progress indicator at the top — logical, familiar, easy to build. We rejected it.

The problem: in a multi-step flow, users lose track of the decisions they've already made. To review or change something, they have to navigate backwards. For a non-technical user estimating a complex property with multiple service zones, that's a UX failure waiting to happen.

A guided chatbot experience solves this natively. Every decision the user makes — which layers to measure, which services to add, what quantities to confirm — stays visible in the conversation thread above. No backtracking. No confusion about where they are or what they chose. The chat history is the progress indicator.

It also scales. New steps, new service types, new decision branches can be added to the flow without redesigning the interface.

The page was structured to maximize readability and encourage users to explore further. Every element was designed to support the client’s affiliate marketing goals.

"In This Blog" Section: Featured at the very start, it includes a table of contents with affiliate links to the products mentioned in the blog. This allows users to decide whether the content is relevant to their needs and provides direct access to products.
CMS Integration: A dynamic CMS was built to handle blog content efficiently, ensuring easy updates and a seamless connection between the blog database and the front end.

THE 4-STAGE FLOW

Stage 1 — Takeoff

User searches a property. Scout renders the boundary on a satellite map. User selects which zones (layers) need measurement — turf, mulch beds, pathways, and so on. No manual input.

Stage 2 — Service Maps

Scout asks what needs to be done with each selected layer. Turf selected? It surfaces mowing, trimming, edging. Mulch beds? Refreshing, weeding, reshaping. Only contextually relevant services appear — nothing else.

Stage 3 — Estimation

Scout builds the estimation sheet automatically — crew rates, production rates, frequency, gross margin — pre-filled with industry benchmarks. Users edit only what's specific to their business.

Stage 4 — Proposal

Scout compiles everything into a client-ready proposal document. Property details, services, pricing, scope — formatted and ready to send. No reformatting. No switching between tools.

FUNNEL ANALYSIS —
WHERE THE REAL PROBLEM WAS

After the beta launch, I tracked the full funnel in Mixpanel and found a metric on fire.

67% of users were dropping off at Stage 1 — without ever entering a property address.

The original first screen had a 2-column layout: instructions on one side, a search bar on the other. Users landed, didn't know what to do, and left before the product had a chance to show its value.

THE FIX

Stripped the 2-column layout entirely.

Replaced it with a single, full-width property search bar — the only element on the screen. Added a sample property with an interactive placeholder so users could immediately see what Scout does without committing to their own data. The goal was one thing: get the user to measure at least one property and experience the product.

Result: 30% reduction in Stage 1 dropoff.

IMPACT

54 paid customers acquired in the first campaign launch, with the product still in beta

30% reduction in Stage 1 dropoff after the search bar redesign
Full public launch: June 30, 2026

54 paid customers acquired in the first campaign launch, with the product still in beta

30% reduction in Stage 1 dropoff after the search bar redesign
Full public launch: June 30, 2026

More Works More Works

Samsung Keyboards

UX Enhancement

Samsung Keyboards

UX Enhancement

Samsung Keyboards

UX Enhancement

Athlabs Branding

Branding & Packaging

Athlabs Branding

Branding & Packaging

Athlabs Branding

Branding & Packaging

Nutrabay App

Product Redesign

Nutrabay App

Product Redesign

Nutrabay App

Product Redesign

Nutrabay Admin Panel

Supplement Marketplace

Nutrabay Admin Panel

Supplement Marketplace

Nutrabay Admin Panel

Supplement Marketplace

Let'S WORK