Anarock genie - Reduced AI agents creation time from 2 weeks to 2-3 mins
Product design & User research
.
Web
.
3 Months
What does Anarock do?
ANAROCK is one of India’s leading real estate services companies, helping real estate builders efficiently sell and manage their property inventory. By leveraging in-house technology and data-driven tools, Anarock streamlines property transactions, making them faster, smarter, and more transparent.
My role
User research, User flows, Brainstorming & exploring different AI tools & Final UI design
Collaborated with
2PMs, 8+ engineers, Business heads & Ops team
Challenges & Pain Points
Key Business Challenges
Lead Generation Timing - 70–80% of online leads are generated during non working hours
No Immediate Response - Agents are unavailable during non working hours to respond or engage with leads
Delayed Follow-up - It takes 20–40 hours on average to contact a new lead
Lead Loss - Delays often lead to losing prospects to competitors
Sales Agent Pain Points
Repetitive Queries - 80–90% of fresh lead queries are similar and repetitive
Inefficient Lead Prioritisation - Limited bandwidth to focus on high-intent leads
Target Pressure - Constant pressure to meet aggressive sales targets
Invoice Delays - Administrative bottlenecks in raising invoices promptly
Solution
To address these pain points, we introduced Anarock Genie — an AI-powered platform designed to create intelligent AI agents that answer lead queries 24/7, share project information, and route high-intent leads directly to sales agents. This eliminates repetitive queries, reduces human agent workload, and enables them to focus on qualified leads, helping them close deals faster.
Product challenge
During the MVP phase of Anarock Genie, the core processes of AI Agents creation and testing were handled using Google Sheets and Docs. Account managers relied on these manual tools to design and train the AI agents. While the adoption of Anarock Genie rapidly increased across India, the process of making a AI agent live often took 2–3 weeks, leading to significant delays in go-to-market speed.
Understanding users & their pain points
To understand the current AI agent creation process and identify the factors contributing to longer deployment times, I talked with account managers and observed them during their bot creation workflows.
Who Are Our Users?
Account managers
Creates AI Agents & integrates with WhatsApp, & Manages the AI Agent. And they manage communication between Anarock and real estate developers.
Sales Team Leaders
They lead the on-ground sales teams and oversee day-to-day operations. And also Monitor Chats between the Lead and AI agent & take actions accordingly
💬
“WhatsApp account setup usually takes weeks from the developer and requires multiple follow-ups for credit, money, and phone number.”
🕒
“Prompt approvals from the tech team usually take time and don’t revert immediately.”
🗂️
“We literally need to upload media files to S3 with no preview. If something goes wrong or we need to update 40–50 media files.”
Challenges in current workflow
These user interviews gave me valuable insights to identify the key pain points across the user journey of creating AI agents & managing them
📄
Project Info Delays
Getting project information—such as pricing, available inventory, and brochure details—from the developer is often delayed.
✍️
Manual Doc Prep
Account Managers often upload brochures to ChatGPT to extract text and manually create training documents.
📤
Bulk Media Uploads
Account Managers often upload 40–50 media files directly to AWS S3, where there are no media previews, making it difficult and hectic to edit or manage the files.
☎️
IVR Detail Delays
To integrate bots on WhatsApp, Account Managers need a phone number and credit balance from the developer — a process that often gets delayed.
🧠
Prompt Review
Account Managers are not technical, so they rely on the tech team to review prompts — which consumes time and bandwidth for both teams.
Ux goals
These user interviews gave me valuable insights to identify the key pain points across the user journey of creating AI agents & managing them
👩🏻💻
Reduce the Learning curve for Account managers
🔄
Eliminate back-and-forth between multiple platforms
🚀
Reduce the turnaround time for bot deployment
Initial explorations
I have analysed how users can create AI agents on other platforms, focusing on ChatGPT, Character.ai, Relavance AI, and Notebook lm. I aimed to understand how these platforms approach training documents, prompting
Product visuals
The project is under NDA, so I can't reveal final designs & major design decisions, but i can show them in 1:1 call
Final AI agent creation flow
🫢
Impact
Reduced the time required to create AI agents and deploy them on mandates. (Currently it takes ~3-4 mins)
With faster agent creation, we successfully deployed AI agents across India. As of today, 48 agents are live, engaging with leads. 🚀
Sales agents focused on high-quality, fast-converting leads, while AI agents efficiently managed other leads and customer interactions. 🚀
%
Of Queries handled by AI agents currently.
+
Leads registered through genie agents in Chennai
+
AI Sales agent created & implemented all over India