Business Overview

The client, CarSoup, operates a leading online automotive marketplace for advertising, buying, and selling cars of all types from various dealers based on different requirements. The aggregation platform synchronizes daily with partner inventory to offer customers up-to-date listings, comparison tools, reviews, buying advice, and financial services.

Under a commission-based model, the CarSoup marketplace streamlines the search experience, and once a user finds a match, the platform redirects them to the dealer’s website. This empowers dealers to control the customer experience, finance requests, and lead conversion.

The client identified a significant point of friction: while the platform offered powerful search capabilities, the filter and configuration setup was quite complex and unintuitive, discouraging new users from using it. This ultimately led to users getting insufficient car search results and the client losing conversions and missing revenue opportunities. To address this issue, they decided to build an AI-powered site search capability that would:

  • Elevate user experience by simplifying the search process through a conversational interface
  • Improve the search results by utilizing more advanced features of the conversational search engine, which are omitted by users during manual search

Project Scope

NIX was engaged to develop a smart AI semantic search that precisely classifies the user’s intent and requests relevant clarifications through seamless, human-like interactions. The goal of the engine is to efficiently utilize the client’s complex search engine API by responding to user inputs and adjusting search parameters based on the conversation.

2 Project scope
Frame 1000004588

Challenge

Many shoppers use slang terms when searching for certain cars, but the LLM often fails to interpret those informal inputs and map them to structured search parameters without specific context. The main challenge was the gap between how users naturally describe vehicles and how databases categorize them.

Solution

Our team engineered a sophisticated AI agent application hosted on AWS Cloud to significantly improve the vehicle discovery process, while also addressing client challenges in smart, innovative ways.

Intelligent Orchestration and Integration

We developed an advanced decision‑making framework leveraging AWS AI Agents Service together with the Azure OpenAI Service. This multi‑agent system orchestrates the full capabilities of the client’s search API, dynamically configuring the best search parameters based on the user’s conversation.

To ensure a cohesive user experience, we integrated the engine via FastAPI and developed a custom, responsive React chatbot interface embedded directly within the platform’s UI. We further reinforced this complex ecosystem with a custom LangSmith logging and monitoring application. This provides total visibility into the AI’s decision-making paths, allowing us to proactively prevent pipeline bottlenecks.

4 solution
5 Solution

Two-phase Semantic Optimization

To bridge the gap between structured API logic and shoppers’ natural language, we implemented a phased optimization strategy:

  1. First, we analyzed the platform search statistics to compile a list of automotive slang and synonyms. By integrating this library directly into the core prompts, we provided the AI with the necessary context to translate informal queries.
    Impact: 20% increase in search accuracy KPIs
  2. Next, we added a separate agent to perform reasoning and match ambiguous user inputs with the actual search parameters, like car model, color, etc. This is designed to act as a semantic bridge, analyzing user queries against our evolving database of slang and intent.
    Impact: An additional 15% increase in KPIs after moving from static prompts to an active reasoning agent

Key Gains

  • 01

    Enhanced search

    Improved the accuracy and relevance of search results, enabling users to find desired cars more efficiently.

  • 02

    Improved user experience

    Quicker, more precise results led to longer user sessions, increased engagement, and higher satisfaction rates, fostering customer loyalty in the competitive online market.

  • 03

    Reduction in search abandonment

    The AI engine effectively guided users who previously dropped off due to the tricky manual filter configurations.

  • 04

    Improved market competitiveness

    The platform now offers a unique “digital concierge” experience that differentiates it from standard automotive aggregators, fostering long-term customer retention.

Outcome

Collaborating with NIX, the client implemented the smart semantic search engine that transforms the user journey from a tricky filtering task into a streamlined, conversational experience. By successfully bridging the gap between search API parameters and user intent, the solution has already delivered significant gains in both user engagement and lead quality.

4 solution-1

Success metrics

Up to 40%

increase in user engagement with advanced search features

35%

overall improvement in the relevance of search results

75%

reduction in support tickets related to search filter issues

Team:

Team:

5 experts ( Frontend Developer, PHP Developer, QA Automation Specialist, DevOps Engineer, Data Scientist )
Tech stack:

Tech stack:

AWS Cloud, AWS AI Agents, Azure OpenAI, Python, React, FastAPI, LangSmith

REQUEST A CONSULTATION

Contact us   

Relevant Case Studies

View all case studies

Bridging OpenAI and Amazon Bedrock to Enhance LLM Evaluation Platform

Internet Services and Computer Software

Success Story Bridging OpenAI and Amazon Bedrock to Enhance LLM Evaluation Platform image

AI Hazard Detection for Care Facilities: 98% Accuracy in Safety Threat Prediction

Healthcare

Success Story AI Hazard Detection for Care Facilities: 98% Accuracy in Safety Threat Prediction image

AI Assistant: 70% Faster Support for Enterprise Employees

Social Networks and Communications

Success Story AI Assistant: 70% Faster Support for Enterprise Employees image

Starday Foods: Scaling to 100K Posts per Hour With AI

Food & Beverages

Success Story Starday Foods: Scaling to 100K Posts per Hour With AI image

Driving AI Innovation for a Global Customer Service Leader

Social Networks and Communications

Success Story Driving AI Innovation for a Global Customer Service Leader image

AI-Driven Application for Mental Health Support in the US

Healthcare

Success Story AI-Driven Application for Mental Health Support in the US image
01

Contact Us

Accessibility Adjustments
Adjust Background Colors
Adjust Text Colors