Case Studies | SaaS B2B

Reading stock signals from any company website with agentic AI

AI Web Scraping Stock Analysis Tool

About

A SaaS B2B business working with Cloud Combinator on AWS. The client is anonymised at their request.

Challenge

The challenge had three focus areas.

A manual workflow that would not scale

Deciding whether a company's website reveals its stock levels meant a person visiting the site, hunting through product and category pages, and making a judgement by hand. With a sales team assessing many companies a day, that manual effort was the bottleneck the project set out to remove.

Web scraping is brittle and heavy to run

Traditional scraping typically leans on Docker-based Playwright deployments that are awkward to operate and fragile against dynamic rendering and anti-bot protections. The client wanted the crawling to be intelligent and managed, not a stack of containers to babysit.

Turning a raw crawl into a clear decision

Reaching a page is not the same as understanding it. The system needed to explore a site the way a person would, spot stock-related signals such as inventory counts, availability indicators and catalogue structure, and return a structured yes or no with a supporting explanation rather than a pile of scraped text.

Solution

An employee submits a website URL through an API backed by an AWS Lambda function. The Lambda acts as a thin orchestrator: it launches an Amazon Bedrock AgentCore Browser session and writes a pending status to Amazon DynamoDB, so long-running crawls do not block the caller. The AgentCore Browser Tool provides managed, serverless browser automation, which removes the need for Docker-based Playwright entirely and comes with built-in observability including live view and session replay.

600

Company websites assessable per month at target usage (projected)

Serverless

Managed browser automation, no Docker or Playwright to maintain

Yes / No

Structured determination returned with a supporting explanation

By the numbers:

  • 600 - Company websites assessable per month at target usage (projected)
  • Serverless - Managed browser automation, no Docker or Playwright to maintain
  • Yes / No - Structured determination returned with a supporting explanation
Changes

The engagement delivers a working, end-to-end proof of concept: an URL goes in, an AgentCore Browser session launches, the agent crawls and analyses the site, and a yes or no stock determination with a supporting explanation comes back and is persisted. Acceptance is on demonstrating that full workflow and confirming the AWS environment is suitable for the client's AI-powered web scraping and stock analysis tool.

  • API-driven workflow — An URL submitted through a Lambda-backed API returns a structured yes or no with an explanation, saved automatically to DynamoDB.
  • Managed browser automation — Amazon Bedrock AgentCore Browser replaces Docker-based Playwright, with built-in live view and session replay for observability.
  • Intelligent analysis — A Strands agent on Amazon Bedrock navigates each site, finds stock signals and handles dynamic rendering and anti-bot barriers.
  • Tracked and auditable — Every submitted URL, determination and explanation is persisted in DynamoDB, giving the client a queryable history for audit and reporting.
  • Handover — A CloudFormation template, a runbook and a reference architecture, plus a knowledge-transfer session, leave the client's team able to operate and iterate independently.

Built as a proof of concept, the tool is designed with a clear path to production: Amazon Cognito can be added to open it up beyond internal employees, an asynchronous pattern with Step Functions or Fargate can absorb longer crawls beyond Lambda's execution limit, and the same agentic foundation can be pointed at metrics well beyond stock. The client now has a serverless, AWS-native base to build the next generation of its web intelligence on.

AWS Stack

Amazon Bedrock AgentCore Browser Tool

For managed, serverless AI-driven web crawling.

Amazon Bedrock

With the Strands Agents framework for reasoning over crawled content.

Amazon DynamoDB

For tracking submissions, determinations and explanations.

Amazon S3

For AgentCore Browser session recording.

AWS CloudFormation

For repeatable, automated deployment of the whole stack.

Amazon CloudWatch

For real-time monitoring of Lambda and browser sessions.

YOU MIGHT LIKE

Related success stories

View all case studies

Case Studies | Insights

Utilising Language Recognition, Speed, and Enhanced Security to Make Social Media a Force for Good

  • Here, we take a detailed look at how the Cloud Combinator team collaborated with another cutting-edge AI service provider that provides intelligent systems to “make social media more social” for brands and users alike.
  • Arwen AI is a UK-based startup specialising in AI solutions to manage and enhance brands’ social media interactions. Founded in 2020 by Matt McGrory, Dr. David Cole, and Joel Bailey, Arwen. AI focuses on using AI to automatically detect and remove spam, toxic comments, and other unwanted content from social media platforms.
  • The team at Arwen have three core products. ‘Moderate’ is focused on identifying and removing toxic content from social media channels. ‘Engage’ helps brands identify and engage with meaningful conversations on social media, and ‘Customize’ allows brands to apply bespoke algorithms to their channels - creating an even more effective moderation and engagement.
Read more
CONTACT US

Ready to turn AI into impact?

We'll help you spot the highest-value opportunities, reduce risk around your first AI initiative, and define a clear path to results from day one.

Why talk to us:

Outcome-driven recommendations

AWS-recognised delivery expertise

Risk-aware AI adoption

Clear next step, not a sales pitch

Start with a focused 20-minute conversation about your goals — no pressure, no commitment.

We use cookies

This website uses cookies to enhance user experience and to analyze performance and traffic on our website.

See our Privacy Policy for details.