AWS Builder Center
How We Built ALXO: Catching Scope Creep with Amazon Bedrock Claude 3 & Serverless AWS Infrastructure

How We Built ALXO: Catching Scope Creep with Amazon Bedrock Claude 3 & Serverless AWS Infrastructure

Learn how Team Arixen built ALXO — an automated scope creep detector and change order generator powered by Amazon Bedrock Claude 3, DynamoDB, S3, and AWS Amplify. Discover how AI and serverless cloud architecture eliminate unbilled software work and save agencies 20-30% in lost revenue.

How We Built ALXO: Catching Scope Creep with Amazon Bedrock Claude 3 & Serverless AWS Infrastructure

By Ilakkiyan J & Team Arixen | Built for #AWSBharathBuildsTour #BharathBuilds #WeMakeDevs



💥 The Problem: The Hidden Cost of "Quick Favors"

In software development, freelancing, and agency consulting, scope creep rarely happens inside formal SOW contracts. Instead, it happens casually:
  • A late-night message on Slack: "Hey, can we quickly add a multi-currency payment checkout?"
  • A WhatsApp chat: "Just one more design revision before launch!"
  • An email thread: "We need real-time ERP inventory sync included in this sprint."
Before developers realize it, a 3-week sprint blows out into 6 weeks, swallowing 20% to 30% of billable revenue. Reconciling this manually is awkward, confrontational, and usually given away for free.
To solve this, Team Arixen engineered ALXO — an enterprise-grade SaaS platform that parses client conversation exports, detects out-of-scope demands using Amazon Bedrock, logs verified items into an authoritative Scope Creep Ledger, and generates legally binding Change Orders in seconds.



🏗️ AWS Multi-Layer Cloud Architecture

ALXO is built on a 100% serverless, multi-layer AWS cloud architecture designed for high availability, security, and low latency.
+----------------------------------+
| User Browser / Client |
+----------------------------------+
|
v
+----------------------------------+
| AWS Amplify Cloud |
| Next.js 14 Web & API Router |
+----------------------------------+
|
+-----------------------------+-----------------------------+
| | |
v v v
+--------------+ +--------------+ +--------------+
| Amazon S3 | | DynamoDB | | Amazon |
| Raw Files & | | Projects & | | Bedrock |
| Exports | | Ledger Items | | Claude 3 |
+--------------+ +--------------+ +--------------+

Key AWS Services Used:

  1. Amazon Bedrock (Claude 3 Haiku):
  • Evaluates incoming message batches against original baseline scope contracts.
  • Assigns a classification confidence score ( CONFIDENCE_THRESHOLD = 0.70 ).
  • Categorizes extra deliverables into labor cost estimates and timeline delays.
  1. AWS Amplify:
  • Hosts the Next.js 14 full-stack application (App Router, Server Actions, API routes).
  • Provides global CDN distribution and continuous deployment pipelines.
  1. Amazon DynamoDB:
  • Serverless NoSQL storage for projects, ledger line-items, and user audit logs.
  • Low-latency single-digit millisecond query performance.
  1. Amazon S3:
  • Secure bucket storage for raw client conversation exports ( .txt ,  .csv ,  .json ).
  1. AWS Cognito:
  • Role-based access control (RBAC) supporting Admin, Developer, and Demo accounts.



🤖 How the Bedrock AI Engine Works

When a user uploads a client conversation thread, ALXO initiates a batch evaluation pipeline:
  1. Baseline Ingestion: The baseline Statement of Work (SOW) is retrieved from DynamoDB.
  1. Context Framing: Amazon Bedrock receives the SOW along with raw conversation snippets.
  1. Structured Classification: Bedrock outputs JSON classifications containing:
  •  isScopeCreep : Boolean flag
  •  category : (e.g., Feature Expansion, Third-Party API, Design Overhaul)
  •  confidence : Confidence rating (0.0 to 1.0)
  •  estimatedHours : Recommended billing impact
  •  justification : Detailed explanation for the client
  1. Offline Fallback Engine: If cloud connectivity is disrupted, ALXO seamlessly falls back to local heuristic rules without failing the user session.



🧪 Real-World Impact: The 15 Scenario Benchmark Suite

To validate ALXO across diverse industries, we built a 15-scenario benchmark suite covering:
  • Mobile Apps & E-Commerce: Healthcare portals, virtual reality property tours, POS systems.
  • Fintech & Security: Payment gateway additions, penetration testing re-scans.
  • Enterprise Software: SAP/Workday HR sync, IoT fleet telemetry.
In testing, ALXO identified over ₹4,50,000 ($5,400 USD) in unbilled scope additions across sample benchmark projects that would have otherwise been given away for free.



🚀 Live Demo & Open Source Repository




👥 Built by Team Arixen

Thank you to AWS India, WeMakeDevs, and the #AWSBharathBuildsTour team for empowering builders!
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article