Creating text and images for every new piece of content meant repeating the same manual work, over and over. One prompt needed to do the job of many.
Client
A consumer products company wanted to use generative AI to automatically create content, text and images together, from a single input, integrated directly into their existing software rather than run as a separate tool.
Challenge
Most AI pilots never make it past a demo, they work once in a controlled test, then fall apart under everyday use. This client needed something that could actually hold up in production: turning one simple input into multiple connected pieces of content, reliably enough to trust without someone checking every output, and running as a lightweight, low-cost tool that didn’t rack up expenses when nobody was using it.
Solution
We built a serverless setup on AWS that only costs money when it’s actually in use, connected to OpenAI’s API to generate the content itself. Rather than treat testing as a final step, we iterated on it throughout the build, so the tool could be trusted in daily production use instead of staying a one-off demo. We built a simple web interface so anyone on the team could generate a full set of content, text and images together, from one input.
Results
From Demo to Daily Use.
The client didn’t just get a proof of concept, they got a tool running in production today, generating a full set of content from a single input and saving hours of repetitive work. We also tuned the system’s startup performance so results came back faster for anyone using it.
Want me to hold off finalizing this one until you can confirm any real security specifics, or is the “production, not a pilot” framing enough on its own for this case study?
Technical Details
Built on AWS using Lambda, CodePipeline, CloudFront, API Gateway, and S3 for a serverless, pay-per-use infrastructure, integrated with the OpenAI API. Backend built in Java, frontend in Handlebars.java, with Lambda cold-start optimization to reduce response wait times.
What’s Next
The client is looking to enrich the underlying model with their own dataset and add more content generation options going forward.
Spending Too Many Hours on Repetitive Content?
If your team keeps recreating the same kind of content by hand, we can help you automate it without the infrastructure headache.
Date
04/2023
Languages
Java
Typescript
Frameworks
OpenAI API
Handlebars.java
Scrum Development
Cloud
AWS Lambda
AWS Pipeline
AWS Codefront
AWS Route53 SDK
Tools
IntelliJ IDEA
AWS Codecommit
Discuss Your Project
Great things happen when good people connect. Leave us your details, and we’ll get back to you.
By sending the information in this form, you agree to have your personal data processed according to A-CX’s Privacy Policy and Cookie Policy to handle the request and respond to it.
Related Case Studies
Microsoft: Zero Trust Security for Windows 365
Microsoft Entra Private Access enables secure and seamless access to internal applications by isolating and protecting Windows 365 virtual machines from unauthorized network access.
Consumer Products Company: An AI Assistant Trained on Its Own Data
We helped a client create a custom LLM to encourage their employees to explore AI to improve the efficiency of different business practices.
Global Financial Services Firm: Secure Connections Across Public Clouds
Find out how we helped a company secure its public cloud services. This case study details the challenges faced and how our expert team implemented cutting-edge technologies to maintain confidentiality and integrity of sensitive information. See how we protected against cyber threats and ensured secure communication for this company. Read…
