Caylent Services
Artificial Intelligence & MLOps
Apply artificial intelligence (AI) to your data to automate business processes and predict outcomes. Gain a competitive edge in your industry and make more informed decisions.
Z5 Inventory Achieves 5X Efficiency Gains in Healthcare Inventory Management with AWS AI
Artificial Intelligence & MLOps
Generative AI
At a glance
Z5 Inventory leverages AI-driven transcription and data logging on AWS to automate healthcare inventory accounting, improving processing efficiency by more than 5X and accelerating video to inventory data conversion from days to mere minutes.
Solution Implemented
→ AI-powered transcription using Amazon Transcribe enabled rapid, accurate conversion of inventory video data.
→ Structured inventory data generated using Amazon Bedrock and Anthropic Claude 3.5 Sonnet, integrated seamlessly into Z5’s existing backend systems.
→ Robust and scalable AWS cloud infrastructure (Lambda, Step Functions, SQS, S3) designed for simultaneous, high-volume video processing.
→ Ensured rigorous security and compliance standards via AWS encryption and best practices.
Outcomes Expected
→ Improving per-item processing efficiency by more than 5X, enabling Z5 to reinvest in operational enhancements and customer experience improvements.
→ Accelerated video processing times from approximately 45-60 minutes to just 1-2 minutes per video.
→ Reduced transcription and data logging costs by 92%.
→ Improved inventory accuracy, reducing human error rates.
→ Enhanced customer satisfaction through real-time, highly accurate inventory data visibility.
→ Enabled scalable growth, positioning Z5 for future expansion into pharmacy management and audio-only transcription.
Optimizing Inventory Management to Transform Healthcare
U.S. hospitals discard approximately $5 billion worth of expired or overstocked medical supplies annually due to inadequate inventory tracking systems. Managing inventory accurately in the healthcare industry isn't just important—it's essential.
Traditionally, hospitals relied on large teams—sometimes 40 or more employees—to physically count and catalog medical supplies. This manual process could stretch across multiple weeks and incur millions of dollars in logistical expenses. Moreover, it was prone to human error, causing inaccuracies and delays in obtaining accurate counts.
Founded in 2015, Z5 Inventory aimed to disrupt these traditional inventory management practices. Initially, Z5 streamlined the inventory management process by deploying smaller teams equipped with GoPro cameras to record video counts, which remote workers then transcribed digitally. This innovative approach significantly reduced logistical complexity and transcription timelines from weeks to just days.
Although these innovations represented significant progress, Z5 recognized that there were still opportunities for further disruption, aiming to eliminate the $5 billion wastage experienced annually by the U.S. healthcare system.
Automating Transcription and Data Entry with AWS AI and ML services
The recent emergence and maturity of AI technologies created a groundbreaking opportunity for Z5 to accelerate inventory counting dramatically further. By automating video and audio recognition with advanced AI solutions, Z5 recognized they could achieve even greater efficiency gains.
Partnering with Caylent and leveraging AWS services like Amazon Transcribe and Amazon Bedrock, Z5 implemented AI-powered transcription that transformed hours of video footage into structured and highly accurate inventory data. A process that previously took a minimum of 45-60 minutes per video has now been reduced to mere minutes, with the added advantage of processing multiple videos simultaneously.
Unlocking more than 5X efficiencies in operation
This technological advancement delivered measurable business transformations, including:
- More than 5X more efficient inventory counting processes that unlock resources to scale faster and provide better outcomes to customers.
- Significantly accelerating inventory processing times, delivering near-instantaneous results and greatly enhancing customer satisfaction.
- Improving accuracy by minimizing human errors traditionally associated with manual processes.
Furthermore, this improvement in revenue efficiency is tangibly impacting the company’s bottom line and profitability.
Currently, medical and surgical supplies in hospitals can only be accounted for with up to 85% accuracy due to inconsistent cataloging. Z5 expects to remediate industry-wide concerns and elevate accuracy levels to a targeted 95-98% by normalizing these inconsistencies with AI, analyzing multiple sources of videos to corroborate catalog tagging.
Moving forward, Z5 Inventory expects to expand this AI-powered inventory solution into new avenues such as pharmacy management, where standardized product identifiers promise even higher accuracy. Additionally, Z5 aims to transition to audio-only inventory transcription, accelerating data processing even further and mitigating HIPAA compliance risks associated with video recordings.
Technical Challenge
Z5 Inventory needed to overcome costly, slow, and error-prone manual inventory processes to enable scalable growth and profitability. To achieve this, the company looked to transition from manual transcription methods toward a highly automated, AI-driven approach.
Their desired goals from the new system were:
- Accelerating inventory counting and transcription processes dramatically.
- Minimizing manual transcription to enhance accuracy and ensure consistency in inventory records.
- An advanced infrastructure capable of processing multiple inventory videos simultaneously to handle significant volumes efficiently.
- Ensuring seamless, automated integration of inventory data directly into their backend management system for immediate usability.
To achieve these objectives, Z5 Inventory required a partner with deep expertise in AWS cloud infrastructure and AI/ML technologies.
“The AI solution we’ve implemented with Caylent and AWS has helped us unlock a new level of accuracy and efficiency in healthcare inventory management. What once took weeks of manual effort can now be done in just days—or even minutes. Caylent’s expertise and hands-on partnership were instrumental in making this vision a reality, ensuring a seamless transition to a smarter, faster, and more scalable solution.”
Brandon Burris
CTO
Solution
Caylent and Z5 Inventory evaluated multiple AWS AI offerings, quickly identifying Amazon Transcribe for its exceptional accuracy and efficiency. Leveraging Amazon Transcribe, Caylent rapidly implemented an AI-driven transcription solution capable of converting inventory data from lengthy videos into structured text in mere minutes, reducing transcription times from 45-60 minutes to 1-2 minutes.
Data Structuring with Amazon Bedrock
To effectively structure the transcribed data, Caylent integrated Amazon Bedrock, specifically leveraging Anthropic Claude 3.5 Sonnet to intelligently format transcribed audio data into actionable JSON inventory records. This integration allowed instant updates directly into Z5’s existing backend inventory management system, improving speed and significantly reducing errors.
User Interface and Experience Enhancements
Recognizing the importance of seamless integration and a user-friendly experience, Caylent collaborated closely with Z5’s UX/UI team. Together, they integrated AI-generated inventory data directly into Z5’s existing user interface, creating a streamlined workflow that requires minimal intervention.
Scalable Cloud Infrastructure
Caylent designed a robust, scalable cloud infrastructure utilizing AWS Lambda, Step Functions, SQS, and S3. Terraform was used to deploy infrastructure and it was seamlessly integrated into Z5’s existing CI/CD workflows.
This infrastructure is capable of processing millions of inventory counts annually and simultaneously handling thousands of inventory videos. By leveraging these scalable AWS services, Z5 Inventory now possesses the capacity to significantly expand its operational capabilities and confidently accommodate future growth.
Security & Compliance
Caylent rigorously implemented AWS encryption and security best practices. They validated that no sensitive patient health information (PHI) or protected data were included in processed videos, ensuring industry compliance standards were maintained.
Implementation Process
Caylent maintained proactive, iterative communication throughout implementation, providing frequent demos and updates for transparency. When Z5 experienced internal team changes, Caylent provided comprehensive engineering support, ensuring uninterrupted progress and successful implementation.
