Featured Projects
CavTech Roadmaps a CyberSecurity Plan for a Houston area school
Project Type: Physical Security and Cybersecurity Assessment
Client: Small Houston-area independent school
Total Time: 2 weeks
Overview
When a Houston-area independent school needed a clear understanding of its physical and cyber risk exposure, CavTech was engaged to conduct a focused security assessment. Within one week, CavTech identified immediate vulnerabilities, implemented high-priority fixes, and delivered a practical roadmap to help school leadership strengthen security over time.
The Challenge
The school needed to validate its existing security posture across both digital systems and the physical campus environment. Like many small education organizations, the client needed fast remediation for obvious risks, practical guidance for longer-term improvements, and leadership-ready reporting that could support investment decisions.
CavTech’s Approach
CavTech began with a rapid review of the school’s cyber environment to identify exposure points and configuration weaknesses. Immediate issues were patched or reconfigured where possible, reducing risk while the broader assessment continued.
The team then performed an on-site physical security assessment, reviewing the campus environment and identifying best-in-class recommendations for camera placement, wireless coverage, access visibility, and overall security readiness.
Solutions Implemented
• Patched and updated configurations to close immediate cyber vulnerabilities.
• Added DNS, DMARC, and SPF records to strengthen email authentication and reduce spoofing risk.
• Implemented limitations for known problem geographic areas and sites using the client’s existing hardware.
• Assessed physical security controls, including camera coverage and wireless network footprint.
• Provided formal cybersecurity awareness training materials to help staff prevent and identify attacks.
Deliverables
At the conclusion of the engagement, CavTech produced a clear readout documenting what was found, what was fixed, and what should be addressed next. The client also received a board-level presentation summarizing findings, completed implementations, future roadmap items, and risk mitigation return on investment.
Results
In just two weeks, the school moved from uncertainty to a clearer, stronger security posture. Immediate vulnerabilities were addressed, physical and wireless security considerations were documented, and leadership received a prioritized roadmap for continued improvement.
Impact
The engagement gave the client actionable visibility, measurable risk reduction, and a practical path forward. By combining cyber remediation, physical security review, staff training, and executive reporting, CavTech helped the school improve protection for its students, staff, systems, and operations.
CavTech Builds AI-Powered Quote Automation for an East Houston Industrial Product Supplier
Project Type: AI Development
Client: Medium to large industrial product supplier
Total Time: 3 months
Overview
An East Houston industrial product supplier was experiencing unprecedented growth, creating a rising volume of email-based product requests that were difficult to process manually at scale. CavTech was engaged to assess the workflow and build an AI-powered process that could identify incoming requests, extract key product details, and present them in an easy-to-manage format.
The Challenge
The client needed a way to manage increasing quote and product request volume without adding full-time staff for manual email triage and data entry. The existing process required team members to review each message, interpret the request, identify relevant SKUs or product needs, and move the opportunity forward. As demand increased, this manual workflow created a risk of slower response times, inconsistent categorization, and missed opportunities.
CavTech’s Approach
CavTech designed a fully customized AI-powered quote automation application built around fine-tuned Meta Llama 4 variants. The solution used AI for email intent classification, request categorization, and SKU or product extraction based on the client’s historical data.
Rather than relying on AI to determine pricing, CavTech routed the structured AI outputs into a deterministic process that can retrieve accurate pricing and product details directly from the client’s ERP system through secure APIs. This approach combines the speed of AI with the reliability of system-of-record data.
Solutions Implemented
• Built a customized AI-powered quote automation application for email-based product requests.
• Used fine-tuned Meta Llama 4 variants for intent classification, request categorization, and SKU/product extraction.
• Passed structured outputs into deterministic workflows for accurate pricing and product detail retrieval from the ERP system via secure APIs.
• Generated draft quotes and placed them into a review queue for human oversight.
• Captured human corrections and quality feedback to support future fine-tuning cycles and ongoing model improvement.
• Provided the initial basis for a CRM-style interface supporting feedback loops and opportunity management.
Deliverables
CavTech delivered an easy-to-use user interface that organized incoming requests, supported human review of draft quotes, and helped manage opportunities through a practical CRM foundation. The solution also included a feedback loop so corrections and quality input could be captured and used to improve future AI performance.
Results
Over the three-month engagement, CavTech helped the client transform a manual, email-driven quote intake process into a scalable AI-assisted workflow. The system reduced the need for additional staffing as request volume increased, while keeping human oversight in place for accuracy and quality control.
Impact
The engagement gave the client a repeatable process for managing growth without proportionally increasing operational headcount. By combining AI-based classification and extraction, secure ERP integration, human review, and continuous feedback, CavTech helped the supplier improve speed, consistency, and scalability across its quote request workflow.