AI Agent Operational Lift for Its (internet Testing Systems) in Baltimore, Maryland

Why AI matters at this scale

Internet Testing Systems (ITS) operates in the specialized niche of custom internet-based testing and assessment platforms. Founded in 1997 and headquartered in Baltimore, Maryland, the company provides software solutions for test development, delivery, and scoring to credentialing bodies, educational institutions, and professional associations. With 201-500 employees and an estimated annual revenue around $45M, ITS sits in the mid-market sweet spot—large enough to have meaningful data assets but likely constrained in R&D bandwidth compared to tech giants.

At this size, AI is not a luxury but a competitive necessity. The assessment industry is being reshaped by generative AI, which can author test content, score essays, and even simulate conversational exams. For a company with a 25+ year legacy, the risk is that nimbler, AI-native startups erode market share by offering faster, cheaper, and more adaptive testing solutions. However, ITS holds a critical advantage: decades of proprietary testing data and deep domain expertise in psychometrics. Applying AI to this data moat can create defensible, high-value features that are hard to replicate.

Three concrete AI opportunities with ROI framing

1. Automated item generation and test assembly. Developing high-quality test questions is labor-intensive, often requiring subject matter experts (SMEs) to write and review each item. By fine-tuning large language models on existing item banks and source curricula, ITS can auto-generate plausible distractors and entire question stems. This can reduce content development costs by 40-60% and cut exam refresh cycles from months to weeks. ROI is direct: lower SME fees and faster time-to-market for new assessments.

2. AI-enhanced remote proctoring and security. The shift to online testing has made exam integrity a top concern. Computer vision models can detect gaze aversion, additional faces, or unauthorized devices in real time. Audio analysis can flag whispers or keyboard sounds. By offering a tiered proctoring service—from fully automated flagging to live review—ITS can reduce the proctor-to-candidate ratio from 1:50 to 1:200+, slashing operational costs while maintaining security. This also opens revenue from high-stakes certification markets that demand rigorous monitoring.

3. Predictive analytics for candidate success and test optimization. Using historical performance data, ITS can build models that predict candidate readiness, recommend personalized study paths, or identify items that are too easy or culturally biased. This transforms assessments from static measurement tools into dynamic learning instruments. The ROI here is strategic: stickier client relationships through value-added insights and the ability to upsell analytics dashboards.

Deployment risks specific to this size band

Mid-market firms face unique AI deployment risks. First, talent scarcity: attracting and retaining ML engineers is difficult when competing with Big Tech salaries. ITS should consider partnering with specialized AI consultancies or using managed ML services to bootstrap capabilities. Second, technical debt: a platform built in 1997 likely has monolithic architecture that resists integration with modern AI APIs. A phased, strangler-fig pattern refactoring is essential to avoid destabilizing core exam delivery. Third, regulatory exposure: assessments are high-stakes; biased AI scoring or proctoring errors can lead to lawsuits and accreditation loss. Rigorous fairness testing, explainability tools, and human-in-the-loop safeguards are non-negotiable. Finally, data governance: test-taker data is sensitive. ITS must implement robust anonymization pipelines and consider on-premise or VPC-hosted models for clients with strict data residency requirements. Starting with low-risk, internal-facing use cases like item generation can build organizational AI fluency before tackling customer-facing features.