Guest: Glenn Russell (LinkedIn)
Company: Egen
Show Name: An Eye on AI
Topic: AI Infrastructure

Enterprises are eager to adopt AI, but the business case often gets muddled by hype. Glenn Russell, Global AI Practice Lead at Egen, joined Swapnil Bhartiya to explain how leaders should evaluate AI costs realistically—and why total cost of ownership (TCO) is the true benchmark.

“While LLMs are probabilistic in nature, their cost isn’t,” Russell said. Token prices may have fallen year over year, but overall costs are climbing. Usage patterns, platform pricing changes, and scaling effects are driving TCO upward. He cited moves by companies like Cursor and Anthropic to eliminate unlimited tiers as proof that providers themselves are adapting to higher real-world costs.

For enterprises, the answer is to anchor AI investment in measurable business outcomes. Russell shared the case of a major financial services firm that used human-in-the-loop generative AI to double throughput—from 20 to 50 transactions per day. That KPI, tied directly back to initial investment, delivered a clear and defensible ROI.

He stressed the importance of planning consumption models in advance. Leaders should ask: how many tokens will this use case consume, what does that look like at scale, and how does it map against KPIs? With the right expertise, these questions can be answered before AI reaches critical workflows.

Finally, Russell recommended starting small. Low-risk pilots provide a controlled way to test assumptions, validate ROI, and fine-tune systems before expanding. “It’s possible to plan up front while you’re still small, while you don’t have AI in your critical path,” he noted.

For decision-makers, the lesson is clear: AI adoption isn’t free just because tokens are cheap. Success depends on disciplined cost modeling, tying investments to outcomes, and proving value incrementally.