ISG, Telstra, Cresta, and ScorebuddyCX Expose AI’s Contact Center Test

What happened

ISG has published its 2026 Contact Centers Buyers Guide, a benchmarking report used by enterprises to evaluate contact-centre technology providers. The guide lands alongside recent moves from Telstra, Cresta and ScorebuddyCX, and together the developments signal a shift in how contact-centre AI is being assessed — away from headline capability claims and towards governance, workforce integration and demonstrable return on investment.

According to CX Today's reporting, buyers and analysts are now scrutinising AI tools deployed in customer service settings less for what they promise in a demo and more for how they perform once embedded into live operations: how well they are governed, how they fit alongside human agents, and whether they produce measurable value once deployed at scale.

The report and the associated vendor activity point to a maturing market in which contact-centre AI is judged against operational and compliance criteria, rather than treated as a standalone innovation showcase.

Why it matters

Contact centres have been one of the most active testbeds for generative and agentic AI, but the emphasis on governance and workforce fit suggests the market is entering a harder, more accountable phase. Enterprises are asking sharper procurement questions — not just "can this AI handle the interaction?" but "can we govern it, audit it, and prove it improves outcomes for agents and customers alike?"

For leaders running digital transformation programmes, this is a signal to treat contact-centre AI as an operating-model decision rather than a point-tool purchase. Vendors that can evidence measurable value and safe, well-governed deployment are likely to pull ahead of those still selling on demo-stage promise alone.

The Renascence take

The interesting story here isn't that AI in contact centres is being tested — it's what it's being tested against. Buyers guides and vendor moves like these are quietly rewriting the scorecard from "can it talk like a human" to "can it be trusted, managed and proven."

Most organisations still buy contact-centre AI the way they'd buy a chatbot demo — on the wow factor of a single interaction. But the real test happens in week twelve, not day one: does the tool make agents faster or just busier, does it reduce escalations or quietly create new ones, and can compliance actually see inside the black box when something goes wrong. Operators who treat governance and workforce fit as launch-day requirements, not post-launch clean-up, will get more defensible ROI — and fewer nasty surprises — than those chasing the flashiest capability on the sales floor.