InfoQ’s latest AI engineering presentation treats AI adoption as an organizational maturity problem, not just a tooling rollout. Quotient CEO Lizzie Matusov argues that many teams are spending heavily on AI without seeing matching improvements in software delivery.
The presentation introduces a five-stage framework for engineering organizations and focuses on where teams get stuck. Its core point is that usage metrics such as prompts, tokens, or seats do not prove that AI is improving outcomes across the software development life cycle.
For engineering leaders, the practical shift is to look for bottlenecks: unclear requirements, weak review loops, fragile delivery systems, or misaligned incentives. AI coding assistants may help individual developers move faster, but that gain can disappear if the surrounding process cannot absorb or verify the work.
The format is a conference presentation rather than a product release, so it should be read as guidance and analysis. Still, it reflects a growing theme in enterprise AI: the hard part is no longer buying access to models, but redesigning workflows so that AI output becomes reliable, measurable software delivery.