AT&T data chief outlines top AI trends for 2026

AT&T’s chief data officer Andy Markus expects 2026 to be defined by small, specialised AI models working alongside the big headline systems. In his view, large language and reasoning models will run overall “agentic” workflows, while fine-tuned small language models will take on narrow, repetitive tasks where accuracy and efficiency matter. He argues that companies will focus less on generic tools and more on how well their own data shapes these models and drives measurable results.

Markus also predicts AI-assisted coding will become the standard way to build software, shrinking development stages from weeks to minutes. AT&T says it already uses these tools to assemble data products in about 20 minutes while enforcing its usual quality, security and compliance checks. Faster coding, he says, will support on-demand apps that AI agents can adapt as requirements change, sitting alongside traditional, slower-to-update software. Mobile operators are expected to lean into this shift by selling more AI services such as model fine-tuning. Across industries, Markus expects tighter focus on hard metrics for AI accuracy, cost and speed, with pressure on businesses not just to use AI, but to prove it delivers consistent, measurable value.