AI agents drive enterprise rollouts, spark cost concerns and new tool releases

A wave of new AI‑agent products is hitting the market, promising to automate multi‑step tasks across business functions. ASUS launched the free Zenni Claw assistant, which offers guided installation, ready‑made work, travel and life assistants, and local‑cloud routing to balance privacy and speed. Stanford researchers unveiled Biomni, a “co‑scientist” biomedical AI agent that can read literature, generate protocols, run analyses and propose hypotheses, cutting weeks of work down to minutes.

At the same time, companies are confronting the hidden expense of AI‑agent workflows. Executives report that token consumption – the amount of data processed by models – has risen faster than per‑token prices, inflating budgets despite cheaper models. Experts suggest optimizing model choice, caching, and pruning context to curb spending. In the contact‑center space, Cresta introduced a Training Simulator that uses real conversation data to let agents practice high‑stakes calls in a safe, AI‑driven environment.

Enterprise data‑migration also sees an AI boost: SNP’s Kyano Lorna adds an agentic AI layer to SAP transformation projects, providing real‑time recommendations, risk detection and 24/7, multilingual assistance while preserving auditability. Analysts warn that growing reliance on agentic AI may erode human judgment, likening it to a consulting “con” that could diminish organizational expertise. Finally, developers are advised to manage API usage with project‑specific keys and detailed logs to avoid billing and security issues.

These developments highlight both the accelerating adoption of AI agents in diverse sectors and the emerging operational, financial and governance challenges they bring.