Main content

Partner stories

The LLM tax - why your AI agents are wasting tokens and how to stop it

Celonis' Manuel Haug argues that every time an AI agent makes an unnecessary call to an LLM, they leak a little money. Those trickles can quickly add up to a flood. To stem the flow, enterprises must give their AI agents access to the right operational context – so they know how the business runs, instead of having to guess.

what man in business suit, throwing money in the air, in front of blue sky

The hiring results we measured when we deployed agentic AI

AI budgets are growing but return on investment is harder to find. Workday's Allison Joyce explains what her team learned building AI into its own hiring process, what they're seeing across customers, and how they govern AI in a domain that doesn't forgive mistakes.

Blue background with abstract images of people in white but one in red. Black magnifying glass with AI in the lens.

Cyber resilience in the age of AI - why recovery now defines defense

Prevention and detection still matter, but they are no longer the measure of a strong defence. Patrick Smith, Field CTO for EMEA at Everpure, argues that in the AI era resilience is defined by how quickly an organisation can recover - not whether it can keep attackers out.

Back view of young businessman looking at curved blue screen with cybersecurity concept image and blurred office plaza behind © NicoElNino via Canva.com

More stories

"There's no perfect way to forecast AI spend" - Ensono's CFO on governing costs when vendors keep changing the unit

Ensono's Scott Grossman and Jim Piazza describe governing AI spend in a market where Microsoft reports in credits, Snowflake reports in credits, and no cross-industry standard forces them to stop.

Financial forecast chart analyzing. Graph reflects data behavior over historic period. Statistic research concept © VideoFlow - Shutterstock

Don’t mention the war! (1/3) - how traditional defense firms, Big Tech and frontier newcomers alike are fighting to service the new AI military complex

War, what is it good for? Well, making a lot of money for a lot of hi-tech firms. And given the current macro-political instabilities around the world, there's a lot of opportunity to step up to support the services that protect the nation.

AI amry

Atlassian research coined the AI efficiency paradox. DX's Q2 data now finds it inside engineering, with the Developer Experience Index falling

DX has published its third quarterly report on AI's impact on engineering. This one required rebuilding the methodology from scratch. Deputy CTO Justin Reock walked me through what changed and what the numbers show about where the return on AI is going.

DX Q2 2026 AI Impact Report

Did we build the engine before we worked out the physics? What's the pathway to take after the LLM transformer?

Does AI have the cart before the horse? Zuzanna Stamirowska is CEO of Pathway, a company has just been valued at $500 million on the strength of a post-transformer architecture built on first principles that learns after deployment without all the messy baling wire and duct tape of LLMs built on transformers.

priorities

The LLM tax - why your AI agents are wasting tokens and how to stop it This article is sponsored by: Celonis logo © Celonis

Celonis' Manuel Haug argues that every time an AI agent makes an unnecessary call to an LLM, they leak a little money. Those trickles can quickly add up to a flood. To stem the flow, enterprises must give their AI agents access to the right operational context – so they know how the business runs, instead of having to guess.

what man in business suit, throwing money in the air, in front of blue sky
Loading
A grey colored placeholder image