A steerable, token-efficient reasoning model, now available on the platform
Kari StehrSoftware Engineer
Carlton KozeyApplied Scientist
Neil DhruvaEngineering
Julie MillsPMM
Bonnie QuigleyMarketing Technology
:quality(80))
A snappier, more steerable experience that’s better at following user instructions.
Why it matters
Enterprises are moving from experiments to production, and the bar for AI output has changed. Teams expect answers they can act on, grounded in the systems where work actually happens, with permissions respected end to end.
That shift puts pressure on the context layer. Without it, models over-fetch, reason in loops and still miss the signal that matters most to the business.
What we learned
Start with the workflows that already have clear owners and measurable outcomes.
Invest in shared context once, then reuse it across every surface.
Measure quality and cost together, not one at a time.
We will keep sharing what works as more teams bring AI into their daily operations.
