SAGE
SAGE is a four-stage loop for improving how AI platforms present your brand: Setup, Analyze, Generate, Engineer.
SAGE is a four-stage loop for improving how AI platforms talk about your brand. The letters stand for Setup, Analyze, Generate and Engineer. You set up what to measure, find out where you're losing, make the changes, and then build them into how your team works, before going round again.
Its value is the order. Many teams jump straight to writing content, without knowing which questions they're losing or why. A loop like this makes you measure first, and check afterwards whether the work did anything.
The four stages
Setup is choosing what to track: the questions customers ask, the platforms they use, the markets you sell in and the competitors they weigh you against. The aim is a starting point you trust.
Analyze is reading the answers to find the gaps. Which questions don't name you? Which sources do the platforms read that leave you out? What do they get wrong about you?
Generate is doing the work: new or better pages, outreach to the sites the platforms read, corrections to wrong claims.
Engineer is making it stick. You measure what changed, keep what worked, and turn it into a routine, so the next round starts from a stronger place.
An example
Harrow & Pine sets up 40 prompts on tea gifts across two markets. The analysis shows it's rarely named for corporate gifts, and that the answers lean on three gift guides that don't mention it. It writes a corporate gifts page and pitches the three guides. Six weeks later it checks which prompts moved, and adds corporate gift pitches to its monthly PR plan.
Where the work happens in AnswerOrbit
You set up in Topics and prompts and Competitors. The gaps show up in Sources and Claims. Recommendations lists what to change, and its Impact tab takes readings two and six weeks after you mark one done.