Work IQ: the context behind the answer
The same model.
Different results.
Start with an AI model. It brings what it learned in training. It doesn't arrive knowing how your company works.
The model is the starting point.
Your work creates
the context.
That context already exists in the work people do. Start with email and meetings, where priorities and decisions take shape.
Add the work, not just the document.
A file is only
the beginning.
Now take a pricing proposal. The question is practical: what changed, and what still needs a decision before it goes to the customer?
A familiar task gives the context a purpose.
Find what matters.
Carry it forward.
There may be many similar decks. Work IQ uses your current active work, called a working set, to narrow the search to the one that matters now.
Current work helps identify the right context.
Assemble context.
Skip the chain.
With traditional APIs, an agent may retrieve the file, then its versions, then comments, then related messages. Each lookup adds work.
Retrieval can become a chain of calls.
Bring what helps.
Leave the rest.
Finding context is only part of the problem. Too little can leave a crucial gap and force another search. Too much makes the model read irrelevant material.
More context is not automatically better.
Connect the model
to the work.
Return to the pricing proposal. The useful answer depends on the document, its history, the people involved, and decisions recorded elsewhere.
Keep the task at the center.