But for many IT managers, telling their organizations to be innovative is about as helpful as telling someone to have a nice day. It's pretty easy to imagine a few eye rolls among people who have already watched their organizations jump on other expensive and not-always-successful management theory bandwagons.

That's why jump-starting a culture of innovation goes far beyond pinning slogans on walls. For one thing, it's important to create a process for innovation that's distinct from linear project management efforts, which, as Andrews points out, are optimized for decisions, not for trying out new ideas.

This means replacing familiar steps like planning, analysis, design, construction and quality testing with new ones like problem framing, status quo questioning, observation synthesis, hypothesis mapping, prototyping and feedback-gathering.

Key to the process is directing your efforts at a question or problem that you're trying to resolve. Although you want people to think outside the box, you need to impose some type of constraint or else you get what Wycoff calls the "popcorn effect" -- wild ideas bouncing around with no purpose in sight. "The old mind-set was to create a bunch of ideas, get out of the box and let the creativity flow," she says. "But now we realize it's better to spend a lot of upfront time defining what we want to do and exacting criteria for success."

For instance, when developing Partners' innovation program, Woodward polled directors on long-standing business issues they wanted to address. He came up with four: improving patients' telephone interaction with the hospital, establishing a regional data exchange for hospitals to share information, providing better trend data on chronic illnesses and learning how to commoditize Partners' IT infrastructure.