(FORTUNE Magazine) – If you scan the reams of "best advice" in the preceding pages, you'll notice a pattern: Many of the key advice givers are older and wiser bosses. No surprise there. It's the managers in their 50s and 60s who have had time to develop the most valuable knowledge and experience. But few large companies seem to prize that wisdom anymore. Intent on cutting costs, many employers are trying to get rid of

people over 50, despite rising age-discrimination litigation. That's an exceedingly shortsighted policy. By forcing out the employees with the most experience, companies may be inadvertently pushing critical knowledge out the door before it is shared with the next generation. They'll probably regret it before long, since demographics suggest that business is facing a dangerous brain drain from voluntary retirements alone. And those folks' lost smarts can cost an awful lot to replicate.

Consider the chilling example of the National Aeronautics and Space Administration. Way back in the 1960s it spent $24 billion (in 1969 dollars)--and at one point employed 400,000 people--to send 12 astronauts to the moon. But in the 23 years since the Apollo program ended, the engineers who carried crucial know-how in their heads, without ever passing it on to colleagues, have retired or died (or both). At the same time, important blueprints were catalogued incorrectly or not at all, and the people who drew them are no longer around to draw them again. So to fulfill the Bush administration's promise to return to the moon in the next decade, NASA is essentially starting all over again. Estimated cost to taxpayers in current dollars: $100 billion.

Luckily a few smart organizations--among them General Electric, Dow Chemical, and Northrop Grumman--are working on ways to capture their oldsters' knowledge and disseminate it to younger workers before it's too late. What they're doing may offer a blueprint for other companies. It also reveals how the rules of the career game will change, both for those close to ending their working lives and those just beginning them.

We've all seen the statistics. By 2010 more than half of all workers in the U.S. will be over 40. Tens of millions of baby-boomers turn 60 this year, and the decade ahead will see vast numbers of people retiring, or at least leaving their current full-time careers. According to David W. DeLong, a research fellow and head of the workforce unit at the Massachusetts Institute of Technology's Age Lab (and the author of Lost Knowledge: Confronting the Threat of an Aging Workforce), companies most at risk from the coming brain drain are those "with established, traditional cultures where people have spent 20 years or more. Many of these companies have downsized and cut out the middle, so you now have these bifurcated organizations: old white guys nearing retirement, a lot of young people in their 20s and early 30s, and not enough people in between."

As the "old white guys" depart in droves, plenty of human resources managers fear that the younger workers won't be ready to step up and run the show. Boston-based consulting firm Novations Group recently surveyed 2,900 HR people and found that only one-third are confident that they have enough talent in the pipeline to keep their businesses humming as boomers bow out. "There's an imminent leadership crisis at many big companies," notes Paul Terry, who worked on the survey. "They have less management bench strength than at any time in memory." At companies with nonexistent or inadequate systems for transferring critical knowledge, says MIT's DeLong, "younger managers are going to find themselves thrown into positions that they're unprepared for. Some will no doubt find creative ways to succeed. But many are likely to experience burnout and failure, resulting in even more turnover that the company can't afford." Gulp.

Logically enough, figuring out how to keep critical knowledge starts with identifying what it is and who has it now. Knowledge-management experts divide critical know-how in organizations into two parts: explicit--the kind of technical acumen that is easily quantified, explained, and stored--and tacit, which is trickier. "Tacit knowledge is more difficult to capture and communicate," says Scott Schaffar, Northrop Grumman's director of knowledge management. "It's what is held in our heads and includes facts, stories, biases, misconceptions, insights, and networks of friends and acquaintances, as well as the ability to invent creative solutions to problems." Explicit knowledge is the kind you can learn from reading a manual, while tacit knowledge tends to accumulate over years of experience.

Mentoring is especially useful for transferring tacit knowledge. Dow Chemical's strategy for growing new leaders, which started in 1995, depends heavily on a program of formal mentoring. For example, Dave Keppler, Dow's chief information officer and vice president of shared services, has half-a-dozen assigned mentees, with whom he meets one-on-one every month to track their progress, and a dozen more to whom he's an informal sounding board. "Getting people to believe mentoring is important and to put their time and energy into it doesn't happen overnight," says Dow CEO Andrew Liveris. "But it helps if you start by standing on a burning platform." By that he means, "if circumstances have created a pressing need"--as when, between 1993 and 1995, Dow slashed its layers of middle management from 12 to six. "So suddenly we had supervisors and managers with 60 direct reports," Liveris recalls. "People's enthusiasm for being assigned a mentor really grew out of how difficult it was to get a bit of a manager's time."

But traditional mentoring alone is of limited value, for the simple reason that it usually exposes mentees to only one person's expertise. So defense contractor Northrop Grumman has created what are called communities of practice--companywide groups that meet, in person and online, to share information. "Social networks, people building connections across divisional boundaries, are where the breakthrough innovations are coming from," Schaffar says. "We're seeing a crop of young, emerging leaders coming out of these groups."

One such is Tamra Johnson, 26, an aerospace engineer in El Segundo, Calif., who is working on the Prometheus project--NASA's name for the vehicles that will orbit Jupiter's icy moons. The average age of Northrup Grumman's aerospace engineers is 54, so the company has been hiring young rocket scientists like mad, taking on 2,000 new employees in the past two years. In 2003, Johnson started a community of practice for new hires, which now has about 600 members. The group has brought in recently retired project managers to teach seminars and has arranged for young engineers to "shadow" older ones, following them around for a day or a week to soak up scads of know-how, both explicit and tacit. "Senior management has asked us for feedback on how the company can improve assimilation of new hires," Johnson says. "So the knowledge transfer really goes both ways."

Still another mechanism for transferring expertise is something General Electric calls "action learning teams." Action learning teams put people together from several disciplines--manufacturing, sales, marketing, legal, finance--to solve particular problems. The handpicked young managers who participate, along with older and presumably wiser colleagues, "get exposed to very big projects and issues, but with a safety net. You learn by doing, and you get continuous feedback on your performance." One advantage of the teams, says Bob Corcoran, GE's chief learning officer, is that they "encourage people to learn a lot about a lot of things, not just their own jobs. It reduces the likelihood that when boomers do retire, you'll be left saying, 'Gee, old Alex was the only person here who knew how to do this.'"

GE combines such programs with deliberate fast-tracking of the most promising young people, giving them "stretch" work assignments designed to teach them to think on their feet. Says Corcoran: "We expect Gen Xers to take on a lot more responsibility a lot more quickly than before." Todd Wyman, now 37, was supply-chain manager for GE's $1 billion Mexican manufacturing operations at the ripe old age of 33. He now runs the global supply chain for GE's rail-transportation business. Wyman, who has held nine jobs since he joined GE right out of college, has been on action learning teams that took him to Romania, Belgium, and Qatar and at one point had him shadowing salespeople around Latin America to get a firsthand look at customers' needs and problems. "Giving people really challenging jobs early in their careers is important for companies that want to retain them," he says.

But the most obvious way to retain older people's critical knowledge is to, well, retain older people--at least part-time, until they've had a chance to teach others what they know. One obstacle to that approach: rigid federal regulations governing defined-benefit pensions, which keep workers from collecting any payments until they quit their jobs altogether at age 62 or 65. The Treasury Department and the IRS are mulling new rules, which may take effect next year, that would allow people to phase into retirement gradually. Companies could pay out partial pension benefits while employees continue to work a part-time or flexible schedule. Under the plan, employees (who would have to reach the age of 59½ to be eligible) would still continue to accrue pension benefits despite working fewer hours, and their smaller paychecks wouldn't affect the formula for determining final benefits, Treasury officials say.

Many employers are waiting for the new regulations to take effect before they think seriously about phased retirement, says Valerie Paganelli, a senior consultant who worked on the Watson Wyatt study. "You'll see a lot more of it a year from now," she predicts. "Both companies and employees will have to adapt. The trouble is that boomers often retire because they want to do something different, and they think they have to go elsewhere to do it. The challenge is to find ways to channel that restlessness within the company, so that you don't lose them altogether." Employers who do it well, Paganelli says, "will have a tremendous competitive advantage in the long run."

Happily for employers, phased retirement appeals to a lot of people of baby-boom age and beyond. A recent Watson Wyatt poll of 1,000 people at or near retirement age found that most like the idea of bowing out gradually rather than all at once. In a poll conducted for FORTUNE by TheLadders.com-- a job site for executives earning $100,000-plus a year--51% say they plan to keep on working after they retire from their current careers. By contrast, only 16% aspire to the more traditional ideal of dropping out of the workforce entirely. "Golf gets boring," says GE's Bob Corcoran. He envisions a future, not far away, in which "people aged 65 to 80 will share a job with someone else or work 'core hours,' ten to three, or work part-time and take extended leaves to share their expertise with nonprofits." He says GE has offered those options to employees in mid-career for more than a decade now, and "these people have produced great results for us. So why not at the end of a traditional career as well?"

A few innovators are way ahead of the game. Consider, for instance, Charles Stark Draper Laboratories in Cambridge, Mass., which does high-tech research for the federal government and private industry and snagged the No. 1 spot this year on the American Association of Retired Persons' annual list of the best employers for people over 50. Says Jeanne Benoit, Draper's HR chief: "The average age of our employees is 42, and we have lots of people here who are old enough and could easily afford to retire."

Still, Draper's scientists, and the critical knowledge they carry in their heads, stick around, in large part because Draper offers them huge flexibility--six months off a year or an eight-hour-a-day job three days a week, or whatever works for them. Felsa Satlow, 61, is an engineer who spent 31 years managing projects like the software that keeps track of GPS satellites. Last May she "retired" and took the summer off, and then she came back to work 20 hours a week with full benefits. Satlow spends most of her time recording critical knowledge. "I'm writing down how I planned and ran the projects I managed, spelling out what the steps are, so that other people can replicate what we did and continue to build on it," she explains. "This will probably take me three years. Beyond that, I'll stay as long as I find it interesting."

Another approach companies are taking: pooling resources. For example, in 2001 people at Procter & Gamble started worrying that boomers retiring en masse would decimate their R&D department. Eli Lilly, with which P&G had done some joint research projects, had the same concern. So the two companies formed Yourencore.com, an online network of retired and semiretired research scientists and engineers that matches available brains with short-term R&D projects. Two more employers, Boeing and National Starch & Chemical, recently joined the group, which draws on a database of 470 scientists.

The folks who sign on with Yourencore are paid an hourly rate for their project work and often do their own consulting and teaching on the side. Harvey Algeri, a 64-year-old industrial engineer in Cleveland, has the kind of "retirement" that will doubtless become the norm before long. He teaches two days a week at a local community college and takes on consulting projects, including one so far for Eli Lilly through Yourencore. "It's great because in my own consulting practice I have to drum up work myself, which takes time away from my teaching," he says. Yourencore takes care of the matchmaking and the paperwork--and, says Algeri, "I like having limited projects with a commitment to firm start and end dates."

Of course, there's nothing stopping any company from forming its own in-house version of Yourencore to tap the knowledge of former senior employees. Trouble is, once those employees have retired, there's no guarantee they'll want to come back, even part-time. Are you ready to lose what it's taken them a lifetime to learn? If not, now is the time to get serious about preparing for the brain drain ahead.

FEEDBACK askannie@fortunemail.com