(FORTUNE Magazine) –

In late July 2004, Time Warner's America Online unit assembled 25 or so of its top managers for a meeting at the Ritz-Carlton hotel in Midtown Manhattan. The session, held in a narrow conference room in the hotel basement, was blandly billed as a "strategic offsite." But in truth, though no one came right out and said it, the executives were there to plot AOL's last stand.

Just before lunch, AOL's vice chairman, Ted Leonsis, and Mike Kelly, president of the media networks group, delivered a sobering assessment: Customers were defecting in droves from AOL dial-up accounts to broadband. Competitors like Yahoo and MSN were only getting stronger. To stay relevant--and to stay in business--AOL would need to build a Yahoo-like Internet portal. And to make that portal attractive to users and advertisers, AOL would have to offer up its rich content for free.

Leonsis, a gregarious, sometimes unpredictable sort (the majority owner of the Washington Capitals hockey team, he once shoved a fan who taunted him at a Caps game), was particularly impassioned. He cajoled his audience and cheered the proposed strategy. He also prophesied doom if AOL failed to act. "It was like he used every personality he had," says one executive present at the meeting.

What Leonsis was advocating was nothing less than a redefinition of AOL. Many of the executives were stunned to hear such heresy from one of the few holdovers from AOL's glory days. Another vice chairman (and former Time Warner CFO), Joe Ripp, rose in opposition: In his view such a plan would simply accelerate the demise of AOL's still-lucrative dial-up business. Why should customers pay $24 a month for content that would be available free? Other execs argued that AOL had turned a corner, and with ads returning to the Internet, the unit didn't need such a drastic overhaul.

Just as the discussion got heated, AOL chief Jonathan Miller suggested a lunch break and departed with a small group of his top executives, some of whom opposed the free-content scheme. When they returned, Miller made an announcement to the entire group: We are going to do this.

The July session proved a turning point in what has become an improbable and dramatic renaissance. Wasn't it just five minutes ago that, as far as most of the world was concerned, AOL was DOA? Now it is reemerging as a force on the web. More surprising still, it's being courted as an important potential partner for companies that once dismissed the service as Internet for Dummies. Among the suitors: tech titans Google and Microsoft. The incongruity is striking. It's as if Porsche and BMW were suddenly vying for a piece of Buick. But the attractions are considerable: the more than 100 million unique visitors AOL attracts to its network of sites and gaggle of online utilities such as AOL Instant Messenger, MapQuest, and Moviefone. And now there's the free aol.com website, which was quietly launched at the end of September. The site, a portal offering news, search, and listings, is already piquing advertiser interest.

The technorati's newfound love for AOL is not yet matched by Wall Street or by skeptical executives within Time Warner (which also owns FORTUNE's publisher). Neither will soon forget AOL's calamitous plunge from more than $200 billion in value before its disastrous acquisition of Time Warner to about $5 billion for the unit a few years later. At an estimated $10 billion, AOL's worth today is still only a small fraction of its peak--but it's also a dramatic rebound from the bottom of the crater. It is, after all, an operation devastated by the bursting of the tech bubble, traumatized by a financial scandal that cost its parent company $360 million in fines, and hamstrung by a series of execution and strategic bungles, most notably its failure to see broadband as a serious threat to its dial-up business. AOL was scorned as a dinosaur of the early Internet period. But reports of AOL's extinction, it now appears, were greatly exaggerated.

WHEN JONATHAN MILLER, then 46, took the stage at a Hilton ballroom in Manhattan in December 2002, AOL was nearing its nadir. Its U.S. subscriber base, for years an endless upward line, had crested at 26.5 million and was about to sag. AOL was in the process of missing its internal financial targets by hundreds of millions of dollars. A bevy of executives had been ousted, and the press had begun digging into past ad revenue machinations. Miller, who had been brought in a few months earlier from what is now called IAC/InterActiveCorp, where he had run the e-commerce division, was at the all-day analysts' conference to unveil his strategy for saving AOL. His solution? A lower-priced plan for customers who wanted to access AOL via a separately purchased DSL or cable modem connection. AOL would remain a "walled garden"--only members could get access to its content--and other Time Warner divisions pledged to provide articles, music, and videos exclusively to AOL to bolster the service.

The plan landed with a thud. Critics jeered and derided it as a half-measure. More important, it failed to attract new customers (though it did help slow the number of folks who canceled AOL altogether when they moved to broadband).

Miller's debut had been anything but auspicious. But even as AOL's image remained dismal and Miller struggled to find a big-picture strategy, he quietly began implementing incremental initiatives. Faced with lower subscriber revenues and sagging post-boom ad revenue, for example, he cut costs. As a result, in 2004 the unit managed to deliver 17% growth in operating income before depreciation and amortization, the measure Time Warner uses as a proxy for profitability.

Financial analysts yawned. Many wanted AOL simply to squeeze out as much cash from the profitable dial-up service for as long as it could until it faded out of existence. Others thought Time Warner should unload the entire America Online operation to an Internet company such as InterActiveCorp. But Time Warner didn't want to sell at a deep discount after buying at the top of the market. And company executives were reluctant--with good reason--to dump a unit that continued to generate $1 billion a year in free cash flow.

So Miller soldiered on. His opportunity would come in the very realm that had helped send AOL on its downward trajectory in the first place: Internet ads. After a three-year drought, online advertising began to rebound in mid to late 2003, and this time solid companies were buying ads and paying for them with cash. Big marketers started to look at Yahoo and MSN as important outlets for reaching consumers and building brands. Meanwhile, a private company called Google was thriving by selling and distributing ads based on users' search patterns.

AOL salivated at the revived online ad market. But to get a piece of the action, it would need to make some radical changes. Despite its travails, AOL still acted as if the Internet universe revolved around it. For example, AOL had been using a proprietary system called Rainman (remote automated information manager) for publishing advertisements on its sites. It was unlike any other system in the industry--AOL essentially grew up before the Internet did--and that meant marketers had to recreate online ad programs especially for AOL. Fewer advertisers wanted to bother now that the service was no longer king of the Net. So Miller pushed company engineers in 2004 to convert AOL to a standard ad-publishing system, bringing it in line with the rest of the Internet world.

Worse, advertisers hated dealing with AOL. Even after Miller arrived, the ad sales staff was still alienating its customers. "It was their lack of flexibility," says Jeff Lanctot, vice president of Avenue A/Razorfish, one of the largest buyers of online media. "They came to us not with an open mind about what was best for the client but with a predetermined package about what the client should buy." But the tone began to shift after Miller hired Kelly, a Time Warner executive with years of experience dealing with advertisers. (Kelly, it should be noted, is not the Mike Kelly who was previously Time Warner's CFO.) For example, says Andy Sims of online ad agency SF Interactive, AOL has stopped forcing advertisers that buy ads on high-traffic pages to also purchase ads on less desirable locations.

AOL was learning to woo marketers. But no matter what it did, it couldn't shake the underlying problem: the declining user base for the AOL service. Yahoo and MSN's free content meant they could amass audiences dramatically larger than AOL's. By early 2004 a number of AOL executives had started pushing the concept of a free web portal with at least some unique AOL content. They also began to examine ways of using the AOL-owned sites that already were out on the web, such as Moviefone and IM, to drive traffic to aol.com. And to keep people coming back for more, perhaps AOL should offer a free e-mail service, like MSN's Hotmail, that required users to visit a home page to check their missives.

AOL purists resisted. Free e-mail, they argued, would cannibalize subscribers. (A lot of people hang on to AOL accounts just to maintain the same e-mail address they've had forever.) AOL's content, some of it expensive to produce or acquire, sets it apart from broadband competitors, which offer little more than a pipeline to the Internet. And even the most optimistic forecasts for growth in ad revenue seemed unlikely to offset the losses in subscription revenue.

The notion of AOL's new direction encountered almost no resistance, however, from its corporate parents, says Don Logan, who oversees the unit as chairman of Time Warner's media and communications group. Though AOL executives jousted, Logan says, "I don't remember any pushback at all" from Time Warner CEO Richard Parsons. Indeed, the Time Warner board voiced support in April 2004 when Miller presented the broad contours of a plan to win more ad dollars--in part by creating a web portal. Parsons was clearly intrigued. As he put it, according to people who were at the meeting, "What you're telling me is that there is a Yahoo inside AOL."

IT WAS A HOPEFUL moment. But Miller and his team still hadn't settled on just how much content to give away. It remained a divisive issue, one that came to a head at the July meeting at the Ritz-Carlton. Miller, according to people familiar with his thinking, may have made up his mind before the session and simply wanted the rest of his team to hear Leonsis and Kelly's presentation before announcing his intention to put every drop of AOL on the web.

Either way, by the end of the meeting AOL had something it hadn't had in a long time: a clear, unequivocal strategy. A few months later, in November 2004, Miller outlined a key component of his plan in an internal memo. He reorganized AOL into four groups--access, which would manage the dial-up business; digital services, which would pursue new products such as Internet voice calling; AOL Europe; and "audience," which would include aol.com and the other free Internet sites. The move streamlined AOL and eliminated a dreaded "matrix" management system that had snarled decision-making with overlapping responsibilities that enabled countless executives to veto new initiatives.

With the new structure in place, AOL began to move quickly in 2005. It started making more content available online, beginning with its music site. Then, in the spring, came an unexpected opportunity. Miller got a call from a producer of a global event, modeled on 1985's Live Aid, to combat African poverty. Would AOL like to sponsor the concert and have exclusive online rights to it? Miller agreed within an hour. It seemed like the perfect opportunity to show off AOL's commitment to free content and its video capabilities.

Miller's decision gave AOL less than eight weeks to figure out how to webcast nine concerts spanning six time zones. The slightest technical breakdown could have caused a huge public-relations disaster for AOL. "It was a risk," admits the producer, Kevin Wall, "but they pulled it off seamlessly."

The Live 8 concerts on July 2 showed for the first time that AOL could attract a giant audience--without relying on the walled garden--solely with desirable programming. For months afterward music fans and teens swarmed the site to view clips of Beyoncé in Philadelphia and U2 in London.

Some investors sounded as giddy as 12-year-old girls--about AOL. "The real wakeup call, for me, was the Live 8 concert," says Larry Haverty, an associate portfolio manager for Gabelli mutual funds (which own Time Warner shares). "It was a bang, hit-me-over-the-head realization that here was a really strong brand on the Internet, no matter how badly it had been derided by investors."

Competitors took note too. Microsoft, which is building its own search engine to compete with Google, approached AOL about using the MSN search product. AOL, which has a long-standing partnership with Google (AOL was an early investor and uses Google to power searches on its sites), agreed to listen to MSN's pitch and then suggested the parties explore additional ways to partner. Parent companies Time Warner and Microsoft got involved, and soon the companies were discussing everything from shared sales calls to a joint venture.

Google expressed interest, and even approached cable operator Comcast about joining in a bid for part of AOL. (Yahoo also briefly sniffed around.) For Google, a deal would secure its place as AOL's search engine, a relationship that now brings it about $100 million, or 3% of its revenue. The companies all declined to comment. But Time Warner executives are said to be open to any strategic deals that would help boost the valuation of AOL.

So far, talk of dealmaking has done little to boost Time Warner shares, which now trade at around $18, roughly where they've been stuck for two years. What may move the stock, analysts say, is proof that the aol.com strategy has legs. And though it's far too early to issue a verdict on the new approach, at least one preliminary number looks promising: AOL is on pace for ad revenue of more than $1.2 billion this year, up 20% from 2004.

Fulcrum Global Partners analyst Richard Greenfield says he'll be watching to see if AOL can increase the number of AOL-related pages viewed by Internet users. Advertisers use that measure to determine how long consumers stay online. Time Warner's Logan agrees that's critical: "We need sustainable advertising growth, and we need to make aol.com addictive."

That means divining even more ways to hook new users. AOL recently introduced a PC-to-PC calling service similar to Skype, and it appears to be pursuing a sort of online storage locker for photos and music. Fulcrum's Greenfield thinks the new aol.com is a perfect opportunity for Time Warner to get more of its entertainment and information on the web. Time Warner shareholders would cringe at the thought--it sounds suspiciously like the rationale behind the merger of AOL and Time Warner. It's inconceivable that AOL will ever generate enough cash to justify that ill-fated deal. But boosting the stock price by a few dollars--now, that might be a possibility.

FEEDBACK smehta@fortunemail.com