(FORTUNE Magazine) – Jeff Arnold's angel investors all vividly remember their first encounters with the raw, rookie entrepreneur who had this way of making them write million-dollar checks. Atlanta telecom entrepreneur Boland Jones remembers how the fresh-faced kid approached him after a speech: "I want to be like you," said Arnold, who then wangled a meeting and an investment. Lucius Burch, a Nashville venture capitalist, remembers the afterglow: "I told my wife, 'I've just met an extraordinary young man, and I know I'm going to make five times my money on him in three years.'" Jim Gilstrap, a retired securities trader in California, recalls how little he knew about the actual product Arnold was pitching. "I don't care about the cardiac thing," he told Arnold. "I'm betting on you."

Because all three did bet on Arnold's Atlanta heart-monitoring service, they made staggering fortunes when he took the proceeds from its $25 million sale and parlayed them into an Internet juggernaut called WebMD. In little over a year the company's value rocketed to $20 billion. Even by the standards of Internet mania, Arnold operated at a frantic pace, crafting deals nonstop, including a merger with WebMD's main rival, Jim Clark's Healtheon. He persuaded some of America's biggest companies to be partners and investors: Microsoft with a $250 million investment, Du Pont $220 million, News Corp. an 11% stake; the list went on and on. Reaching into every possible nook of online health care--processing claims, purveying pharmaceuticals, promoting wellness, and more--WebMD was the undisputed king of the space.

Then last year the crown tilted. WebMD's stock plunged; the "boy wonder" label frayed; and the last major acquisition Arnold engineered proved his Waterloo. He ceded half of his CEO job to seal the deal and then resigned from that awkward arrangement in mid-October. By all appearances he was just another dot-com casualty, washed up at 30.

Wrong. Even as he left WebMD, the idea for his next thing was germinating. Barely a month after his Oct. 12 departure, Arnold had developed his first pitch. Now, a mere three months later, the enterprise is flying out of the starting gate. Just as he developed an ideal vehicle for the Internet land-rush era, he believes he has now developed a great play for the Internet dust bowl: a combination turbocharged vulture fund and souped-up Web-navigation "delivery engine." It's called the Convex Group.

How does he do it? At a time when the question might be, Would you buy a used portal from this man? Arnold's project has been embraced by highly credible entities. There's the blue-chip consulting firm McKinsey & Co, which has a big strategy role in Convex. There are the Ziff brothers, who are investing some of their publishing fortune. There's Wall Street veteran Eric Gleacher, who is Arnold's fifty-fifty partner in Convex.

At a time when dot-com venture money has widely dried up, Arnold can still open the spigot wide. After he made a presentation to a roomful of fat cats in California at the Del Mar Country Club, his erstwhile angel Gilstrap announced, "I'm betting on this kid again." The kid cleared $9 million that day.

Now, what he presented then may only be the second cousin of the business that goes out the door in March, when Convex is supposed to be up and running. Arnold's business plan is a fluid document, and the details slosh around a lot. His investors care not: In Jeff they trust. "He'll tweak it every day, and that's what you're betting on with Jeff," says Bert Ellis, a WebMD angel who has signed on again. "If shit happens, Jeff will adapt faster than the market."

Ask around more broadly about Arnold, however, and it's almost as if there were two of him out there: the brilliant "front end" Jeff and the dysfunctional "back end" Jeff. Investors' views of him depend greatly on the stage at which they got involved with him. If you are one of the big corporations that became WebMD's investor-partner as it was peaking, your investment is probably underwater and you're probably happy to have CEO successor Marty Wygod unwind your deal. "Jeff is quite visionary strategically," says one such company's liaison to WebMD, "but the flaw is that he doesn't know how to pull it together and operate."

That said, Arnold's helter-skelter assemblage of valuable assets gave WebMD a foundation to make it a rarity among dot-coms: a long-term survivor. Fourteen Wall Street analysts have buy recommendations on WebMD at its current stock price of around $9. A number of them upgraded only after Arnold departed. That's his mixed legacy, and that's why his next gig is a big test. Is he strictly a creature of the Internet frenzy period, when business models trumped business operations and a plausible pitch fetched billions? Or does he have such a keen sense for the zeitgeist of any age that he can play the right chords and profit again? To some the startup Convex is a bit of a grab bag, with such disparate parts as an Atlanta lawn-and-garden Website and a Manhattan Web navigator. But it's no more far-fetched than WebMD was in its infancy; the real question is Arnold's execution.

I was prepared for Jeff's business athleticism but not for his peculiar magnetism. I expected to encounter the lead of What Makes Sammy Run, but what you get is more like Kurt Russell in The Computer Wore Tennis Shoes. He is clean-cut, earnest, unfailingly pleasant, usually clad in an open-collared button-down shirt and, on a dressy day, a blazer. He talks fast, pausing only long enough to check on whether listeners are trailing along in his verbal wake by asking, often, "You know what I mean?" (pronounced: youknowwhatImean?). He's a great salesman, and he seduces with facts and figures. He can make them sing and dance and then congeal into a vision. It will be a big one. Think of him as a fresco painter, one of his serial investors told me. He can start with a blank expanse of plaster and make opportunity emerge in bold colors, sketching the riches available and the route to attaining them.

The fact is that Arnold may be a creature of the new economy, but he operates most comfortably and effectively in the old one. He has never been one of those disingenuous dot-commers saying, "It's not about the money"; he is an Internet entrepreneur with the soul of an industrialist. When his ship came in with WebMD, he bought a mansion fit for a prince. (A previous owner was one.) For WebMD's 2000 Super Bowl party, he hired Elton John to perform. When he figured a special golf outing would help seal his Microsoft deal, he brought a group from Seattle to the 1999 Masters and gave it the run of Augusta National, where WebMD's vice chairman, Billy Payne, the Atlanta Olympics organizer and a pretty atypical dot-commer himself, was a member. (Arnold didn't stay for the whole tournament; he had to excuse himself to do another deal.)

Seldom has so much startup money been raised with so little venture capital. FORTUNE 500 CEOs, out to sea in the foggy new world of business, saw him as a lighthouse beacon. "We walk in there and say, 'Here's the problem and here's the solution and here's the deal,'" he says. "These companies were all struggling about what their e-business initiatives were going to be, and we made it very easy for them to do deals with us, because we gave them solutions, prepackaged."

It wasn't all so neat in practice. Some deals were undercut by subsequent deals. McKessonHBOC severed relations when WebMD didn't honor its deal's exclusivity terms, according to sources familiar with the matter. Some partners were sorely disappointed at the gulf between vision and execution. "They did not deliver on what we'd expected them to deliver on; they were spread way, way too thin," says Richard Tarrant, CEO of a health-care data-networking company called IDX Systems. Arnold acknowledges that conflicts over exclusivity arose as deals piled up. To the charge that he was overextended, he pleads guilty with an explanation: "We were out assembling the assets."

WebMD's corporate partners also got restive because there turned out to be one important constituency to which Arnold couldn't make the sale: the customers--in this case, doctors. Microsoft and Du Pont committed $330 million between them to subsidize doctors' subscriptions to WebMD, and the doctors still weren't interested. For one thing, health-care experts say, Arnold didn't really understand their needs or the system's true nature. For another, there was no "there" there, in terms of WebMD's content and utility. "Even today, we have doctors who look at WebMD's practice portal and go, 'Where's the beef? How's this making life easier for me?' " says Craig Lanway, chief information officer of Hill Physicians Medical Group, a doctors' network in San Ramon, Calif.

Well, as they say in Atlanta, fiddle-de-dee. "I didn't really care if the doctor believed in or liked the system," says Arnold. "I cared if I could sell the CEOs...because once the assets were assembled, then you could prove to the doctors it was worth their time. But until you had all the pieces in place, it didn't jell. It didn't make any sense to them, youknowwhatImean?"

The trouble is, the pieces never really did come together under Arnold, which is a rap he'll have to overcome and one reason a 60-year-old Wall Street veteran named Eric Gleacher figures big in Convex. Gleacher is the governor on the engine that Arnold lacked at WebMD. Once the head of Morgan Stanley's M&A unit, he now has his own firm, Gleacher & Co., which handled a lot of WebMD's deals.

It was Gleacher who planted the seed of this new venture, on the very day Arnold resigned from WebMD. Gleacher sat Arnold down and told him that these times reminded him of the early '90s, when American business was smarting from its '80s excesses. In the void created by people running away from junk bonds and LBOs in those post-Milken days, there had actually been many opportunities for the brave. "You're going to get a bunch of phone calls, and in my opinion, don't go take another job," he recalls saying. "I think there is a different opportunity here. You have this anomaly where you've had a psychological meltdown in an industry and economic segment which is growing and is extremely important. Why not step back and figure out how we can take advantage of that because you've got your skills; we've got our skills."

When Arnold did get one of those phone calls--from a recruiter at AOL--he acted. "We don't want to go there for a job interview," he told an associate. "Let's go there with an opportunity."

As Arnold pondered, he saw two intriguing plays: one, to fix the Web's broken advertising model and two, to pluck gems from the dot-com rubble. But how to create some plausible ideas for a business in the mere two weeks until the AOL meeting? The answer was right in Atlanta, where a former WebMD colleague introduced him to McKinsey & Co.'s Internet specialists. "We just started talking about the idea and the opportunity, and right away they said, 'Well, let's start work,'" Arnold recounts.

What they'd come up with by Nov. 15, when Arnold was to meet with AOL, was a company with two main components. Part one: Convex would develop a sophisticated delivery engine that would enable advertisers to reach consumers on a far more targeted and economical basis. Arnold knew that for a lot of big advertisers, the Internet thrill was gone. They'd grown disenchanted with paying for banner ads that yielded scant results. The smart Web navigation device that Arnold proposed to develop would discern Web surfers' interests, based on where they alighted, and suggest relevant related Websites. Say you're on the Internet, looking at BMW sport-utility vehicles on that company's Website. The navigation device would enable a pop-up window to appear on your browser that would recommend other comparable SUV sites, say Mercedes; reviews of luxury SUVs in Car and Driver; and other sites of interest to such a shopper. Arnold saw two big selling points. Convex would assure an advertiser of a well- targeted customer prospect, and it would be paid by the advertiser only for actual hits on the company's Website--a CPV (cost-per-view) model, in McKinsey-speak. Another virtue of the delivery engine would be to drive traffic to online categories that Convex aimed to dominate, bringing us to...

Part two: consolidation plays. Just as he'd developed WebMD to dominate the online health-care space, Arnold saw the raw potential in other categories. It was a buyer's market for struggling dot-coms that had burned through their cash and faced two choices: sell cheap or die hard. Convex would identify the best spaces, "roll up" the best players, spruce up their operations, and sell them at a handsome profit. Arnold particularly liked the idea of developing e-tailers and flipping them to big retailers, to be their turnkey online subsidiaries. He saw Convex striking "put" deals, whereby a retailer would buy an e-tailer on certain terms at a certain point in time if Convex could meet certain targets.

Arnold's meeting with AOL in mid-November didn't lead to a deal, but it turned out to be a starting point for the new company. While he saw a major portal partner eventually being an important part of his strategy, he felt he should get some assets in place first. This was a departure from the WebMD approach, of course, but it reflected the counsel of his graybeard partner, who often reminded Arnold of the changed times. "It was all about speed then," says Gleacher, "and it's a different environment now."

It sure is a different Internet world. Treacherous too. Let's play devil's advocate and consider the many things that can go wrong for Convex. Arnold isn't the only vulture picking over the carrion. There are plenty of other opportunistic buyers, some with considerably more spending power. Eco Associates, out of Austin, Texas, has a $100 million fund, and it has invested $30 million to $40 million in four troubled dot-coms, including WebMD's old rival, Drkoop.com.

Arnold has limited his startup capital to $20 million and as a consequence limited his company's ambition--at the outset, at least. He was very keen on getting into eGovernment at first but, when he found it would take a good $100 million to develop the space, he backed off. Instead he is fertilizing the lawn-and-garden category for a fraction of that and, well, you get what you pay for. Convex has acquired Planetgarden.com, the B2C arm of a Georgia garden-supply chain, with the idea that it will provide the starting point for a big rollup. This has been a particularly difficult e-tail segment, however. A competing site, Garden.com, once considered the leader in this sector, folded in December after piling up losses of $75 million. Its assets were purchased recently by Wal-Mart and Burpee for $4.4 million. Arnold, optimistically, points to that distress sale as evidence that big retailers are interested in buying dot-coms and turning them into subsidiaries.

The Convex delivery-engine business is a good idea, but for that very reason it's occurred to plenty of other smart people. Is Arnold really telling us the nascent navigator company with which he's got a tentative deal, New York-based Annotate.net, has found a magic bullet? "If they can convert on this, they can be really successful, but so far no one can achieve high click-through rates, mostly because of the limitations of personalization software," says David Halprin, a senior analyst at the market research firm eMarketer.

What exactly is the connection between a sophisticated delivery engine and a category-killer dot-com? Arnold says the delivery engine will drive traffic to Convex-owned Websites, making them more valuable. But there is another, simpler connection: Arnold and Gleacher already had personal financial stakes in Planetgarden.com and Annotate.net, respectively. Arnold insists that they searched the whole bargain bin for distressed dot-coms before settling on their own properties. "The idea that either Eric or I would cast a whole business opportunity around a million-dollar personal investment is ridiculous," he says. "It's meaningless compared to what we're creating here."

Those two companies will eventually be joined by others, of course, but Arnold will not be on any WebMD-style pace. He's doing a lot more due diligence now before agreeing to a deal. In this environment, Convex needs to be profitable from the outset and it can't afford surprises. "There's no margin for error now," he says, and in this buyer's market, "no need to rush."

But if the becalmed market is his friend in driving hard bargains with distressed dot-coms, it is his enemy in terms of striking lush deals with advertisers. Two years ago CEOs were ready to leap at anything Internet; now they're inclined to drag their feet. In the words of WebMD angel Jouko Rissanen, Arnold was "a pastry chef--10% substance and the rest hot air, until he had the assets to cook a full-course meal." Now he must be a mason, building brick by brick and impressing the customer with results.

That wasn't his strong suit before, and, though Convex is no longer a two-man band and is building a management corps, even some of Arnold's biggest fans wonder privately: Can Jeff really adjust to the times? Is he really on to something big again or just on the rebound? They seem almost protective of him, for, despite his riches and accomplishments, he is still a work in progress. The kid is 31 now and has two kids of his own, but he's still apt to lead his Convex troops out of a boardroom and onto a basketball court for a pickup game. If anything, he's gotten even more compulsive about being in constant motion, preferably aboard a private jet. The best hope is that Gleacher really can be a governor on Arnold's high-RPM engine, lest Convex veer opportunistically but dangerously down a slippery road.

Even so, Arnold remains a pied piper to the smart money and some smart people. He got $20 million from individual investors without breaking a sweat. He so captivated a supersalesman named Wil Slagle with his sales job on Convex that Slagle, who'd begun a breakfast in Atlanta pitching Arnold to invest in his startup, ended the meal by signing up for Convex. Arnold offered to make him the company's sales chief, but added they couldn't talk much longer because he had a flight to catch in 40 minutes. "I'm coming with you," said Slagle. "I don't want to miss one minute of this company." And so, with full faith in Arnold and no change of clothing, Slagle picked up his briefcase and joined the road show.

It was quite a diverse group he encountered as he clambered aboard the jet: some McKinsey consultants, some WebMD alums, some dot-commers...and the alpha to them all, Jeff Arnold. Soon they would be soaring high over the Eastern Seaboard and the rubble of dot-coms, off to New York City with the kid to make a bundle off the apocalypse.

REPORTER ASSOCIATE Jessica Sung

FEEDBACK: jhelyar@fortunemail.com