(FORTUNE Magazine) – On a cool, crisp Washington evening in mid-February, AT&T Chairman Mike Armstrong strides beneath the ecstatically muraled cupola of the Library of Congress, an edifice so lavishly decorated that Boris Yeltsin once asked how such a building could exist in a country that had never had czars. Armstrong is here to present a $3.5 million gift that will help the Library put its archives online. As he enters, an elderly woman struggles to his side and greets him ebulliently. "I'm just thrilled to meet you!" she says. "I took this off my wall, and I want you to have it." She hands Armstrong an original photograph of her grandparents. It takes him a moment to realize that the man in the photo is Alexander Graham Bell.

It's a weighty legacy Armstrong has inherited, but you wouldn't know it from his nonchalance. Looking a bit like a cross between Daddy Warbucks and Star Trek Captain Jean-Luc Picard, Armstrong, 60, exudes good humor, energy, and confidence. His arrival at AT&T 15 months ago seems to have drawn a collective sigh of relief from the staff, who are cheered not only by his bonhomie but also by the sense that they are finally going somewhere, like passengers on a train that starts moving after being stuck in a tunnel.

Reaching the destination...that's another matter. Armstrong has made deals at a furious pace, and Wall Street, after nearly a year of neutrality, has approved, driving up AT&T's stock price by 35% in the past six months. Now the question is whether Armstrong can execute. His vision is easy to articulate: He wants to sell every communications service American consumers and businesses could possibly want--not just phone service but also video and high-speed data connections. He wants to own the wires that carry those services. He wants AT&T's network to enter the 21st century speaking the language of the Internet, conveying sounds and images in digital form, with far greater efficiency and versatility than traditional phone lines.

This agenda is hellaciously ambitious. It involves nothing less than transforming the $53-billion-a-year giant's business. Armstrong wants AT&T no longer to be a synonym for long-distance calls. He has already spent more than $70 billion, or 1.2 times AT&T's total assets, to make over its personality. He has bought a cable company and a local phone company and a global data network so that AT&T can have a direct line to U.S. homes and businesses and extend its reach around the world. Now he must make all that work while cutting costs and firing legions of staffers, and without spooking Wall Street. He must also lash together units that don't mesh neatly, and harness the egos of their executives.

Driven by Armstrong's design, AT&T sometimes seems to reach greedily in all directions, like the railroads of robber-baron days. In fact, it is bent on becoming an octopus--an imperative that flows inexorably from its peculiar mix of customers and from a single decision Armstrong made early in his tenure. While AT&T's long-distance rivals get most of their revenues from business customers, AT&T gets nearly 60% from its 75 million U.S. residential customers. Armstrong decided that to serve consumers profitably, especially in the local phone business, AT&T must have a direct physical connection to the home, the way it did before spinning off the Baby Bells 15 years ago.

Under rules set forth by the Telecom Act of 1996, the cheapest way to get such a connection is to rent space on the local phone network--or at least it would have been until the Baby Bells tied up AT&T in court over how such arrangements should be priced. The disputes gall Armstrong, who says, "I don't want to wait until the Bells decide to turn on a switch" before being able to offer service to a customer.

Meanwhile, the Bells are sapping AT&T's profits by charging it access fees to complete calls. Armstrong worries that if he waits for regulators or the courts to change the balance of power, AT&T might become vulnerable to an MCI-WorldCom-like merger, the kind in which a little company swallows a big one.

These ambitions and fears help explain the urgency with which Armstrong operates, like a downhill skier who knows he must court disaster to win. Says AT&T Labs chief David Nagel: "It's rare to see a CEO who so readily exposes himself to risk." Armstrong's bravado also helps account for the words Nagel and his colleagues use to describe him--scary, demanding, taskmaster, and maniac. It's a revealing sign of the psychic distance Armstrong has brought AT&T that those characterizations are meant as compliments.

So now comes the hard part--the part where Armstrong and his lieutenants must meld the pieces he has bought into a whole. As Gary Hamel, chairman of Strategos Ventures in Menlo Park, Calif., observes: "A load of testosterone and a lot of pressure from shareholders will get you through big deals, and then what?"

If Charles Michael Armstrong has shown an appetite for risk in his brief tenure, no one who knew him before is surprised. In his last job, as CEO of Hughes Electronics, he parlayed to success one of the biggest all-or-nothing crapshoots in U.S. business history--the creation of the digital broadcast satellite industry. The gambit depended on such imponderables as whether $200 million satellites that took four years to design and build would survive launch.

Armstrong was born in Detroit in 1938, growing up not in the leafy suburbs but in the city proper, where his father was a manufacturer's rep for suppliers to the auto industry. The almost obsessive discipline he brings to his endeavors comes from his mother. He first ran into a wall when he entered grade school. A left-hander, he quickly learned to write--only in the wrong direction. He says: "I can remember my mother sitting with me at the dining room table evening after evening [making me write] 'Left to right! Left to right!'" Armstrong studied classical piano, and his mother had him at the keyboard for an hour a day, six days a week, for seven years. If he missed a session, he had to play for two hours the next day. "We all tend to rise to the expectation of the people we love," he says. "My mother's attitude was that if you want to do it, you can. That instills a tremendous confidence--combined with a good dose of fear of failure."

The confidence carried over to high school. Armstrong was only 5-foot-6 and 126 pounds when he arrived, but he wanted to try out for the freshman football team. The coach wouldn't even let him on the field, saying he was sure to get hurt. Armstrong recalls: "I was so mad that I bought a set of weights and started eating five meals a day." By the time he was a junior, he had grown six inches and bulked up to 180 pounds. He not only made the team, but became captain as well.

Equanimity in the face of disdain became a theme in his life. Anne Gossett, his girlfriend since he was 14, wanted to follow Armstrong when he enrolled at Miami University in Ohio, but her father wouldn't let her. When Armstrong politely asked why, her father replied, "Because you're probably going to end up selling popcorn in Tiger Stadium." That didn't deter the couple. After bringing home a few manifestly unsuitable suitors, Anne transferred to Miami in her junior year, and they married after graduation.

Besides knowing how to wait out adversaries, Armstrong has a gift for winning over people from the podium. He delivers speeches with a thespian's aplomb, often without notes, pausing for effect and with attention to the rhythm of the words. As President Leo Hindery Jr. of TCI, the cable giant AT&T bought, puts it, "Mike is like Elmer Gantry in disguise. He's a gifted spokesman for the company and the stock."

Armstrong's ease masks meticulous preparation that dates from a career trauma at IBM, where, just out of college, he took a job in sales. In one of his first training sessions he had to demonstrate a punch-card machine in a two-day competition staged to simulate a customer sales call. "I was awful," he remembers of the opening day. "I was nervous. I hit the wrong buttons. I fed the cards wrong. It was just a disaster." Worst of all, he says, he didn't deliver his pitch with conviction. His instructor told him to go back to his motel room, stand in front of the mirror, and keep talking until he believed the reflection. He practiced till two in the morning, and the next day he won the competition. He still rehearses for hours the night before a speech, often running through it several times with Anne as his listener.

He takes his hobbies just as seriously. As CEO of AT&T, he has inherited the highest-profile golf tournament in corporate America: the Pebble Beach Pro-Am, which AT&T hosts. In this year's tourney, says AT&T President John Zeglis, "we both played so poorly that we're going to golf school in Florida." Armstrong insisted.

Armstrong worked at IBM for 31 years, and it shaped him as a leader. IBM's legendary CEO, Tom Watson Jr., had idiosyncratic ways of inspiring younger managers, even after he retired and became honorary chairman. When Armstrong failed to answer a question to Watson's satisfaction one day, the boss emeritus chewed him out over the phone. Watson's wife overheard the tongue-lashing and called to apologize for her husband, inviting Armstrong to the house for lunch. When Armstrong arrived the next day, he got one of the best presents a boss' wife could ever give a guy: a used motorcycle. In fact, the motorcycle was Watson's old Honda, which he sold to Armstrong for $100 as a conciliatory gesture after his wife bought him a new Harley. Armstrong has been a motorcycle fan ever since.

Armstrong had plenty of success at IBM--increasing the company's dominance in Europe, for example. But he talks more readily about the hard lessons he learned. Newly in charge of IBM's personal-computer business in 1983, he chose a single manufacturer for the disk drive on the third generation of PCs. When the manufacturer faltered, IBM couldn't deliver. The mistake hastened IBM's loss of primacy in the PC market. Armstrong says: "I'll never single-source a major component again in my life." It remains the darkest blot on his resume.

That's not the only cautionary lesson he learned at Big Blue. He saw it badly misjudge how much it would be hurt by Ross Perot, who quit in 1962 and went on to create the modern computer-services industry. By taking over management of data centers for corporations, Perot's Electronic Data Systems came between IBM and its customers and in some cases steered them to other suppliers.

Armstrong fears that the same phenomenon could repeat itself in telecom as multinationals seek help managing their global networks. AT&T had been trying to turn network management into a lucrative business when he joined. In his first meeting with Richard Roscitt, head of AT&T Solutions, its network-management subsidiary, he green-lighted expansion plans. Roscitt recalls how vehemently Armstrong insisted that Ma Bell not repeat Big Blue's mistake: "He said, 'I'm not going to wait and let the same thing happen on my watch.'"

The new CEO realized when he arrived in 1997 that he had to put AT&T's house in order before he could accomplish anything. He sold businesses that had nothing to do with telecom, like the Universal Card. But more important, operations were a mess. At about 28% of sales, costs were notoriously bloated and had stayed high despite draconian staff cuts in previous years. Armstrong promised Wall Street he would reduce the figure to 22% within two years. He's well over halfway there, cutting $1.6 billion in costs last year and pushing up income from operations a very impressive 22%. It was keeping his promise about cutting costs that helped trigger the run-up of the stock.

Whether the stock will continue to rise is another matter. AT&T faces the dual problem of slow sales growth and fast-rising debt. CFO Dan Somers points out that revenue growth is improving from quarter to quarter; sales grew 4.8% in the fourth quarter of 1998, up from 0.7% in the first quarter. That is progress, but telecom is an industry in which the race may not go to the tortoise.

It's when you parse the revenue figures that you really see how desperately AT&T needs to cultivate new markets. Last year, long-distance service accounted for 86% of AT&T's $53 billion in sales. The long-distance component was flat vs. 1997; a 4% gain from corporate customers was offset by a 4% decline in residential service. Only because of double-digit growth in wireless, Internet, and outsourcing did AT&T's overall sales rise 3.2% last year.

Armstrong's buying spree, meanwhile, has generated a lot of debt. AT&T went to the bond markets this March to sell $8 billion in long-term notes--the largest corporate issue in history. The money helps pay for TCI, the nation's No. 2 cable TV provider, which Armstrong agreed to buy last year for $48 billion in stock, assumed debt, and cash. (The sale closed this March.) To minimize dilution of AT&T's shares and help buoy their price, the company has bought back $10 billion worth since the deal was struck. All told, the ratio of debt to shareholders' equity has risen sharply, from 21% at year-end to 39% now.

To put cloth on the framework of his strategy, Armstrong has assembled a pricey set of assets. Besides TCI, he has acquired Teleport Communications Group, a $500-million-a-year local business phone company, for $13.3 billion; MetroNet, a Canadian phone system, for $7 billion; and the IBM Global Network, which carries data traffic, for $5 billion. He has also signed a joint venture with Time Warner (owner of FORTUNE's publisher) to carry phone calls over the entertainment conglomerate's cable TV systems, and with British Telecom to serve multinationals overseas.

TCI remains by far his biggest, riskiest move. It reaches only 20% of U.S. households--a fraction of the market AT&T hopes to dominate--and yet it is crucial to proving Armstrong's vision of AT&T as the one-stop shop for household telecom needs. TCI and AT&T are as different as merger partners can be--technologically, organizationally, culturally, in the personalities of top management. Making the marriage work will be the ultimate test of Armstrong's skill.

Armstrong sees TCI as the linchpin in his plan to transform AT&T's consumer business. As TCI's networks are upgraded to carry phone calls--a multibillion-dollar process likely to take three years at least--AT&T will start selling packages of service, including local and long-distance phone service, high-speed Internet access, and cable TV, all delivered over the same wire.

The hook for consumers will be a volume discount: the more services in the bundle, the greater the savings. The promise for AT&T is greater still. Explains President Zeglis: "Our residential customers spend $81 billion a year on telecom and cable TV, but we're only collecting $23 billion. That means they're spending $58 billion away from us." Bundling should help AT&T get some of that cash. What's more, if enough consumers bite, AT&T should be able to reduce churn. Of the company's 70 million long-distance customers, two million join and two million drop out every month.

Reducing churn will also help justify the costly revamping of TCI. Every time a cable customer signs up for phone service, AT&T will have to send over a technician to install a Merlin's box of electronics on the side of the house to separate the phone signal from the video feed. Doing so will cost as much as $500 per house. AT&T wants those customers to hang around for more than a few months.

Armstrong's plan was to absorb TCI into a $33-billion-a-year consumer services company reporting to Zeglis, AT&T's exceedingly bright, somewhat wonkish former general counsel. Zeglis, 51, had gained his first operating experience overseeing all of AT&T's services in the five months before Armstrong arrived. The two men were eager to bring TCI into the fold as soon as possible. Armstrong says they were wary of repeating AT&T's mistake with its wireless unit, McCaw Cellular, which continued to operate as a stand-alone business long after AT&T bought it in 1994.

The plan also called for Zeglis' unit to have a tracking stock--a separately traded class of AT&T shares whose value would reflect the assets of the consumer businesses. The rise in the Dow last year had bypassed AT&T, and Armstrong felt investors would pay a premium for shares that tracked its high-potential cable and cellular systems.

Then came the sharp increase in AT&T's share price this past fall. "People finally started to get it in November and December," says CFO Somers. The rise removed the rationale for issuing a new class of stock. Armstrong didn't take long to respond: In January he scrapped the idea of the tracking stock.

With the financial plan in flux, TCI's Hindery saw a chance to lobby for a direct line to the chairman's office. The 51-year-old executive, whose personality shows the mix of charm, irascibility, and iron that seems peculiar to the cable business, takes umbrage at epithets tossed his way to suggest that TCI doesn't belong in mainstream corporate America. "When I joined TCI, people called us 'cable cowboys,'" he says. "I don't dress in cowboy boots. I dress in suits and ties like everybody else." Armstrong and Hindery are quite simpatico, both outgoing, gregarious men with decades of experience running businesses. Armstrong also itched to become more deeply involved in the cable business, knowing that he would be praised or damned depending on whether he could make the TCI deal work.

In February, AT&T quietly announced that TCI won't be part of Zeglis' realm after all. Instead, it will report directly to Armstrong. He explains that with a little time to mull things over, he decided this arrangement makes more sense because, as head of Hughes, he had competed against TCI with DirecTV, the digital satellite broadcaster. "John doesn't have any experience in cable," Armstrong says, "and it's something I know about." Instead, Zeglis will add AT&T's fledgling international operations to his portfolio.

The new arrangement leaves Armstrong with a potentially risky divide. Rather than reporting to the same person, the consumer unit is split in two, with Hindery running the broadband-services group--cable TV, local telephone, and high-speed Internet services--and Zeglis overseeing consumer long-distance, wireless, and low-speed Internet service. Hindery and Zeglis both will market the whole range of services to AT&T's residential customers--Hindery from the vantage of the local operator who knows where to dig up a street without severing the power lines, Zeglis ("the bundling king," Armstrong calls him) as the national telemarketer who will solicit customers from afar. If Hindery signs up a cable customer for long-distance service, he will have to buy minutes from Zeglis. If Zeglis signs up a long-distance customer for cable service by offering two free pay-per-view movies a month for a year, the two honchos will have to sit down and hash out who pays for that discount.

Eventually the halves of AT&T's consumer business--Hindery's and Zeglis'--will have to unite. Hindery's scenario puts him very much at the center. "Probably most of the energy will migrate over to wireline," he says. "AT&T sells long-distance to two-thirds of the households in the country. I sell video to two-thirds of the households in my territories. As we offer local telephony, over time long-distance becomes part of the facilities-based service."

While the TCI gambit plays out, AT&T will also develop other ways to sell telecom bundles to households. Ventures with Time Warner and other cable companies will let it serve customers outside TCI's territory. For areas where cable doesn't reach, AT&T is experimenting with technology that can create the equivalent of a local phone line with radio waves. In a few places, it will use a new technology called DSL to turbocharge copper phone lines rented from the Baby Bells.

The goal in all cases is to help AT&T reduce its dependence on long-distance revenues. As is often the case in telecom, the corporate market is showing the first signs of the new order. According to Michael Keith, head of the business services group, long-distance revenues as a percentage of the total business market are withering away, declining from 75% of revenues in 1996 to 56% last year. Keith expects the figure to fall in a few years to 35%, with data, wireless, and international services making the greatest gains.

Armstrong's other big deals--the joint venture with BT and the purchase of IBM's data network--finally give AT&T a place at the global smorgasbord. Till now, corporate customers have been able to dial into AT&T's data network only inside the U.S. This puts AT&T at a disadvantage to carriers like MCI WorldCom, whose data network extends throughout Europe. Tapping into the IBM network, the joint venture with BT will build a new high-speed global data network of its own, with services in 100 major cities.

Putting these pieces together may be even tougher than assimilating TCI. As Roscitt of AT&T Solutions says: "We've got to peel off a piece from IBM and make it a piece of AT&T." Roscitt is counting on the unified network to help him quintuple outsourcing revenues to $5 billion over the next four years. But the IBM network comes with 20,000 business customers, and for arcane technical reasons, the hardware and software that lets them patch into the network is entirely different from the links into AT&T's data network. AT&T has to decide on a single standard and retool the other piece. In the meantime, Roscitt must be patient. "We'll keep the IBM network intact for some time," he says. "It's too big to mess with."

In teaming with BT, Armstrong hasn't exactly picked the fleetest of partners, either. As one consultant says, "If you had a scorecard of how telcos have taken advantage of deregulation overseas, at the bottom and next to bottom would be AT&T and BT." While Armstrong can make decisions quickly on his own, moving with his leg tied to a partner is another matter. For example, the companies just appointed David Dorman, the highly respected former chief of PacTel, as CEO of the joint venture. When did they start interviewing him? This past August. Dorman now must adjudicate a very ticklish set of decisions, namely which big business accounts will be managed by AT&T, which by BT, and which by the joint venture.

Even if the venture flourishes, Armstrong's plan for overseas is the sketchiest part of his strategy. Unlike the Baby Bells, AT&T has not invested in overseas carriers, much less ventured abroad to run a regular or cellular phone company. It has thus missed a lucrative opportunity: Wireless operator AirTouch, for instance, recently sold out to Britain's Vodafone for $55 billion, partly on the strength of its foreign franchises. Armstrong says he isn't interested in running a phone company abroad but will take a close look at wireless licenses that become available overseas in the next few years. AT&T and BT are also rumored to be making a big telecom investment in Japan; the company has no comment.

Armstrong knows that anytime a boss makes far-reaching changes, he is bound to provoke disputes among his lieutenants. Few CEOs encourage dissent as freely as Armstrong; it is a hallmark of his style. AT&T Labs chief Nagel says: "Most bosses hate conflict. Mike is delighted when he sees us getting at each other." CFO Somers remembers the religious war preceding the TCI purchase--one camp arguing that AT&T should spend billions to buy high-capacity cable systems, another arguing that it should rent space on the ordinary copper phone networks of the Baby Bells. "We'd have great debates," says Somers. "There were no sacred cows."

Armstrong's one commandment is that his managers make and keep promises about what they will deliver. Nearly all his executives describe him as a demanding boss, even if his true nature took a while to reveal itself. Somers says: "You could get lulled into thinking everything's okay by there not being a whole lot of 'Jesus Christ! What the hell's going on? Why are revenues off?' If you walk away thinking he's accepted that, you could be in deep trouble." Strategy chief John Petrillo says: "He doesn't suffer fools easily. Mike is confrontative and demanding. It's okay to disagree. It's not okay to sit and smile. He can sniff out pretty damn quickly when someone's bullshitting him."

Armstrong also shows a deft hand in the care and feeding of large egos. Nowhere was that more evident than in the way in which he defused Hindery's anger over a report that the board had ordered company investigators to prepare on him while AT&T was checking out TCI. Hindery was offended; his reaction found its way into a Wall Street Journal story about his rivalry with Zeglis. "In my industry, at my level, reports aren't done about people," Hindery says now. Armstrong simply told his new lieutenant to chill out. "Look," he remembers saying, "there's no reason to be concerned, because it's not individual. It happened to me, and it's probably happened to everybody I know who has taken a top position in a corporation." It's not that Hindery isn't annoyed anymore; he is. But Armstrong has enough confidence that the difference of opinion leaves him utterly unbothered, so it becomes no big deal.

Armstrong's reign has not been without defections. Jeffrey Weitzen, former manager of the business group, left to become president of computer-maker Gateway 2000. Gail McGovern, who ran consumer long distance, departed for Fidelity Investments. And one really big fish left for richer waters: Robert Annunziata, founder and CEO of Teleport Communications Group, the local telco Armstrong bought in his first big deal. Only last November, Armstrong had put Annunziata in charge of AT&T's $22- billion-a-year business telecom group. In March, Annunziata suddenly quit to become head of Global Crossing, an undersea-cable startup. Consultant Mark Bruneau of Renaissance Worldwide in Boston says: "Annunziata was pushing AT&T to address important questions, like How will consumers and businesses communicate in 2005, and therefore what businesses does AT&T need to be in?"

Still, those who remain seem energized. None of them are likely to forget how close AT&T came to squandering the legacy of the man in that photo Armstrong accepted at the Library of Congress. If they share anything, it's a need to believe that AT&T is finally off its heels and moving forward. No wonder they regard Armstrong as something of a savior, as the one who "converted a slow, hierarchical organization with a utility mentality into a nimble enterprise," as the "the right man in the right place at the right time," in the words of--actually, in the words of this magazine. Those phrases appeared in 1993, and they described Robert Allen, Armstrong's predecessor, who nearly ran AT&T into the ground, or so people say today. That's one legacy Armstrong doesn't want to inherit.

REPORTER ASSOCIATE Julie Creswell