(FORTUNE Small Business) – One breezy morning last January, Brown Williams, the chairman of Princeton Video Image, walked into his office in Lawrenceville, N.J., picked up his New York Times, and saw his own company name on the front page. Startled and a little worried, Williams started to read. The article described how on New Year's Eve CBS News had used Princeton Video's computer-imaging technology to construct a fiction: A billboard advertising CBS had appeared behind Dan Rather's famous mug as he sat in front of millions of viewers, welcoming in the new millennium.

But the CBS billboard wasn't really there. It was a virtual image, artfully inserted by CBS and neatly concealing NBC's inconveniently placed Astrovision video screen. In addition to marveling at the technology, the Times was alerting readers to what could be a new low in media deception. "I think it raises ethical questions for CBS," one critic told the Times. Dan Rather agreed. He was troubled that CBS would compromise the credibility of his news operation. But for Williams, the front-page coverage was fabulous. Princeton Video (PVI) and its technology had broken into a national forum with a powerful illustration of its potential. "I thought it was an incredible turning point," Williams recalls.

Or at least a critical juncture. PVI's groundbreaking technology, born in a New Jersey basement ten years ago, is turning up everywhere from major-league baseball games to episodes of Bewitched broadcast in Mexico. If you're a regular viewer of CBS' The Early Show, you see L-VIS (for live video insertion system, pronounced "Elvis") at work every morning. It stamps the show's logo (Dan Rather be damned) everywhere from the awning of the Plaza Hotel to the mugs in the Crate & Barrel window shown during the show's fade-in. And if you watched a Phillies or Braves game on TV this season, you (but not the fans at the park) might remember a Century 21 ad behind home plate. With all this business, PVI revenues have soared, to at least twice last year's $1.2 million, and Kevin Wenck, a PVI investor based in San Francisco, expects the company to turn a profit--its first--in about two years. PVI's CEO, Dennis Wilkinson, is even more confident. He believes PVI will pull it off this year.

Meanwhile, this still-small company has helped start a media revolution that it hopes will turn the worlds of entertainment, sports, and advertising upside down. That seems almost inevitable. Consumers are distracted and increasingly wary of the blizzard of messages aimed at them. So the appeal of a more subtle--and invasive--technology is undeniable. But the very power of the approach makes it, as the CBS News incident suggests, controversial. And the industries that have the most to benefit, from Hollywood to Madison Avenue, have proved hesitant. For virtual images to be as pervasive as Williams bets they will be, networks, advertisers, and sports leagues, among others, will have to do business in a whole new way. PVI has a lot of friends to win and people to influence.

And that's on top of its other challenges. From the start, PVI has been a bear to wrestle into shape. Technology, financing, management, and marketing all continue to pose problems. How the PVI team has coped so far is an illustration of the complexity of bringing to market even the most promising product. And it offers an object lesson any tech company can appreciate: If you dally too long in the startup phase, you leave dangerous openings for the competition. In the next few years, PVI will show if it can sustain its new sense of momentum.

PVI's story dates back to the summer of 1989, when physicist Roy Rosser was watching a broadcast of the U.S. Open from his home in suburban New Jersey. A tennis lover, Rosser was frustrated by how commercial breaks interrupted the continuity and tension of the play. He also noticed that the screen was filled with wide-open spaces--on the backboards around the court and on the court itself. Hmmm, he mused. You put corporate logos in those open spaces--visible only to television viewers, and changing as often as sponsors demand--and you get advertising that doesn't interrupt the game but introduces a massive new source of revenue.

Rosser kicked the idea around with Martin Leach, a computer-savvy friend in London, and then with Williams, who had briefly run a spinoff of the Princeton Plasma Physics Lab, where Rosser worked. A 20-year veteran of RCA, Williams approached former co-workers at Sarnoff Corp., formerly an RCA company. Everyone on the growing team was confident and excited.

But it took three years for Williams, Rosser, and team, working mostly out of Williams' basement, to raise the $3 million needed to fund a prototype. "We had the damnedest time trying to get the venture community interested in this," Williams recalled recently, speaking in a modest corner office decorated with the baseball caps he has picked up since PVI began its move into the sports world. "The people who were sophisticated about the TV business believed it was technologically impossible to do. The technologically sophisticated people thought the TV advertising business was too prosaic" to be a fertile field for investment.

Eventually, the fledgling company convinced some investors that the technology could go places, and over the next three years it raised $20 million. Major kinks in the technology had also been worked out. In 1996, L-VIS had its national debut during the Sugar Bowl broadcast. Everything went smoothly. But it was followed by a resounding silence.

"As it turns out, the TV industry was doing just fine without PVI," says Williams. His team realized that "the inertia of the industry had become a major problem," he says. Something else was clear too: The self-described "technology guy" needed help with marketing. In early 1997, Williams hired Douglas Greenlaw. He was a "wonderful guy," says Williams, and as a former president of Multimedia Inc. (which owns TV, cable, and newspaper operations), he seemed the perfect man to push PVI into the market. But a little more than a year later, Greenlaw left, and PVI was searching again. Williams says that Greenlaw left for "personal reasons." Enrique Senior, a managing director at Allen & Co. and a PVI board member, is more blunt. Greenlaw simply wasn't effective, he says; his stint with the company was "wasted time." (Greenlaw, now chief executive of Banyan Worldwide's Internet subsidiary Switchboard Inc., didn't return several calls for comment.)

But with CEO Wilkinson, the company seems to have hit its stride. A former top exec at Primestar, the direct-broadcast satellite company, Wilkinson has signed agreements with CBS, the NFL (for virtual ads in international feeds of Super Bowl 2001), the Indy Racing League, and Jack Nicklaus Productions. And PVI has expanded beyond sports, placing ads for Harrah's and Nordstrom during the Grammys in March. PVI is also working closely with companies such as RealNetworks to use PVI technology in Internet-distributed programming. The most promising L-VIS application is one that will let Internet advertisers send targeted ads to individual consumers' homes. Say you and your three kids are watching an old episode of Seinfeld over Internet videostream. You might see Jerry standing on a New York street corner next to a new Volvo station wagon; your single brother across town will see him standing next to a Jaguar.

Although Wilkinson has dramatically picked up the pace for PVI, he's got his work cut out for him. While the company was losing time because of its marketing naivete and with its former CEO, competitors got more than a foothold. The biggest rival, Sportvision, was launched just two years ago by three top News Corp. executives. They had made a splash while at News Corp.'s Fox unit with the FoxTrax hockey puck, which made the puck visible to TV viewers as it flew across the ice. Sportvision relies on revenue from "game enhancements" such as that one rather than advertising. It was the first to use a virtual first-down line in football broadcasts. ESPN is now using Sportvision's "first and ten" technology. PVI--which is trying to increase its business in this area--is supplying CBS with something similar. But it's fighting a Sportvision suit over patent rights. PVI itself has sued another rival, five-year-old SciDel, over a different patent. Both cases are scheduled for argument in early 2001. Wilkinson is clearly eager to settle the cases, but Sportvision and SciDel don't appear to be. Sportvision won't comment. A SciDel executive says the company has not infringed on any PVI patent. These cases are far from fatal. "It's one of these murky situations where in the end, cooler heads will prevail and they'll get on with their lives," probably cross-licensing their products, says Jeff Davis, director of research at SmallCaps Online, an online investment bank.

But the litigation is putting a crimp in PVI's already constrained cash flow. The company's revenues are climbing nicely, up 135% in the nine months ending March 31, to $1.9 million. But losses accumulated over the past ten years total $48 million. And going to the market for cash isn't a likely alternative, because the stock has lately bounced around--and dipped below--its 1997 initial public offering price of $7 a share.

Probably the biggest wild card in this game is the one that--so far--interests journalists more than it does advertisers or stock analysts. CBS makes good use of PVI in its Early Show, but when it comes to the Dan Rather controversy, CBS spokeswoman Sandy Genelius sounds defensive, even months later. She won't offer a CBS newsperson or executive to talk about it. "You can talk to me," she bristles. "Fire away." But then she's pretty opaque on the subject. "There was a lot of internal discussion both before [PVI technology was used on the Rather show] and after it was used," she says. CBS has no plans to use it on a news show again.

If such discomfort grows, it could discourage potential users of PVI's technology. The new rules of a virtual age are creating new sensitivities, and it's hard to predict where they will lead. PVI executives make a classic argument--that technology is neutral--and contend that market forces make any abuse unlikely. Sure, PVI has the potential for doing "all kinds of terrible stuff," says Williams, such as putting fans in the stands who aren't there. But virtual imaging itself is no more than a communications tool, he argues, like the ability to edit an interview filmed by one camera so that it appears to have been filmed with more. He doesn't think this incarnation of subliminal advertising raises ethical problems. And he says debate about it is healthy, encouraging "people to use these kinds of technologies in ways that aren't deceptive." Advertisers and programmers aren't going to do anything that would hurt their consumers or reputations, he argues, because that wouldn't be in their own self-interest. Other PVI partisans are even more diffident. The CBS incident was much ado about nothing, says Allen & Co. partner Senior. "My view is, 'Get a life,'" he says.

Even assuming that most media players agree with Senior, PVI's biggest challenge is to get more of them to use its technology. To illustrate how easy this job of persuasion should be, Brown Williams notes that his 15-year-old daughter uses a VCR to tape episodes of Buffy the Vampire Slayer without the ads--then passes tapes to her friends. Newer technologies, such as TiVo and ReplayTV systems, make it even easier for viewers to evade advertisers' messages. Whether media executives like it or not, they know that this resistance to ads carries an obvious mandate: Make commercials (or whatever we will call them) unavoidable. And that should mean good things for PVI. "Sports advertising in the U.S. alone is a $10 billion business," says Davis of SmallCaps Online, who has a buy on PVI stock.

But such logic isn't yet enough, especially in a transitional era when traditional advertising still holds a lot of power. For one thing, how do you price PVI's services to make sure everyone gets a piece of the revenue pie? The NFL's annual Super Bowl ad bonanza is a case in point. Advertisers pay $2 million apiece for the 30-second ads. To protect the advertising punch of those commercials, the NFL prohibits signs actually installed in the stadiums that are visible to television viewers. With that rule in place, how could they allow virtual signs? "In a sense, they could be looked at as diluting the ads the networks are selling," says Dennis Lewin, the NFL's senior vice president of broadcast and network TV.

Hollywood raises other questions. In March 1999, UPN ran a test for PVI technology. It slipped ads for products such as Blockbuster Video and Evian into an episode of--fittingly--Paramount's sci-fi series Seven Days. The images were unobtrusive unless you were looking for them. Sure enough, the public was either oblivious or unconcerned. "Not one viewer called. No one said boo," recalls Layne Britton, a former UPN executive.

So why is no one is rushing to repeat the effort? One reason: Advertisers seem to prefer placing their products during the original taping of a show, says Paul McGuire, a UPN spokesman. Back to the pricing problem. How do you construct a model to pay for the ads and reap the benefits? How do you price the placing of, say, a tube of Colgate toothpaste in a sitcom? Who gets what cut of the revenues?

Wilkinson says that PVI has recently come up with a revenue stream that works for broadcasters, using the price of a 30-second spot as a base. He won't name names, but he claims that several shows this fall will use the technology. All big Hollywood studios are "actively discussing" it, confirms Barbara McMahon, an independent agent who arranged the Seven Days test. "It's not if" it's going to happen, she says. "It's when."

That's the kind of prediction PVI loyalists like Roy Rosser want to hear. Rosser, the founder who was first inspired by a tennis game, has learned a lot in the past decade. Raised in Zimbabwe, he even had to master the rules of baseball to put some early ideas into practice. Now a consultant to PVI and starting a new company, he says he's excited about his technology's future. "It's taken a lot longer than we originally thought," he marvels. "We were pretty confident. Maybe it was ignorance. We didn't realize quite the full level of the challenge."