If Silicon Valley hadn't been invented and you wanted to find a hotbed of high-rolling business entrepreneurs, there's one place you probably wouldn't look -- inside American Telephone & Telegraph Co.

Not the stolid, hulking phone company.

That's exactly the viewpoint that AT&T Vice President William P. Stritzler found as he toured the country's venture capitals 18 months ago.

Stritzler's job, then and now, was to tap the ideas of the most creative innovators inside AT&T's huge empire and help build a series of new businesses for the company.

Most of the experts told him he was wasting his time, Stritzler recalled.

Big business isn't where new ideas come from. The kinds of people who invent home computers, video games and monoclonal antibodies won't tolerate the rules, red tape and ego-strangling chains of command in the giant corporations, conventional wisdom says.

"When I started, wandering around talking to the entrepreneurs, the one theme I heard was, I was working for a dinosaur," Stritzler said. He could look far inside AT&T for the geniuses, the type-A firecrackers and the long-shot gamblers who start successful new ventures, but he wouldn't find them, Stritzler was told. "It had been bred out of us," he was told.

Stritzler doesn't mind repeating the warnings now, because he is starting to assemble the first bits of evidence to show the skeptics they were wrong.

AT&T recently announced the first new product from its venture technologies campaign: a portable testing device that can identify defective spot welds in automobiles, bridges, buildings and nuclear power plants, for example.

Another new venture is about ready to be hatched. AT&T engineers have developed a computer system that can project images of near television-picture quality on a personal computer terminal. The invention, called a Video Display Adapter, will be "one of the more significant microcomputer products of 1985," according to Computer & Electronics magazine.

While a personal computer's digital display shows only one or two distinct levels of brightness and maybe seven or eight different hues, the AT&T invention can project an emormously greater range of colors on a computer screen, in color-postcard quality, Computers & Electronics magazine said in its April issue. It would permit personal computer users to skim through illustrated shopping catalogues or view pictures of homes for sale, for instance, or greatly improve the appearance of videotext services, the magazine noted.

In all, there are six such projects in AT&T's pipeline, including a computer system intended to solve the recordkeeping problems of large hospitals and another that applies AT&T's know-how in teaching computers to understand voice commands.

Each venture has been developed by a team of AT&T employes who dreamed up the idea and then pushed it far enough uphill to win Stritzler's go-ahead and financial support from his $20 million venture budget.

That money permits the venture sponsors to develop prototypes in close contact with potential customers. The last step is winning the backing of the managers of AT&T's existing lines, to bring the ventures within the bureaucracy.

How do Stritzler and his staff decide which ventures to support?

There isn't a cookbook formula, he says. The projects have to have a solid connection to some technological or market strength AT&T already possesses -- the metal testing device is about as far afield as Stritzler's team will go. The ventures have to be potentially big enough to justify the effort, but small enough or risky enough to fall through the cracks in the AT&T organization.

The personality of the project sponsors counts a lot, Stritzler said.

For one thing, they had better be determined. The metal tester proposal was rejected repeatedly -- "once a month for eight months," Stritzler said, but the sponsor kept at it, refining the idea and trimming development costs.

Once, when Stritzler grilled the sponsor about his qualifications for managing a large-scale product development operation, the best he could come up with was the time he managed a church flea market. But his account of how he did it helped sell Stritzler.

The inventors of the Video Display Adapter agreed to risk part of their AT&T salaries to help pump capital into the project. In return, if the project succeeds, they stand to recover the amount they risked and, potentially, a lot more in profit-sharing.

The successful sponsors share an audacity and enthusiastic belief in their projects, a willingness to step outside the protection of the AT&T bureaucracy, and a lot of confidence in themselves, Stritzler said.

His greatest concern is not that the ideas won't emerge. There are plenty of entrepreneurs inside the company -- more than he can help.

The worry is that AT&T may lose patience with the process. It takes six or seven years for a new venture spawned by a big company to begin making significant profits, Stritzler said. But venture managers in big companies last about half that long, Stritzler has concluded.

If he's right, AT&T will be an exception. He's betting that its long-standing belief in technology will overcome the inertia of one of the corporate world's largest bureaucracies.