This is the thinking computer's man

He started out in 1990 with £2,000 and a lot of enthusiasm. Today, Mike Lynch is at the helm of a $1.2bn business

Mike Lynch is a soft-spoken, polite individual. Some have remarked on his resemblance to The Master, the fictional adversary of Dr Who, but in real life he's much less forbidding. He's slightly deaf (from nights as a student spent programming synthesisers for rock bands) and passionate about his otter hound dog, Gromit. He's enthusiastic about Tesco's Internet delivery service and numbers among influences Blackadder and an American cartoon that features mice trying to take over the world. All in all, a man who doesn't take life too seriously.

Mike Lynch is a soft-spoken, polite individual. Some have remarked on his resemblance to The Master, the fictional adversary of Dr Who, but in real life he's much less forbidding. He's slightly deaf (from nights as a student spent programming synthesisers for rock bands) and passionate about his otter hound dog, Gromit. He's enthusiastic about Tesco's Internet delivery service and numbers among influences Blackadder and an American cartoon that features mice trying to take over the world. All in all, a man who doesn't take life too seriously.

Except when it comes to business. At 34, Lynch is a player to be reckoned with; his company, Autonomy, is now worth $1.2bn, second only to Freeserve in the UK in market capitalisation. He has come far since his fireman father advised him "not to get into a job where you run into buildings on fire". At Cambridge, he took a summer job at a large electronics firm, and recalls: "It was deadly boring. I enjoyed the technical work, but there was no enthusiasm, no excitement."

He started a PhD, researching neural networks - how to get computers to recognise things automatically - and developing the work of Thomas Bayes, an 18th-century clergyman who spent his spare time formulating theories about probability. Lynch began to think about commercial applications. But back then in 1990, there was no established path.

"A lot of venture capitalists didn't really understand technology. They weren't very impressed," says Lynch. "I ended up being lent £2,000 by an eccentric impresario in a bar in Soho - I think he saw in me the enthusiasm of youth." Thus began Cambridge Neurodynamics.

"The big problem when you start," he recalls, "is that you have a better mousetrap theorem - this idea that the world will beat a path to your door. In reality, they buy the acme mousetrap. You begin to learn about customers, about selling to markets. We only had £2,000, so we had to learn quickly. The most precious commodity was time. A lot is about admitting you don't know what you are doing, and trying to find people who do; having an open mind."

In 1996, Autonomy became an offshoot of his original company as Lynch began to concentrate on the problem of getting computers to recognise words. Computers may work in ones and noughts, but Lynch perceived that they needed to be able to distinguish, like humans, between varying shades of grey. "The way our technology works is to look at words, and understand the relationships, because it's seen a lot of content before. When it sees the word star in the context of the word film, it knows it's got nothing to do with the word moon. Because it works from text, it can deal with slang and with different languages."

Marketing this was the next big battle. "We knew we had to be successful in America. It was a question of Go West, young man, go to San Francisco and be ignored," says Lynch. "They found it hard to believe that anyone from England could have anything powerful. People often use proxies to measure companies, so rather than actually looking at the product, they will ask who your VC [venture capitalist] is. If you don't have one of the star VCs, they won't even look at your product."

The PR cart may come before the techno-horse, but, says Lynch, even finding a PR firm was a battle. "My marketing director and I decided the only way to crack this was to be outrageous and do what the Americans do - make an incredibly over-the-top pitch about how we were going to take over Europe. They realised we had a good story, and worked on how we were going to explain it."

He found the "cold-hearted schmooze" harder. "There are these pseudo-social gatherings, where people will be watching others over your shoulder and will walk away in mid-sentence. When you talk to seriously capable Americans, they are more European; its the upper-middle management who are the most ruthless. I am very shy and find it hard to talk to people, and in the early days that was difficult. One trick is to become incredibly English - as soon as you use that accent, they assume you're intelligent."

Back in Britain, Lynch, who took Autonomy to Easdaq last July, has become an elder statesman of the industry. He commends Gordon Brown for making "brave" tax reforms which will aid enterprise, but bemoans the lack of a decent technology market and of angel money - small investments in start-ups in the UK.

The most exciting development on the cards, he says, is Wireless Application Protocol, which will allow consumers to use their mobile telephones for anything from buying groceries to online banking.

Crucially, says Lynch, this development will happen faster in Europe than America, because of the former's congruence of phone standards. "This is the most interesting place for technology; the rate of change here is faster than I've ever seen, and were still on the cusp," he says.

"In the States, the mobile phone standards are a complete mess. Europe has a key secret position because here, they are coherent. The problem is to work out the killer application."