Exclusive: Appian's Marc Wilson warns of AI pilot pitfalls

Artificial intelligence might be the most talked-about technology in enterprise today, but it's also one of the most misunderstood.

Appian's Chief Executive Ambassador and co-founder, Marc Wilson, sat down with TechDay, during his visit to Sydney, to discuss the confusion.

"Many executives tell me the board said 'we need to do AI' - as if it's a dance move," Wilson said "They want to say they've done AI, like it's some kind of status symbol."

Wilson, who helped establish Appian in 1999, spends most of his time travelling the world meeting clients, partners and government agencies to help them unlock real value from their AI investments. He spoke with us ahead of Appian's 'Around the World' event in Sydney, and didn't hold back on the state of AI uptake across enterprise.

According to a widely cited MIT report, around 95% of AI pilots fail to scale or deliver impact.

Wilson believes the core reason is a lack of focus on business process. "Too many initiatives start with the tech first," he said. "They're doing the cool stuff for cool stuff's sake. But if you're not connecting it to something meaningful - something tied to strategy or a real business priority - it's just not going to last."

Why AI needs process

For Wilson, AI is a supporting act, not the star. "We think AI needs process to deliver value," he explained. "You're trying to make something your organisation does better, faster or cheaper. That means changing how the business operates - and that's process."

The error, he said, is treating AI as a bolt-on tool or a standalone brain. "AI is a complementary technology, not a standalone one," Wilson argued. "The fears people have - that AI is making every decision - aren't what's actually happening. Most of the real value is coming from embedding AI inside workflows."

He shared a personal example: using Google's Gemini to generate multiple-choice questions for his daughter's history revision. "It's great. Saves me an hour. But it's not groundbreaking," he said. "That's personal productivity AI. What we're seeing succeed in business is far more boring - but way more valuable."

'Boring AI' gets the job done

Wilson calls it "boring AI" - intelligent document processing, compliance support, automated reporting, and workflow optimisation.

"It doesn't make your teenager's eyes light up," he said. "But it's saving millions of dollars, cutting processing times from days to hours, and freeing people up to do more human work."

Australia, he noted, is leading the charge in this space. "We've seen some of the most advanced production uses of AI here - on par or better than other regions," he said. "It might surprise people, but we're seeing strong adoption across government and regulated industries."

From pilot to production

Partners, Wilson said, are crucial to moving AI beyond the pilot phase.

"The channel brings expertise that's close to the customer. They understand the data, the decisions, the history - and how things should be done," he said.

He highlighted a recent success story involving Appian partner Roboyo and the National Injury Insurance Scheme Queensland (NIISQ). "Roboyo built a solution with Appian that embedded Gen AI into the invoice processing workflow. It used to take over a week - now it's under 24 hours," Wilson said. "They're achieving almost 100% accuracy, even better than humans."

That speed, Wilson added, has a very human impact. "When you're seriously injured and waiting on care, delays hurt. Getting that invoice processed faster isn't just efficiency - it's improving lives."

Australia's tight-knit advantage

Australia's channel ecosystem is uniquely placed to deliver those results, Wilson said. "The community here - partners, customers, even government - it's much more close-knit," he explained.

"There's a mutual sense of mission. And that means our partners are more embedded, more aligned, and often more trusted."

Appian is investing locally to keep up. "We're expanding our training and enablement, adding technical support in-country, and making sure our partner teams grow with us," Wilson said. "Our Asia-Pacific headquarters is here, and this region is our fastest-growing globally."

What does good look like?

Asked what success with AI will look like in 2026, Wilson said two things: continuous process optimisation and agentic AI.

"We'll see a shift from one-off AI projects to ongoing improvements - spotting bottlenecks and fine-tuning," he said. "And we'll move from standalone AI tools to intelligent agents embedded in process."

He added, "It won't just be a personal agent booking your flights. It'll be multiple agents working in concert inside business workflows. Directed from the beginning, constrained appropriately, but freeing teams to focus on high-value decisions."

Domain over hype

Wilson also urged organisations to value domain knowledge over flashy tech skills alone. "Our customers aren't using Appian for just one or two things anymore," he said. "They're using it for dozens, even hundreds. So our partners' domain expertise - knowing the industry, the business - is becoming more and more important."

In the end, he said, successful AI comes back to the basics: have a goal, measure the value, and make sure AI is part of a larger plan. "Appian projects are never really done," he said. "There's always a next step, a new improvement to make."

And if those improvements are "boring"? All the better.

"Boring is what's going to fund the future," Wilson said.