AI Won't Fix a Broken Process - It Will Break It Faster

The idea is seductive: deploy AI across the organization and watch productivity multiply overnight.

That temptation, and its risks, came up in a conversation at Wave by Vento, where Keegan Sard, founder of Keegan Ventures, a Monaco-based strategy consultancy serving enterprise clients across the US, Europe and the Middle East, laid out why most companies get AI adoption backwards.

Sard has spent over fifteen years moving through project management, implementation consulting, and strategy work across Australia, Asia and now Europe. He went independent in 2014 after growing tired of templated strategies handed down by large firms that didn't survive contact with on-the-ground reality.

His core message is simple, if unfashionable: AI doesn't fix bad processes. It amplifies them.

"You gotta go and change the processes because if you put a jerrycan on a fire, it's gonna blow up."

Listen before you implement

Sard's approach to entering an organization is deliberately slow. He wants to hear from the CEO, then the management team, then the people actually doing the work in regional offices - because their view of what's broken is often different from leadership's.

He cites a moment from the conference itself as validation: hearing how a design team approached a major car brand's legacy not with plans on paper but with a book capturing history and intent before translating it into product. For Sard, that's the model - understand before you build.

He's applied the same logic inside client organizations. At one company with 6,000 employees, he set up a dedicated inbox so any staff member could flag frustrations or ideas that might never reach senior leadership. He points to OpenAI's internal "friction" channel, which escalates recurring pain points straight to Sam Altman's team, as the mirror image of the same instinct.

"I think in consulting, we are just too quick to jump in and make changes without listening first" he said.

AI is a tool, not a shortcut

Sard is adamant that AI should not be sold — internally or externally — as a headcount reduction story.

"A team of five can now do what a team of 50 used to do," he said. "So I argue you're not hiring anymore, but you're not firing either." The reframe matters to him: efficiency and productivity gains, not layoffs, should be the language leadership uses.

He's equally skeptical of the idea that AI access alone produces expertise. His go-to example: handing the same design brief to Claude, once with a professional designer prompting it and once without. The designer wins every time, he argues, because knowing what to ask for is still the scarce skill.

"AI empowers you to do the things that you couldn't do before."

Regulation is treating AI as something it isn't

Sard's sharpest criticism is reserved for European regulatory approaches to AI. He sees no meaningful distinction between AI as a tool and other tools businesses have long used without special restriction.

"There's no difference between AI as a tool and Excel as a tool," he said, arguing that watermarking and framing AI as an uncompetitive advantage misunderstands how tools have always worked - from formulas on paper to spreadsheets to now.

He also pushes back on bubble narratives circulating in the industry. With only 2% of the world currently paying for an AI license, he believes adoption is far from its ceiling: "We're nowhere near the bubble yet. We're eighteen months, two years behind a bubble in my opinion."

Reasoning still belongs to humans - for now

Sard doesn't dismiss concerns about AI's effect on human reasoning, raised elsewhere at the conference. But he frames the risk differently: the issue isn't whether people use AI, it's how.

He described pitching a casino client by first drafting ideas himself, then running them through a structured AI exercise — modeling red team, black team and white team perspectives, as a McKinsey-style consultant might. The result, he says, is consistently useful: AI surfaces two or three ideas he hadn't considered, sharpening rather than replacing his own thinking.

"The guys doing AI well are leveraging it to get to a better solution," he said. "It just makes you better. It enhances what you already got."

He's less optimistic about the generational picture, predicting large shifts in what younger people will or won't experience — from driving to entry-level jobs - as automation and cost dynamics shift. He points to Elon Musk's suggestion that compute abundance could eventually fund some form of universal basic income, though he's agnostic on what people do with that freedom, from reinvesting it to "going to Bali and surfing all day."

Positioning expertise without claiming it

On personal branding, Sard is direct about what doesn't work: claiming AI expertise in a market flooded with self-declared experts. His approach instead is demonstration over declaration.

"I don't say that I'm good at AI. I show them what I've done with AI," he said. He builds live, interactive presentations rather than static PDFs specifically to prove the point — if a proposal can't show the time savings and process it used, he argues, it probably isn't as advanced as it claims.

He's also tested distribution directly: in 2025 he committed to posting daily on LinkedIn until he reached a million impressions, which took three months. His takeaway was less about virality and more about discipline - and about avoiding what he calls "AI slop," the wave of generic expertise claims the AI boom has produced.

Choosing clients who are ready to listen

Sard's enterprise focus is deliberate. He favors organizations that already know they need to change and that give him access to boards and senior leadership, where influence is possible. He's particularly drawn to sectors where he hasn't worked before - transportation, sports, banking, wine — because an outsider's perspective tends to land differently than another specialist in the same niche.

What he looks for in a conversation, enterprise or otherwise:

• Willingness to ask for a second opinion rather than confirmation
• Openness to cross-sector thinking instead of templated playbooks
• A readiness to fix process before layering on new tools
• Genuine curiosity rather than transactional networking

The takeaway

Sard's argument isn't anti-AI - it's anti-shortcut. The organizations getting real value from AI, in his telling, are the ones willing to do the unglamorous work first: listening to frontline staff, fixing broken processes, and treating AI as leverage rather than rescue.

"You gotta go and change the processes because if you put a jerrycan on a fire, it's gonna blow up."

The gap between AI's potential and its actual use, by his estimate, is still measured in years, not months. For founders and enterprise leaders alike, the lesson is less about which model to adopt and more about whether the ground underneath it is ready.

Interview

• Keegan Sard — Founder, Keegan Ventures

Event

Wave by Vento — October 2026