AI Readiness & Real ROI: Building Model-Agnostic AI That Actually Delivers

Telmo Silva kicks off Tech Data & Reality with John Galbraith and Brock Rowlands, co-founders of Evermore AI, a product-agnostic AI training organization born from the manufacturing floor. Together, they cut through the noise around artificial intelligence and get to what actually matters: people, processes, and data readiness. From spreadsheet graveyards to data sovereignty, this episode is a grounded conversation about what it really takes to make AI work inside a business.

Questions We Asked

• Is AI genuinely transforming how teams work, or are most companies just scratching the surface with individual ChatGPT use?
• What are the real signs that an AI implementation is delivering business impact, and what are the red flags?
• With geopolitical tensions rising, how should companies think about data sovereignty and taking back control of their data?

What We Learned

• AI is a team sport, not a solo tool. Most companies use AI individually. Real business impact only comes when it is deployed at the team level with leadership buy-in.
• Spreadsheets are the litmus test. The number of spreadsheets in a business is a direct indicator of where AI can and should go next. Every spreadsheet is a process waiting to be operationalized.
• You can’t automate what you haven’t documented. Before touching AI, companies need to understand and map their own processes. If it lives only in people’s heads, AI will stall.
• Data sovereignty is no longer optional. The US Cloud Act, geopolitical instability, and rising distrust of hyperscalers are pushing companies, especially in Canada and Europe, to repatriate their data and own their AI stack.

From the Factory Floor to the AI Frontier

John Galbraith didn’t start as an AI evangelist. He spent a decade building advanced production scheduling software for manufacturers, writing lines of code to solve complex operational problems. When ChatGPT launched, his CEO walked in and told the team to call every customer: the old way was over. Reluctantly, John leaned in. What he discovered during customer onboarding wasn’t a technology problem or a product problem.

We’re not in a technology gap. We’re not in a product gap. We’re actually in an understanding gap.
John Galbraith, Co-founder, Evermore AI

That realization is what led him to call Brock Rowlands and co-found Evermore AI: a product-agnostic training organization built entirely around the human side of AI adoption.

Teaching AI Like You’d Teach Email

Evermore’s approach strips away the mysticism. Whether a company uses Microsoft Copilot, ChatGPT, or something else entirely, the fundamentals are the same: identify the problem, find the data, understand who holds the knowledge, then build repeatable workflows around it. The tool is almost secondary. What matters is that teams learn to ask the right questions and know where to get the answers. That clarity, Brock argues, is what makes an AI implementation stick, and what makes it replaceable with any future tool without starting from scratch.

Spreadsheets as a Diagnostic Tool

One of the episode’s sharpest insights is the simplest: count the spreadsheets. Every spreadsheet in a business is an undocumented process, a database whose logic lives inside someone’s head. Forty years of enterprise software have standardized roughly 80% of business operations. The remaining 20%, the stuff that makes each company unique, still runs on Excel and tribal knowledge. That is exactly where AI has the most to offer, and where Evermore focuses its work. The signal that AI is taking hold? Fewer spreadsheets, and people doing more strategic work with the hours they get back.

Data Sovereignty: From Afterthought to Boardroom Agenda

The conversation shifts toward a topic Telmo flags as deserving its own full episode: data sovereignty. For Canadian and European businesses, the question of where data lives, and who can access it, has moved from technical footnote to strategic priority. The US Cloud Act makes clear that data residency is not the same as data security. Brock makes the case for repatriating data: running your own servers, owning your source of truth, and building AI on top of infrastructure you actually control. It is, he argues, not only more secure but ultimately cheaper than paying every SaaS vendor for AI features one by one.