AI in Business: The Real Questions Quebec Decision Makers Need to Ask Now

July 2025, by Catherine Héroux and Tommy Thierry

Let’s be honest: here in Quebec, just like elsewhere, AI is everywhere. It’s almost instinctive now – we want to automate, predict, optimize. Many organizations dive in enthusiastically, sometimes without a clear strategy. And inevitably, many run into a hard truth: results don’t come as fast, or as strong, as promised.

While Quebec is internationally recognized for its AI expertise, through institutions like Mila, IVADO, and the non-profit Numana, one big question remains: how can we drive true business value from these tech investments?

Where things get stuck

1. Massive investments, uncertain gains

According to Numana, Quebec businesses have increased their digital budgets by 7–10% per year since 2021. Yet productivity gains don’t always follow. Remote work, automation, predictive analytics… are often deployed in silos, limiting their real impact across the organization.

2. Limited visibility on tech costs

Cloud, generative AI, on-demand services… This new technology model promised agility and savings. In reality, many Quebec executives admit they don’t know who is using what, at what cost, or with what impact. Complex cloud billing and scattered initiatives eat away at profitability.

3. Compliance, privacy and cybersecurity: ever-rising demands

AWith the implementation of Law 25, data privacy compliance is non-negotiable for companies. The result? Growing investments in cybersecurity and data governance, sometimes at the expense of innovation projects.

4. A technology debt too often ignored

The IT “patchwork” keeps growing: APIs, SaaS tools, disparate databases. The result? Fragile, unscalable architectures. And in an environment where AI requires clean, consistent, accessible, well-structured data, this tech debt directly hinders innovation.

Crucial note on data:
Many companies are slow to centralize their data in well-governed lakes or warehouses. The promise of AI relies on the ability to leverage robust, standardized, and accessible data, which requires significant cleaning and alignment efforts. It’s a major undertaking, but it is possible to move forward in parallel by building targeted use cases on well-managed partial datasets. The challenge of “ubiquitous data” – the ability to access accurate, real-time information regardless of its source – will, in our view, be the key differentiator for companies over the next five years.

5. Business value that dissipates

Yes, AI and automation tools can save time. But who actually benefits? If an organization doesn’t rethink its processes or business models in depth, ROI remains limited.

How to turn AI investments into measurable impact

Quebec organizations that generate real value from their tech initiatives share several concrete approaches. Here are the best practices we see, whether you’re a growing SME or a large public organization.

1. Implement strong tech cost governance

Through consumption metering and FinOps approaches, some Quebec companies – from financial institutions to critical infrastructure providers, have recovered hundreds of thousands of dollars in unnecessary spend. The goal: pay only for what creates real value.

For example: a distribution company discovered that a pay-per-use predictive AI tool was only being used at 8% capacity.

2. Manage AI projects like evolving business products

Every AI project should be managed like a business product: with clear targets, a lifecycle, and ongoing ownership. Gone are the days of “implement and hope it works.” Measure, adjust, capitalize.

3. Prioritize transformation by high-impact domains

The classic mistake? Trying to automate everywhere at once. The better approach? Choose one strategic domain, supply chain, customer service, operations, and transform it fully with AI.

For example, according to Updata, agri-food companies adopting AI or automation systems typically see ROI within 12 months, including:

15–25% reduction in production waste
40% average reduction in inspection costs
15% increase in overall productivity

4. Focus on your people, not just your algorithms

AI isn’t magic. Without supporting internal teams, tools aren’t adopted or used optimally.

The most inspiring leaders we see invest in continuous learning, adapt their team structures, and create internal AI champion networks, and they’re the ones achieving the most sustainable results.

So, what now?

President / CEO

- Align every AI initiative to a strategic business objective: cost reduction, customer experience, productivity, etc.
- Demand clear performance indicators, reviewed quarterly.

CIO / Head of IT

- Avoid scattershot projects; group initiatives by functional pillars.
- Prioritize systems supporting high-leverage activities.

CFO / Finance Director

- Request cloud consumption accounting.
- Decline investments without measurable short- or mid-term returns.

HR / Training Leaders

- Launch “train-the-trainer” AI programs.
- Promote data and augmented decision-making skills across all levels.

Bottom line: AI is a business strategy, not just an IT initiative

Every dollar invested should generate measurable value. That requires discipline, visibility, and – above all – alignment with what truly matters to your organization.

At Era, that’s exactly what we deliver: AI projects rooted in your priorities, without jargon or unrealistic promises. We tell you what works, what doesn’t yet, and how to achieve true organizational performance.

Want to talk with local, neutral experts? Reach out. We know the terrain.

Sources:

Public data from Numana (CEFRIO) on digital transformation in Quebec
IVADO / Mila business case studies
Law 25 and recommendations from the Quebec Commission on Access to Information
IT indicators from the Quebec Treasury Board