CGI: Why AI Adoption Faces Gaps Despite Growing Investment

Despite rising investment in AI, a critical implementation gap persists. 

According to CGI research, many organisations struggle to move AI projects from proof-of-concept to production, stalled by policy and process gaps, legacy systems, persistent talent shortages to implement and maintain AI solutions as well as limitations in existing data foundations. 

Here, AI Magazine sits down with Diane Gutiw, PhD, Global AI Research Lead at CGI, to analyse this friction.

Drawing on insights from CGI’s 2025 Voice of Our Clients research, she diagnoses why enterprises are stuck and details the necessity of building robust, responsible AI frameworks: not as brakes, but as the essential engine for achieving scalable, competitive advantage.

What’s stopping most companies from moving AI projects from proof-of-concept to production?

When we look to our 2025 Voice of Our Clients research – covering 1,813 discussions with 1,477 CXOs across government, banking, retail, manufacturing, energy, insurance, healthcare and communications – we see a significant momentum when it comes to AI: 36% of the organisations engaged in the discussions are now implementing traditional AI and 26% are implementing Gen AI. 

Yet there’s still a critical implementation gap for many organisations due to three barriers that constantly emerge. 

First, legacy systems and technology constraints affect 46% of organisations. These enterprises are trapped in siloed systems that are not integrated, creating data fragmentation that undermines AI deployment whether you're running a bank, hospital or government agency. 

While you can develop pockets of AI solutions or models you cannot scale AI and achieve real value with a patchwork infrastructure as like human reasoning the more information you have available the better decision you can make – gaps in information result in incomplete AI outputs and insights. 

Second, are organisational challenges. Companies are missing the operating models, policy frameworks and governance structures required to manage AI risk and validate secure AI solutions. 

In addition to the governance challenges 69% of clients interviewed noted they are struggling to hire relevant talent; the resource and capability gaps are real and widening. 

Third – and this is where many stumble – companies haven't built the adaptive foundations necessary to scale for production AI. They jump to use cases without strengthening their data quality and accessibility, infrastructure scalability and/or data management and governance frameworks – which is like building a skyscraper on unstable ground. 

Likewise, many companies which are moving ahead have moved beyond disconnected experiments with AI and are now focusing on building integrated ecosystems where data, governance, infrastructure, talent and security work together.

Our data suggests that organisations with holistic AI strategies show 6.6x higher Gen AI maturity – and those with robust data strategies see 5x higher maturity. In other words, being digital leaders isn’t just about being smarter, leaders tend to be more systematic. 

Leaders are recognising that success doesn’t come from any one tool, technology or quick win. It comes when organisations adopt a pragmatic approach to scaling innovation. 

It also comes with an ecosystem mindset, viewing AI not as a single technology but as an integrated network of capabilities, enablers and policies tightly tied to data and data management, are not only more mature in their AI capabilities, but also achieve more meaningful outcomes. 

For example, agentic AI, by the nature of the tool(s), is moving this dial forward – by focusing on solving specific problems and adding efficiencies and change management into clearly defined workflows and value-based outcomes.

How can businesses build responsible AI frameworks that don’t slow down competitive advantage?

This question reveals a fundamental misconception: that responsibility and speed are at odds. CGI data proves otherwise.

Responsible AI isn't a brake on innovation: by providing clear guardrails for use of new technologies, it becomes the engine that enables sustainable competitive advantage.

CGI adopts this practice in its Responsible Use of AI framework that is built on one critical principle: governance by design, not governance as an afterthought. 

When you embed responsible AI principles from day one – aligned to organisational values and ethical principles – you actually accelerate deployment because you avoid the costly rework, regulatory penalties and reputational damage that come from bolting on governance and mitigating risks of bias, reliability and data management later in the process.

Across our 1,813 client discussions spanning every major industry, the pattern is identical. 

Organisations that move fastest are those that establish guardrails to enable innovation that are aligned to their values and regulatory requirements from the outset, with humans in the loop throughout. 

CGI’s framework integrates responsibility into the entire AI lifecycle – from “Envision” where we articulate a responsible human-AI future, through “Explore” with ROI-led use cases under responsible governance, to “Engineer” where we build governance into operating models and finally “Expand” where we scale responsibly.

The competitive advantage comes from trust. 

When customers, citizens, employees and regulators trust your AI systems because they see transparency, fairness and accountability built in from the start, adoption accelerates. 

For example, in banking, responsible frameworks around lending prevent discriminatory outcomes that would trigger regulatory action. In healthcare, governance around AI leveraged for diagnostics and treatments ensures patient safety. In government, frameworks for citizen-facing AI-assisted services protect privacy and ensure equitable access.

The key is viewing responsible AI not as a compliance checklist but as a strategic enabler. 

Our framework empowers clients to realise AI benefits while mitigating risks and shaping a positive human-centric future. The organisations succeeding today reject the false choice between speed and responsibility, recognising that lasting competitive advantage requires both.

Which AI investments deliver the fastest return and which require longer-term commitment?

Understanding the ROI timeline is critical for resource allocation and stakeholder management. 

We see distinct patterns across industries and our 2025 Voice of Our Clients data shows the maturation happening: 36% of organisations are now implementing traditional AI (up 9% year-over-year) and 26% are implementing Gen AI (up 13%). 

The key to achieving ROI and value is to identify the intended benefits at the outset, designing an approach to monitor and measure the value and then ensuring that value is achieved in production or adjusting early.

From initial analysis, fast returns from AI tools tend to come from operational efficiency and customer-facing automation where the intended value is built into the design. 

This could be in the form of process automation, intelligent customer service and support, fraud detection in financial services, predictive maintenance in asset-intensive industries and intelligent document processing across government, insurance and healthcare. 

These initiatives share common traits in that they are well-defined problems, already have good quality existing data sources, clear processes for automation and it easy to measure cost savings from KPIs.

Longer-term investments often represent foundational layers – what we call “Engineering” in our four Es approach. 

This is when organisation commits to the infrastructure that will bring them scaled success; things such as building adaptive operating models, establishing enterprise data governance, investing in organisational readiness, creating scalable cloud infrastructure and developing AI literacy. 

All these processes take time but with them in place they multiply the value of every use case you deploy. 

For instance, a bank that invests 18-24 months building a unified data platform can then deploy dozens of AI use cases rapidly. A healthcare system that focuses on rethinking how we work with AI and leverage organisational change management can scale AI-assisted diagnostics across specialties. A government agency that establishes enterprise-wide governance can safely innovate and deploy citizen-facing AI services.

Transformational investments are ones that completely reshape business models and create new revenue streams. They use AI to fundamentally change value propositions or transform the way a business orchestrates complex processes. These of course require more planning, investment and strategic buy-in across the company.

Rather than just throwing tools at an existing problem, organisations that are achieving value from investments in AI examine their processes, their players and their intended outcomes – and are rethinking how they achieve their goals.

Without the proper culture in place, there is a risk some programs may fail. 

The strategic approach is portfolio management: balance quick-win use cases that build confidence and fund the journey with foundational investments that enable scale, while selectively pursuing transformational opportunities aligned to your competitive position.

What ecosystem partnerships are essential for companies without in-house AI expertise?

Strategic ecosystem partnerships are essential for organisations pursuing AI at scale and the talent dimension is particularly acute. 

CGI’s 2025 Voice of Our Clients research reveals that 69% of clients continue to have difficulty hiring IT talent – and the AI skills gap is even more severe. 

Organisations are competing for a limited pool of AI specialists at a time when demand far outstrips supply across every industry sector. But it's not just about finding Gen AI and agentic model engineers, data scientists and data engineers with the right technical credentials. The real challenge – and what separates successful AI initiatives from failed experiments – is finding professionals who combine deep and applied  AI expertise with genuine industry knowledge. 

These are people who understand the operational realities of your business, who can walk a factory floor and identify where computer vision could catch quality defects – or spend time in a call center – and recognise where natural language processing could transform customer service or shadow clinicians and see where AI could augment diagnostic accuracy. 

They're professionals who aren't afraid to get their hands dirty on the shop floor, in the branch office, at the warehouse or alongside frontline workers.

This means that organisations need practitioners who can bridge the gap between theoretical AI capabilities and practical business problems. 

These individuals speak both the language of algorithms and the language of operations. 

It’s the scarcity of this hybrid talent – technical depth combined with domain expertise and operational pragmatism – that makes partnerships with firms that have both AI capabilities and industry experience essential. 

It's not enough to work with pure-play AI vendors who bring algorithms but lack context. Without access to the ability to bridge domains, organisations are confused about where to place their bets and money.

You need partners who have spent decades working in roles across your industry, who understand existing workflows, your regulatory environment, your competitive dynamics, your operational constraints and your business model.