By Kellyn Coetzee, national head of AI & analytics, KINESSO Australia
Despite the hype, tangible returns often fail to match the headlines. Research indicates 11% of companies report significant EBIT gains from AI, and just 26% move beyond proofs of concept, prompting the question: is AI all sizzle and no sausage?
Today’s models, when effectively integrated and applied, can unlock remarkable value that is still vastly underexploited. Microsoft defines the next phase of AI, Artificial General Intelligence, in terms of financial achievement as generating $100 billion in profits, yet many marketers are still struggling to extract meaningful gains from the AI tools already at their disposal. It’s like owning a Ferrari but never taking it out of first gear. How do we shift from promise to measurable profit?
A major hurdle is scaling from proof-of-concept to enterprise-wide deployment. Only 22% progress beyond prototyping, with just 4% delivering substantial value. Data management nightmares arise when information is locked away in disparate systems, requiring cleaning and enrichment before it’s AI-ready. Unclear ROI is another barrier, as nearly half of IT leaders struggle to prove AI’s worth due to fuzzy use cases and undefined KPIs.
Kellyn Coetzee
The skills gap also plays a role, with data scientists, ML engineers, and broader AI literacy across non-technical teams needing alignment. Adoption stalls when employees feel threatened or fail to see how AI aids their daily roles. Furthermore, resource intensity and “last mile” rollout pose significant challenges. Moving beyond superficial pilots requires significant investments in time, money, and talent, while rolling out AI across multiple functions demands a well-coordinated plan to ensure teams truly understand and measure the benefits.
These obstacles underline why merely offering AI “snacks” or one-off use cases does not translate into an experience that can drive enterprise-wide growth. For AI to become a sustainable engine of value, organisations need to move away from “snacks,” teach their workforce to “cook,” and invest in continuous learning.
A Blueprint for Extracting Growth
It’s not easy, but it can be done, and it has been done. The following five pillars form a comprehensive recipe for turning AI’s potential into widespread adoption and profit:
1. Strategy
A robust, long-term commitment to AI must begin at the top. Senior leadership should embrace the transformative capacity of generative AI; develop an enterprise-wide roadmap that prioritises initiatives based on value, feasibility, and risk; appoint credible, empowered AI leaders to champion initiatives and ensure the company’s strategic goals align with AI objectives.
I’ve worked with visionary leaders whose foresight in applied AI helped champion a future-proof workforce. Strategic commitment is the cornerstone of success.
2. Operating Model for Adoption and Scaling
Companies need a balanced AI portfolio that spans quick wins and deeper transformation by establishing a centralised team that synchronises AI efforts across functions; using agile methodologies for rapid, iterative delivery; and setting up funding frameworks that support nimble, high-impact projects.
Equally important is cultivating an environment where non-technical personnel appreciate both the potential and risks of generative AI. Relevant KPIs and performance metrics must track and nurture the value created.
3. Technology and Data
A robust data and IT infrastructure is the backbone of any successful AI initiative. Organisations must integrate data from varied sources and maintain strict governance; embed rigorous testing and validation in release processes with protocols for human oversight as needed; and develop modular, reusable components that encourage continuous improvement and faster problem resolution.
A well-defined data strategy is indispensable—think of it as your “kitchen pantry” of ingredients that support the organisation’s AI roadmap.
4. Talent
Transformation is as much about people as technology, and organisations need to focus on behaviour change.
This includes implementing governance frameworks that support end-to-end transformation spanning workforce development, process redesign, and change management; offering tailored learning journeys to build critical generative AI skills from fundamentals to advanced analytics; and defining clear roles and responsibilities for AI execution, ensuring teams are recruited, onboarded, and integrated effectively.
If you want everyone to cook, you need to provide the right recipe books (training) and the right utensils (infrastructure).
5. Risk Management
Responsible deployment is key to harnessing AI’s transformative power without unintended consequences.
Companies must establish guidelines ensuring ethical, legal, and risk-managed AI implementation; instil risk awareness in technical teams to mitigate data bias and other hazards; build auditable models that allow for bias checks, accountability, and transparent assessments, and form an enterprise-wide council or board dedicated to responsible AI governance.
Embarking on AI doesn’t require an overnight overhaul. By starting small, whether acquiring AI-enabled tools or upskilling existing teams, organisations can begin incrementally:
• Build: Invest in tools, infrastructure, and policies that foster secure, accessible AI environments.
• Train: Provide foundational AI instruction, from basic prompting to specialised generative techniques.
• Refine: Audit roles and processes for bespoke use cases, then craft departmental best practices.
• Engage: Continuously inform teams through workshops, videos, or podcasts.
Although challenges remain, the gap between AI’s potential and realised value is an opportunity, not a chasm.
Those who move beyond talk, empower their teams, and blend human ingenuity with machine intelligence will shape a transformative era.