As generative AI gains traction, CIOs find themselves at the forefront of its adoption, balancing the promise of transformative impact with the practical challenges of implementation. This journey is more than a technical undertaking; it calls for a strategic vision on how generative AI can align with business objectives, drive ROI, and create value across diverse sectors, from healthcare to construction. In a recent panel discussion, hosted by DiUS, industry leaders shared insights into how they navigated unique challenges and capitalised on generative AI’s potential, shedding light on the evolving role of CIOs as not just technologists but as strategic and ethical stewards of change.
Generative AI in practice
Our panel featured three industry experts with distinctive applications of generative AI: Mark Green, Associate Professor and Deputy Scientific Director at Monash IVF, who is pioneering generative AI-driven patient support in fertility treatment; Aaron Vanston, CTO and co-founder of BuildPass, using generative AI to streamline construction site management; and Marc Phoa, DiUS’s AI/Machine Learning Practice Lead, guiding clients on adopting generative AI solutions. Despite their diverse applications, a common theme emerged: generative AI must align closely with an organisation’s core business goals, and CIOs are key in steering this alignment.
DiUS partnered with Monash IVF to help deliver a chatbot designed to guide patients through complex genetic counselling, underscoring the consultancy’s strong presence in healthcare generative AI solutions. “Our goal was to give patients accessible, accurate information without replacing the human touch,” Green explained, underscoring the need for generative AI to augment, not replace, professional expertise. This chatbot was tailored to handle routine inquiries, enabling genetic counsellors to focus on complex, personalised issues. Reflecting on his team’s progress, Green noted, “I was genuinely surprised by how much we could accomplish with the time and resources we had. Despite having a very specific focus and defined goals, we achieved significant strides with the chatbot project, delivering results more effectively and quickly than anticipated.” For CIOs, this represents a model of balancing technological benefits with the nuanced demands of patient care.
On the other hand, BuildPass, a construction-focused startup, took a developer-led approach. “We didn’t set out with a rigid plan,” Vanston shared. “Instead, we encouraged our developers to identify real pain points and experiment with generative AI solutions.” This approach allowed BuildPass to integrate generative AI in a way that prioritises convenience and productivity for users without emphasising generative AI as the focal point. Recently, BuildPass secured $7.5 million in a seed funding round led by Carthona Capital, supporting its global expansion of the first generative AI-powered operating system for the building sector. This significant milestone not only fuels the company’s growth but highlights the potential of generative AI to transform traditionally low-tech industries.
Prioritising use cases with the DVF framework
One of the pressing questions for any CIO considering AI adoption is where to start. Phoa highlighted the importance of focusing on problems that align with generative AI’s proven capabilities, carry low risk, and establish a foundation for future AI initiatives. “The most successful strategies prioritise problems that can deliver measurable value quickly and lay the groundwork for scaling AI across the organisation,” he explained.
To support this, DiUS often uses the Desirability, Viability, and Feasibility (DVF) framework to structure the decision-making process. This framework helps CIOs evaluate use cases by asking: Is the solution desirable to stakeholders? Is it viable within the constraints of the organisation? And is it feasible given the current technology and resources? By combining these structured criteria with a focus on immediate, tangible outcomes, organisations can ensure their AI adoption strategy is both impactful and scalable. For example, starting with a knowledge-based search or document summarisation are proven applications that satisfy the Feasibility criteria, offering quick wins that lay the groundwork for the broader generative AI roadmap.
Overcoming internal resistance and building a culture of curiosity
Generative AI’s impact goes beyond technical boundaries, often sparking apprehension among employees concerned about job displacement. At Monash IVF, Green noted that while some healthcare professionals feared generative AI could replace roles, the chatbot was framed as a tool to automate repetitive tasks, enabling staff to focus on personalised care. “We made it clear that generative AI was there to enhance their jobs, not replace them,” Green shared, turning initial scepticism into support.
At BuildPass, Vanston fostered a culture that embraces experimentation and curiosity. “We nurture a culture where experimentation is encouraged. I’ve spoken with developers from other organisations who’ve said, ‘we’re not even allowed to use ChatGPT.’ For us, access to these tools is a given—it’s just part of how we work,” he explained. This open approach encourages the team to freely explore generative AI’s applications, making innovation a natural part of the process and building a stronger foundation for generative AI integration.
For CIOs, nurturing curiosity and inclusion can be essential strategies for transforming resistance into engagement, especially during the early stages of generative AI implementation.
Navigating compliance and sector-specific regulations
The regulatory environment varies widely across sectors, and CIOs must navigate compliance carefully. In healthcare, Monash IVF prioritised patient safety by tightly controlling the data sources accessed by their chatbot, limiting its responses to vetted information to prevent misinformation. “Our goal was to create a safe, informative experience without compromising patient confidentiality,” Green explained. This approach highlights how healthcare CIOs can incorporate safeguards to avoid missteps in sectors where accuracy and privacy are paramount.
In contrast, the construction sector faces fewer regulatory constraints, though Vanston shared that BuildPass anticipates stricter standards in the future. “Even if the regulations aren’t stringent now, we want to stay ahead of the curve in data security and transparency,” he explained. This proactive approach to compliance reflects an awareness among CIOs that as generative AI adoption grows, so too will the need for robust governance and compliance measures across all industries.
Implementing generative AI guardrails and observability mechanisms
Ensuring generative AI behaves in predictable and safe ways is critical. Both Monash IVF and BuildPass emphasised the importance of guardrails and observability in their generative AI initiatives. Green’s chatbot, for example, included strict controls to defer complex queries to a human professional, maintaining patient safety. Observability played a dual role at Monash IVF, allowing for performance tracking and prompt adjustments to improve the chatbot’s responsiveness and relevance.
Marc Phoa elaborated on the broader role of observability in generative AI, emphasising that it aids in monitoring both technical performance and user interactions. “Observability is crucial not only for compliance but also for understanding how generative AI is functioning and where adjustments are needed,” he said. For CIOs, integrating observability mechanisms can help ensure generative AI solutions remain efficient, compliant, and adaptable as they scale.
Managing costs and budgetary considerations
Generative AI adoption brings with it significant budgetary considerations, often extending beyond initial deployment. Vanston’s approach at BuildPass was to use off-the-shelf models in the experimentation phase to validate generative AI’s potential impact before committing substantial resources. “We wanted to ensure that the generative AI solutions were effective before scaling up investment,” he explained. This phased approach can be useful for CIOs, particularly in budget-conscious sectors, where demonstrating value is essential for securing ongoing investment.
Green underscored that in healthcare, compliance requirements drive up costs, making it essential for CIOs to account for ongoing operational expenses. “It’s not just about the initial deployment; it’s about continuous updates and monitoring to keep the generative AI effective and compliant,” Green noted. Planning for these costs upfront is critical for sustaining generative AI initiatives and avoiding budget overruns.
Determining readiness for deployment
Testing is indispensable in generative AI deployment. Before rolling out Monash IVF’s chatbot, Green’s team conducted extensive user testing, involving genetic counsellors and doctors to ensure the chatbot’s responses were accurate and safe. This thorough process helped refine the chatbot and confirm that it could handle real-world scenarios effectively. “We had to be absolutely certain the chatbot would provide accurate information and know when to defer to a professional,” Green emphasised. For CIOs, comprehensive testing helps avoid unintended consequences and fosters confidence in generative AI’s reliability.
BuildPass, by contrast, adopted an agile testing process, deploying generative AI features incrementally to gather real-time feedback and iterate on solutions. This allowed the team to quickly identify and correct issues, maintaining flexibility and alignment with user needs. “Our agile approach supports rapid testing and learning, allowing us to scale generative AI features effectively,” Vanston noted. CIOs can benefit from phased testing strategies to ensure generative AI solutions meet both technical and business expectations.
Measuring ROI: Balancing quantitative and qualitative gains
Evaluating the success of generative AI goes beyond financial metrics, requiring CIOs to consider both quantitative and qualitative impacts. For Monash IVF, quantitative metrics like appointment bookings and genetic test uptake helped measure the chatbot’s direct impact on patient engagement. Additionally, patient satisfaction scores provided insights into the chatbot’s influence on the user experience. “Success meant not just guiding patients but helping them feel supported during a stressful time,” Green shared, underlining the chatbot’s broader impact on patient care.
In construction, BuildPass focused on productivity gains, analysing how generative AI-driven features reduced time spent on repetitive tasks. “For us, success isn’t just about hitting a number but about how much time we save for our users,” Vanston explained. CIOs should approach ROI measurement holistically, considering frameworks like DVF to assess immediate gains while paving the way for longer-term value, including competitive advantage and enhanced innovation capacity.
Navigating the future
For these pioneers, generative AI adoption is an ongoing journey. DiUS, which partnered with Monash IVF on the chatbot project, is seeing a surge in demand for generative AI solutions, particularly in healthcare and financial services. This trend points to the growing relevance of generative AI-driven solutions in regulated sectors, where technology can optimise both user experience and compliance. Vanston envisions a future where BuildPass leverages generative AI to reshape the construction experience, addressing inefficiencies and improving workflows on a global scale, supported by its recent $7.5 million seed funding round led by Carthona Capital. Monash IVF’s Green continues to explore ways to integrate generative AI for patient support, with an emphasis on ensuring each innovation enhances patient care. For CIOs, embracing generative AI is not just about implementing new technology; it’s about strategically leading their organisations into a future where generative AI complements human insight and amplifies impact across industries.