A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by the Construction Tech Review Advisory Board.
Aia
To Harness AI, Understand Its Purpose for Your Organisation
Philipp Gschöpf
Do not let the artificial intelligence (AI) hype limit the technology’s potential for your organization.
With the emergence of effective and accessible AI-powered tools, organizations will be eager to implement AI to accelerate their work processes. Free AI applications - such as ChatGPT – have proven the technology’s ability to produce quality essays and code with a simple prompt.
Using AI can be an exciting journey for an organization but using AI as a solution without understanding the “why” can lead to mixed results. There are cases of AI projects by large organizations displaying discriminative and inaccurate business outputs, which have incurred excessive costs.
The insurance industry provides a useful case study to understand how asking “why” can maximize the technology’s potential. The industry’s products involve many complex users’ needs and processes. These learnings can be applied to other sectors such as healthcare or finance.
Maximizing AI’s full potential calls for a purpose-driven approach by using technology as a tool to address organizational needs. For organizations keen to maximize AI’s impact, there are four ways to avoid falling into the traps of poor AI delivery.
Scaling Based on User Needs
Without understanding an organization’s needs, innovative AI applications may not scale to meet requirements.
Creating a new AI capability from the ground up can be daunting, but this is addressed by first discovering the right requirements and use cases.
Conducting user research is the first step to understanding customer needs and developing a clear use case for AI. In insurance, it is common that customers to struggle with anxiety in the process of claiming large monetary amounts during periods of ill health, which calls for the use case to accelerate such processes. In some Asian markets, AI has enabled customers to receive eligible claims payments in their bank accounts within 3- 20 minutes of making a claim.
After developing the right requirements with consultation from partners and stakeholders, finding the right solution providers is next. AI research partners, with sizable research hubs, can help AI-powered services be built and delivered smoothly.
Making an Impact
Using AI in every solution can be tempting, but organizations must be intentional to avoid having many initiatives with minimal impact.
An agile approach can help organizations avoid this. An agile start-up methodology ensures that teams can prioritize core features and milestone achievements based on qualified customer learnings instead of random AI learning speeds. Through data analytics and a test-and-learn approach, organizations can have the insights to identify impactful features with user demand and high feasibility. This can be observed in the scale-up of Xiao Bang, which is an AI-powered virtual assistant launched by AIA and Xiao-I, which provides personalized customer call reminders in China.
Building Products Rapidly
The industry’s speed and evolving customer needs mean that AI solutions must be built fast. Specific organizational practices can help to accelerate the product life cycle.
Collaboration hackathons are one way to develop complex projects quickly. In these hackathons, competing solution provider teams can be co-staffed with internal go-live relevant personnel, and led by customer experts to build a winning solution quickly.
A modular approach, such as the “70 – 20 – 10” rule, can help. This rule dictates that 70 percent of a solution is deemed reusable once a solution is live in two markets. 20 percent of the solution can then be scaled with minor adjustments, with the remaining 10 percent re-built to tailor to markets.
Driving Adoption and Trust
An AI solution can exist but have low adoption, due to concerns about trust and inaccuracy.
Fortunately, there is a dramatic shift in AI’s application to high-error-cost use cases, as risk-mitigation techniques become increasingly widespread in the AI creator community. This can mitigate the risk of errors in the insurance value chain, which can lead to high costs.
There Is A Need For Frameworks And Policies To Build User Trust Around The Organization, Instead Of The Technology
There is a need for frameworks and policies to build user trust around the organization, instead of the technology. These policies should see irresponsible AI use as a cost to organizations, with the promise that every AI built is suitable, transparent, and can have its outcomes clearly explained.
If done correctly, these standards can be the way for organizations to drive a meaningful conversation on AI in markets that do not have published regulations. Standards are also a great way to encourage more debate during AI creation and enhance customer focus.
Using AI, the Right Way
Organizations must place their purpose and needs first to avoid falling into the traps of AI delivery.
For example, AIA’s approach is to have “Artificial Intelligence at the heart.” Guided by this principle, AIA believes that organizations should ensure AI goes beyond being an end solution, but rather a way to meet the needs of customers and society at large. This principle has served AIA in effectively and ethically implementing our AI-powered claims processes and Xiao Bang products, enabling new customer experiences for our 32 million policyholders.
If organizations want to spread the benefits of AI, the purpose must be at the heart of their AI journeys.
