Let FIT Guide You on Your Next Emerging Technology Adventure
When looking to learn about a new technology, what’s the first step we usually take? We Google it, right?! That may lead to various definitions, articles, and even white papers published by companies looking to engage with us. Such research may yield the type of information we are looking for as far as background but may not answer the questions specific to our needs – how it may apply to our work and how to get started.
In August, the Bureau of the Fiscal Service, Office of Financial Innovation and Transformation (FIT) in partnership with the ACT-IAC Emerging Technology Community of Interest (COI), hosted an Emerging Technology Day. The event included industry, academia and government experts discussing AI’s impact on the financial sector and showcasing examples of projects happening in real-time with insights on how to get started and achieve results.
As a follow up to this Emerging Technology Day, we completed an interview with IBM Partner and AI Practice Leader, Claude Yusti.
- Describe the process you use to work with agencies in identifying / exploiting AI opportunities.
Agencies look for an AI project to showcase the capabilities and impact of AI. The driver behind this thinking is the hype that AI has acquired over the past several years. Often, this results in a short-term pilot that does not gain adoption or produce a useful long-term result.
A more effective approach is to start with a focus on business processes and the challenges that people are having getting work done. By first getting an in-depth understanding of the work that needs to be performed and the data that people are using to do that work, the characteristics of problems that can be helped by applying AI become much more evident.
These can frequently be smaller, mundane looking problems (such as forms recognition, workflow routing and call center help). This type of role for AI can be seen as lower risk and easier for an organization to adopt. The important objectives are to have measurable performance outcomes in mind, an understanding of how and where AI will help work getting finished, and verification that the needed data is available for the AI to work reliably. - What are the key characteristics traits for a strong AI project?
In terms of culture, the project should be evaluated to understand the impact it’s going to have on the daily work of the people that are going to interact with the AI. What is the organization’s appetite for change? Will there be substantial new skills / retraining required or will it be transparent? Is senior leadership endorsing the adoption or taking a wait and see attitude towards the outcome? Preferably, the project is addressing a problem that has a strong mandate to be fixed and the impact to how work gets done has been fully thought through.
Governance is an equally important topic for any organization adopting AI. The long-term cost and investment to sustain AI is higher than the implementation cost. The benefits of the AI solution needs to be properly aligned in order for the project to be sustainable. An integrated model to govern and manage across the different AI projects is much likelier to succeed. - How do you narrow down a use case for AI?
Use cases are best validated as part of a process that starts with the identifying potential opportunities, exploring ideas for potential solutions and then progresses to a business case followed by an early-stage capability or pilot and then production readiness. Clearly, not all use cases get all the way through the process. They will reach a stage where there is not enough support and evidence to continue with the investment.
It is useful to have a repeatable process that is consistently applied and involves a cross-competence team from business, technology and consumers of the solution. This allows standardization of the tools that are used, a more consistent evaluation of value gained and a portfolio view of the total investment in AI to avoid over-investment in some areas while others are neglected.
Properly executed, this process can become a solution factory and a center of competence for artificial intelligence. It avoids relearning lessons by repeating past mistakes and consolidates the organizations competence in artificial intelligence which is often a scarce resource. It also helps keep the focus on the end user of the solution throughout the process. - What benefits and costs do you consider have the best ROI?
The costs and benefits analysis starts with the business goals of the agency. This includes both tactical and strategic objectives. Some use cases may not appear to have high value immediately. But they are well aligned to long-term needs and are worth the investment because they prepare the organization to be successful. An example is establishing a repeatable process for identifying, validating and maturing opportunities.
For the more tactical opportunities, the costs and benefits can include not only direct savings but also flexibility and organizational capabilities. For example, using automation with AI to help in performing repetitive straightforward activities can free up people to tackle more complicated decision-making that is getting backlogged because of a lack of resources. This may not have a direct dollar value savings but it allows the skills of the agency to be more flexible and easily applied to critical needs.
To find out more about the Fiscal Service, visit fiscal.treasury.gov or follow us on LinkedIn, Twitter and Facebook.
