How Artificial Intelligence Can Inform Public-Sector Decisions in Texas

A panel held at this month’s Government Technology* Texas IT Leadership Forum featured leaders from the public and private sectors sharing their thoughts on how public-sector leaders can use artificial intelligence to make more informed decisions.

Moderated by Sean McSpaden, senior fellow of the Center for Digital Government*, the panel featured Tyler Technologies Vice President of National Sales Angela Fultz Nordstrom, Google Public Sector Customer Engineer Flora Huang and Lower Colorado River Authority (LCRA) Enterprise Data Architect Rhiannon Sanchez.

Sanchez opened with a point about how to evaluate the long- and short-term consequences of decisions, positing that controlling inputs and instituting decision-supporting frameworks can be more productive than evaluating decisions based on long-term outcomes, which can have unintended consequences.

“We have our decision-making inputs that go into the moment of the decision, we have the implementation decision, then we have outputs, which are near-term results,” said Sanchez. “Then we have outcomes, which are the longer-term results. And over time, we lose control of those — from implementation to the results of those outputs and outcomes, we're losing more and more control over dictating how that happens.”

Sanchez offered the U.S. highway system as an example, where President Dwight D. Eisenhower decided to fund the Federal-Aid Highway Act to help mobilize troops across the country in the event of an invasion.

“The immediate result is we’re going to start building roads, and the outcome is we now have roads built from coast to coast,” said Sanchez. “Now, we were never invaded by another country. So in those terms, would we say this was a good decision? Well, we all agree it was a great decision … . The outcome was that the U.S. economy benefited enormously because of those decisions.”

Huang identified three common challenges that organizations face when preparing their data for analytics and AI: data being spread across systems, a lack of defined use cases and the ever-shifting knowledge barrier in the face of evolving technology.

On the subject of data spread, Huang pointed out that not all data is subject to the same regulations.

“Where you can have data in different formats, some are structured and from, say, an ERP system,” said Huang. “Some of the data are just managed in a shared folder and governed by the IM permissions. This data can be managed by different business teams and then it could be subject to different compliance and regulation needs, like for example, HIPAA compliance and FedRAMP."

To combat the knowledge barrier, Huang recommended that employers focus on upskilling and assured the audience that becoming a data scientist is not required or necessary to begin experimenting with AI.

Fultz Nordstrom clarified that expectations from constituents are still driving the need for digital transformation, albeit through different delivery methods today than 30 years ago. According to Fultz Nordstrom, identifying the needs and the data remain key first steps, along with refining approaches, validating assumptions and proposing solutions before implementation.

As for how to convince skeptics in an organization to warm up to AI, Fultz Nordstrom suggested looking outside one’s agency for additional validation examples to bring back and continue justifying one’s approach.

“Think about what are the ultimate delivery points,” said Fultz Nordstrom. “Think about who ultimately is going to benefit from this and then frame every conversation from there … . Keep going, keep showing up, keep showing them that this is the new path, and then look for peers outside of your agency to give you kind of that increased energy to keep going at it … that will give you more examples that you can then bring back and say, well, this agency or this state, here's some other examples. And it just continues to validate your message.”

*Government Technology and the Center for Digital Government are part of e.Republic, Industry Insider — Texas’ parent company.