IE11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

What to Know When Selling AI to the Public Sector in Texas

What to Know:
  • A recent Texas Digital Government Summit panel brought public- and private-sector IT leaders together to discuss practical AI.
  • Agencies are evaluating AI investments not on novelty but on measurable impacts.
  • Panelists stressed the importance of understanding agency missions, building trust and offering solutions that agencies can scale across multiple business units.

Texas Department of Public Safety CIO Jessica Ballew speaking into a microphone while seated at a table beside NVIDIA Senior Business Manager Michael Sherwood and IBM Worldwide Leader for Data and AI Solutions Bala Vaithyalingam at the Texas Digital Government Summit in Austin on July 24, 2025.
From left to right: Texas Department of Public Safety CIO Jessica Ballew, NVIDIA NVIDIA State and Local Government Senior Business Manager Michael Sherwood and Worldwide Leader for Data and AI Solutions Bala Vaithyalingam at the Texas Digital Government Summit in Austin on July 24, 2025.
Photo by Chandler Treon.
As Texas public-sector agencies deepen their use of artificial intelligence, vendors looking to support them need to go beyond feature pitches and focus on real-world outcomes.

That was a central theme at a recent panel discussion held during the Texas Digital Government Summit* in Austin, featuring Texas Department of Public Safety (DPS) CIO Jessica Ballew, NVIDIA State and Local Government Senior Business Manager Michael Sherwood and IBM Worldwide Leader for Data and AI Solutions Bala Vaithyalingam.

The panel focused on how agencies are operationalizing AI and what it takes to move from pilot to enterprise-scale implementations.

“You don’t just do AI for AI’s sake,” said Ballew, echoing statements she made during an Industry Insider — Texas member briefing the night before. “Our goal isn’t to implement AI. Our goal is to solve a business problem or improve a business process.”

That perspective carries implications for how vendors engage. Agencies are evaluating AI investments not on novelty but on measurable impacts, whether that’s increasing call center capacity, improving data classification or accelerating internal workflows. DPS, for example, is using AI for image analysis in human trafficking investigations, license pattern recognition, and chatbots for driver license services.

“We can make contact and serve 30 percent more customers every day that are trying to get ahold of us that today can’t get through,” Ballew said, referencing DPS’ new AI-powered call center assistant.

Sherwood noted that across the country, many AI pilots never graduate to production because agencies and vendors fail to align on outcomes and resourcing.

“Don’t implement technology for technology,” Sherwood said. “Find real pain points with your customers, and work on those as starting points. And then implement solutions.”

For vendors, this means showing up with use case alignment already in mind. Panelists stressed the importance of understanding agency missions, building trust and offering solutions that agencies can scale across multiple business units.

“It has to be tied to your strategic initiatives,” added Vaithyalingam. “Don’t chase the hype. Focus on the real problem, the pain points.”

Agencies are also setting clearer expectations around AI governance. DPS requires vendors to submit specific documentation for any product with an AI component, including inventories and high-level system information. An internal AI governance board reviews those submissions in the same way it would evaluate technology against accessibility or cybersecurity standards.

Panelists emphasized that governance is what makes long-term success possible rather than a tedious hurdle.

“The purpose of the brake is not just to slow things down,” Vaithyalingam said, comparing governance to a race car. “It also provides a confidence to drive faster.”

Vendors should expect to address how their tools handle data privacy, human oversight and automated decision-making. In emerging models such as agentic AI, this also means defining which software agents can access which tools and under what circumstances.

And while the technical capabilities matter, so does the organizational readiness to support them. Ballew said DPS is facilitating cross-agency use case workshops to align efforts and reduce duplicative spending. AI platforms that can serve multiple departments — for example, in licensing, recruitment or IT — are more likely to be adopted.

“We try to bring in those groups where we think that they have a common capability that they could take advantage of,” Ballew said.

Sherwood advised vendors to prepare for lengthy conversations around data quality, contract governance and interdepartmental workflows. The agencies most likely to succeed with AI, he said, are the ones that know their data before starting.

Ultimately, the panel called on vendors to act as partners in helping agencies deliver ethical, measurable and efficient AI-driven services.

“Know your data, know your customer,” Sherwood said. “And don’t go it alone.”

*Note: The Texas Digital Government Summit is hosted by Government TechnologyGovernment Technology and Industry Insider — Texas are both part of e.Republic.
Chandler Treon is an Austin-based staff writer. He has a bachelor’s degree in English, a master’s degree in literature and a master’s degree in technical communication, all from Texas State University.