# Texas Capital Bancshares: Responsible AI

Source: https://www.bankingnewsai.com/banks/texas-capital/responsible-ai
Last updated: Nov 17, 2026

*Part of the [Texas Capital Bancshares AI strategy profile](https://www.bankingnewsai.com/banks/texas-capital).*

## How does Texas Capital govern its use of AI?

Texas Capital governs AI through its board and a management program rather than published principles. The board's Technology Committee, formed on January 22, 2025, carries risk oversight of cybersecurity and AI, according to the 2026 proxy, which also says the firm implemented a comprehensive AI governance program with a dedicated working group in 2025 and engaged external firms to conduct AI defense assessments. The annual report first named AI as a risk in the 10-K for 2024, filed February 11, 2025; the 10-K for 2023 does not mention artificial intelligence. That risk factor warns that generative AI models may produce output that is incorrect or biased, that limited transparency makes models hard to assess, and that the company may rely on AI models developed by third parties; the 2025 10-K repeats it, and the Q3 2026 10-Q lists AI among forward-looking risks. CEO Rob Holmes said in April 2026 that governance and risk management are embedded in every stage of development and deployment. An October 2026 AWS blog shows the bank's thinking for agents that act: autonomy granted by a nightly trust score from 0 to 100 and revoked, with rollback, when it falls below a floor of 80. Texas Capital has published no AI principles document. ([SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm), [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742825000036/tcbi-20241231.htm), [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742824000040/tcbi-20231231.htm), [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000010/tcbi-20251231.htm), [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000080/tcbi-20260930.htm), [The Motley Fool](https://fool.com/earnings/call-transcripts/2026/04/23/texas-capital-tcbi-q1-2026-earnings-transcript), [Amazon Web Services (AWS for Industries blog) (archived copy of 2026-10-06; original page offline as of 2026-11-17)](https://web.archive.org/web/20261006020141/https://aws.amazon.com/blogs/industries/trust-earned-autonomy-how-texas-capital-bank-demonstrates-an-ai-agent-that-commits-to-the-core-banking-ledger-then-reverses-itself-2/))

## What governance of AI has Texas Capital disclosed, and when?

In date order. Filings and statements as worded by the source.

| Item | What it says | Source | Date | Sources |
| --- | --- | --- | --- | --- |
| 10-K for 2023 | No mention of artificial intelligence | Annual report | February 13, 2024 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742824000040/tcbi-20231231.htm) |
| Technology Committee | Board committee formed on January 22, 2025 | 2026 proxy | January 22, 2025 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm) |
| 10-K for 2024 | First AI risk factor: 'The development and use of AI presents risks and challenges that may adversely impact the Company's business' | Annual report | February 11, 2025 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742825000036/tcbi-20241231.htm) |
| 10-K for 2025 | Repeats the AI risk factor | Annual report | February 10, 2026 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000010/tcbi-20251231.htm) |
| 2026 proxy | 'Risk oversight by the Board and Committees, including cybersecurity and AI by the Technology Committee' | Proxy statement | March 12, 2026 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm) |
| 2026 proxy | The firm implemented a comprehensive AI governance program and established a dedicated working group; external firms conducted AI defense assessments | Proxy statement | March 12, 2026 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm) |
| Rob Holmes, CEO | 'We have embedded governance and risk management into every stage of development and deployment' | Q1 2026 earnings call | April 23, 2026 | [The Motley Fool](https://fool.com/earnings/call-transcripts/2026/04/23/texas-capital-tcbi-q1-2026-earnings-transcript) |
| 10-Q for Q3 2026 | Lists 'risks related to the development and use of artificial intelligence' among forward-looking factors | Quarterly report | October 22, 2026 | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000080/tcbi-20260930.htm) |

## What AI risks does Texas Capital's 10-K describe?

The risk factor, first in the 10-K for 2024, says the legal and regulatory environment relating to AI is 'uncertain and rapidly evolving' and could require changes in the company's implementation of AI technology and increase its compliance costs. It says AI models, 'particularly generative AI models, may produce output or take action that is incorrect', reflect biases in training data, release confidential information or infringe intellectual property. ([SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742825000036/tcbi-20241231.htm))

It adds that the complexity of many AI models 'makes it difficult to understand why they are generating particular outputs', which makes it harder to reduce erroneous output, eliminate bias and comply with rules requiring explanation of decisions, and that the company may rely on AI models developed by third parties, over whose training it may have limited visibility. ([SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742825000036/tcbi-20241231.htm))

## How does Texas Capital propose to control AI agents that write to systems of record?

The AWS blog of October 2026, a demonstration on synthetic data, gives each decision recipe a trust score from 0 to 100, recomputed nightly, with reviewer agreement weighted at 40 points. A recipe moves from Learning to Competent at a score of at least 70 and 50 production runs, and to Autonomous at 90 and 200 runs; it drops to Competent below the autonomous floor of 80 and straight to Learning on a regression of 15 points or more. On demotion the system reverses the commits it made under the revoked tier. ([Amazon Web Services (AWS for Industries blog) (archived copy of 2026-10-06; original page offline as of 2026-11-17)](https://web.archive.org/web/20261006020141/https://aws.amazon.com/blogs/industries/trust-earned-autonomy-how-texas-capital-bank-demonstrates-an-ai-agent-that-commits-to-the-core-banking-ledger-then-reverses-itself-2/))

The agent can call only a closed set of ten tools, and a Cedar policy at the gateway caps loan mutations at an operator-set amount regardless of what a prompt asks. 'Trust is slow to grant and fast to revoke,' Vibhu Bhatnagar is quoted as saying. The post reports banking-verdict accuracy of 50% (3/6) on its benchmark, which it says is why such a recipe would not reach the autonomous tier. ([Amazon Web Services (AWS for Industries blog) (archived copy of 2026-10-06; original page offline as of 2026-11-17)](https://web.archive.org/web/20261006020141/https://aws.amazon.com/blogs/industries/trust-earned-autonomy-how-texas-capital-bank-demonstrates-an-ai-agent-that-commits-to-the-core-banking-ledger-then-reverses-itself-2/))

Not on the record: an AI principles document, a named head of model risk, or a chief AI officer. Risk leadership changed in September 2026, when Chief Risk Officer David Oman left and Chief Credit Officer David Youngberg took on the role while a successor is sought. ([SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000063/tcbi-20260916.htm))

## What does the Texas Capital profile record on this topic?

| Date | Move | Detail | Sources |
| --- | --- | --- | --- |
| Sep 16, 2026 | Chief risk officer departs; chief credit officer steps in | An 8-K says David Oman left as CRO effective immediately, stating the departure was unrelated to any dispute or to financial condition or controls. David Youngberg, Chief Credit Officer, took on CRO duties and is a candidate in the search for a successor. | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000063/tcbi-20260916.htm) |
| Jul 23, 2026 | Q2 10-Q mentions AI only as a forward-looking risk | The quarterly report lists 'risks related to the development and use of artificial intelligence' among its forward-looking factors, its only AI reference; the Q1 10-Q is the same. | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000059/tcbi-20260630.htm) |
| Mar 12, 2026 | Proxy: AI governance programme and staff AI initiative | The 2026 proxy says the bank implemented an AI governance programme with a dedicated working group, engaged external firms for AI defence assessments, and launched a firm-wide staff initiative in four areas: AI Talk, AI Learn, AI Engage and AI Grow. It places oversight of cybersecurity and AI with the Technology Committee. | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm) |
| Feb 10, 2026 | 2025 10-K repeats its AI risk factor | Carried over from the 2024 report: generative-AI output errors, bias, limited model transparency, evolving AI law and dependence on third-party models; a board Technology Committee formed in Q1 2025 oversees IT and cyber risk. | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000010/tcbi-20251231.htm) |
| Jan 22, 2026 | Q4 2025 call: AI to accelerate productivity | The CFO says technology has improved expense-base productivity for several years and expects that to accelerate as the bank 'more fully adopt[s] AI across the franchise'. | [The Motley Fool (Yahoo Finance)](https://finance.yahoo.com/news/texas-capital-q4-2025-earnings-153932984.html) |
| Jan 22, 2025 | Board Technology Committee formed | The 2026 proxy says the committee, formed on 22 January 2025, oversees information technology, cybersecurity risk and significant technology investments; it held four meetings in 2025. | [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1077428/000107742826000040/tcbi-20260309.htm) |
| May 21, 2024 | David Oman named chief risk officer | Oman, from PwC's financial risk practice, joined on 10 June 2024 as CRO reporting to the CEO and sitting on the Operating Council, responsible for risk functions, enterprise risk management and regulatory compliance. | [Texas Capital](https://texascapitalbank.com/who-we-are/newsroom/news/2024/05/21/Texas-Capital-Names-David-Oman-as-Chief-Risk-Officer) |

## Common questions

### Who oversees AI risk at Texas Capital?

The board's Technology Committee, formed on January 22, 2025, oversees cybersecurity and AI risk, according to the 2026 proxy. In management, the proxy says the firm implemented a comprehensive AI governance program with a dedicated working group in 2025. Chief digital and information officer Mo Jamous, appointed July 13, 2026, has AI enablement and information security in his remit.

### When did Texas Capital's 10-K start mentioning AI?

In the 10-K for 2024, filed February 11, 2025, which added a risk factor titled 'The development and use of AI presents risks and challenges that may adversely impact the Company's business.' The 10-K for 2023 does not mention artificial intelligence. The 2025 10-K repeats the risk factor, and the Q3 2026 10-Q lists AI among forward-looking risks.

### Does Texas Capital have published AI principles?

No. No AI principles or responsible-AI policy document is public. The disclosed governance is the board Technology Committee's oversight, a management AI governance program and working group, external AI defense assessments, and the 10-K risk factor.

### What is Trust-Earned Autonomy?

A governance pattern Texas Capital Bank demonstrated with AWS, described in an October 2026 AWS blog: AI decision recipes earn the right to commit to a ledger through a nightly trust score from 0 to 100, lose it below a floor of 80, and have their commits reversed on demotion. It runs on synthetic data and is a demonstration, not a production workload.
## Deeper on Texas Capital

- [What is Texas Capital's AI strategy?](https://www.bankingnewsai.com/banks/texas-capital/ai-strategy)
- [What technology stack does Texas Capital run?](https://www.bankingnewsai.com/banks/texas-capital/technology-stack)


---

Canonical page: https://www.bankingnewsai.com/banks/texas-capital/responsible-ai
Part of [BankingNewsAI](https://www.bankingnewsai.com/) — a free daily brief on AI in banking, an AI regulation tracker (46 authorities, 278 documents) and AI-strategy profiles of the 120 largest US banks. Markdown versions of every reference page: append `.md` to the page URL; index at https://www.bankingnewsai.com/llms.txt.
