Our verdict
KUNGFU.AI is an AI development, implementation and consulting firm in Austin, Texas, founded in 2017 by Stephen Straus, Steve Meier, Ron Green and Zhangzhang Si. It pairs AI strategy work with custom machine learning and generative AI builds, mainly for healthcare, government and financial services clients. Its site says the team has more than 40 people, and Clutch lists 10 to 49 employees. In its entry for the 2026 HFS Research awards, the firm reported $18.4 million in revenue for the fiscal year that ended March 30, 2026, up from $9.8 million the year before (KUNGFU.AI; Clutch; HFS Research).
Its best-documented system is Instant Decision, an invoice approval model built for the factoring business of Triumph Financial. KUNGFU.AI says the model cut approval time from 178 minutes to about 10 seconds, has processed about $4 billion of invoices and approves more than half of them on its own, while people still review the cases the model is unsure about. The firm also trained the breast cancer risk model behind Clairity's Allix5, which the FDA authorized through its De Novo process, and built a radar-based debris detection system for the F-35 Joint Program Office (Triumph case study; FDA De Novo DEN240047; F-35 case study). In September 2026 HFS Research gave it first place for Best Startup / Emerging Tech Firm in its Services-as-Software Awards (HFS Research).
KUNGFU.AI is worth a call for a healthcare, finance or government organization that wants one high-stakes model built and put into daily use by a small team that handles both strategy and engineering. Buyers who need a large team or a long record of public client reviews will be better served elsewhere.
The right buyer
- Healthcare and life sciences companies that need a model taken through clinical validation and FDA review, as with Clairity's Allix5.
- Finance and operations teams that want one high-volume decision automated, with people reviewing the cases the model is unsure about, as Triumph did for invoice approvals.
- Defense and government programs: the firm has built debris detection for the F-35 Joint Program Office and an AI playbook for the Defense Department's Chief Digital and AI Office.
- Leaders who want the AI roadmap and the build from one team, as Iodine Software got with its use-case ranking and product work.
Reasons to look elsewhere
- Clutch shows no client reviews, so nearly all published praise comes from the firm's own case studies.
- Clutch lists a $100,000 minimum project size, which rules out small pilots.
- With 10 to 49 people on Clutch and more than 40 on its own site, it has a small team for large programs that run several workstreams at once.
- Its published Triumph figures differ by source: the case study says more than half of about $4 billion in invoices are approved automatically, while its HFS entry says 70% of more than $18 billion a year.
Scores
| Criterion | Weight | Score | Basis |
|---|---|---|---|
| Production track record | 20% | 7.8 | Triumph Instant Decision rolled out to all carrier customers in June 2025; Clairity Allix5 authorized by the FDA; F-35 debris detection being deployed to more airfields. |
| Technical depth | 15% | 7.6 | Radar computer vision with synthetic data and on-device deployment; gradient boosted underwriting model; LLM roadmap work for Iodine Software. |
| Industry depth | 10% | 7.4 | Clutch lists 45% medical and 25% government work; its government solutions team signed its first prime contract in 2023. |
| Delivery model and handoff | 15% | 7.2 | Strategy and engineering on one team, with a staged feasibility check before a full build; Iodine praised the handoff; no third-party delivery reviews. |
| Commercial clarity | 10% | 6.0 | Clutch lists a $100,000 minimum; no rates or pricing models are published. |
| Platform independence | 10% | 7.5 | Builds custom models on each client's own data and systems, such as Triumph's invoice history and Iodine's existing software; Google Cloud machine learning specialization since 2021. |
| Stability and scale | 10% | 7.0 | Operating since 2017; $18.4 million revenue in fiscal 2025/26 per its HFS entry; more than 40 staff per its site. |
| Independent evidence | 10% | 7.0 | HFS Research first place in 2026; FDA De Novo record for Allix5; Inc. 5000 listing; no Clutch reviews. |
| Overall | 100% | 7.3 | Weighted average |
Where KUNGFU.AI fits
| Need | Fit | Why |
|---|---|---|
| Regulated healthcare model taken through clinical validation | Strong | Clairity Allix5 De Novo authorization; Clutch lists 45% medical work. |
| One high-volume decision automated with human review of unsure cases | Strong | Triumph Instant Decision covers all carrier customers. |
| AI strategy and roadmap for leadership or an agency | Good | Defense Department CDAO AI playbook; Iodine use-case ranking and roadmap. |
| Small pilot under $100,000 | Weak | Clutch lists a $100,000 minimum project size. |
What clients and leaders say
“From day one, they brought incredible talent and were people I could jam with, think strategically, and genuinely solve problems alongside.”
“The KUNGFU.AI team was very knowledgeable and attentive, I really enjoyed working with them. The integration into our current infrastructure was a breeze, as was the handoff from KUNGFU.AI in general.”
“Prior to ChatGPT, it was very hard to find other services firms focusing on AI, and now it’s almost impossible to find a services firm that doesn’t say they offer AI services.”
Sources (13)
- KUNGFU.AI homepage
- KUNGFU.AI Our Story
- KUNGFU.AI on Clutch
- HFS Research Services-as-Software Awards 2026 entry: KUNGFU.AI
- Triumph Instant Decision case study
- Triumph appoints Jason Heilig chief technology officer of its factoring division
- Clairity breast cancer risk prediction case study
- FDA De Novo decision summary for Allix5 (DEN240047)
- F-35 Joint Program Office debris detection case study
- Defense Department CDAO AI playbook case study
- Iodine Software case study
- Unite.AI interview with Stephen Straus
- KUNGFU.AI appoints a chief marketing officer and chief revenue officer (GlobeNewswire)