Ramon Sanchez
CEO, FounderFounder of Dreszi. Over twenty years across marketing, search, product, and education. Sets the lab's direction and runs applied AI strategy on every build.
Vesti Labs is an applied AI lab in Austin. We design, engineer, and launch AI apps, agents, and custom models. Every method we sell was used first on a product we own.
Most companies hire a strategy firm, then an agency, then a dev shop. Each one hands the next a document, and the context stays behind. An AI product is mostly context: the data, the edge cases, and the reason a model should answer one way and not another.
| What the market found in 2025 | Share |
|---|---|
| Companies that saw no measurable return from generative AI | 95% |
| Custom AI tools built with an outside partner that reached deployment | 67% |
| Custom AI tools built in house that reached deployment | 33% |
MIT Project NANDA, The GenAI Divide: State of AI in Business 2025. Deployment shares are approximate and self-reported, and the authors say they do not show that partnering causes success.
We run three products of our own. Each one tested part of the method before any client paid for it.
| Part | Product | Platform | Status | What it does | |
|---|---|---|---|---|---|
| VL-01 | Interface Audit | Web | Live | Audits a website against 168 peer-reviewed UX principles in about 30 seconds. | interfaceaudit.com |
| VL-02 | Dreszi | iOS | In launch | Wardrobe intelligence. One outfit each morning, already right. | dresz.io |
| VL-03 | Pulso | Partner app | Closed | Signal detection for regulated prediction markets. Its Bayesian thesis became Chevalière. | Case notes coming |
UX insights agencies charge thousands for. 2,098 citations. Zero opinions. Paste a URL and the audit returns findings with the research behind each one.
Client work runs on the same tools as our own products. Every line below has already shipped once.
| Capability | What you get | Shipped first in |
|---|---|---|
| AI apps | Web and iOS products with a model at the core, from prototype to store release. | Dreszi |
| Agents and internal tools | Agents connected to your systems, data, and approvals, with an off switch your team controls. | Pulso |
| Custom models | Models trained on your brand or domain, plus the tests that show they hold up. | Chevalière |
| Product and interface design | Interfaces designed from research and checked against 168 principles before launch. | Interface Audit |
| Applied research | Synthetic panels that test a product decision before you pay to build it. | Maison |
| Launch | Site, search, paid, and AI-answer channels that bring the first users. | All three |
The signet of the lab. A proprietary Bayesian engine, forged in Pulso, that scores every output against the bar set at the start of the project.
Nothing leaves unsealed.Synthetic panels built for one client, one market, one question. They read a product decision before real users have to.
Sharper. More exact. Between them, the founders have shaped work for Lucky Strike, Oreo, Nestlé, Visa, and DeWalt.
Founder of Dreszi. Over twenty years across marketing, search, product, and education. Sets the lab's direction and runs applied AI strategy on every build.
Two decades designing interfaces. Author of the 168 principles behind Interface Audit. Leads experience and product design across Atelier and Maison.
Send the product, the data, and the question. We answer with what we would do first.
The diagnosis and the build plan. You can run it with us or with anyone else.
Capsule for a two-week sprint, Patronage for a standing team, Confection for the full build.
Write three lines: what you want to build, who it is for, and what you have today. We reply within 48 hours, in English or Spanish.
+1 (512) 222-3078
AI apps for web and iOS, agents and internal tools connected to your systems, custom models with their tests, and the design and launch work around them.
A lab ships its own products. We run Interface Audit, Dreszi, and Pulso, so the method you hire was used first on products we own, by the people who will work on yours.
With a first read at no cost. If there is something worth building, the next step is an executive brief with a fixed scope and a fixed price.
Yes. The lab is in Austin and works with teams across the United States and Latin America, in English and Spanish.