Houston startup Saivory has helped brands like Shipley Do-Nuts and Fajita Pete's launch AI-powered ordering. Photo via Unsplash

It’s no secret that restaurant trends move fast and margins run thin. And with the proliferation of platforms like Uber Eats, DoorDash and Easy Cater, customer loyalty is fleeting.

The solution?

How about an AI-powered restaurant technology platform that helps restaurant brands cut back on third-party platforms in favor of driving direct discovery, conversion and loyalty?

Enter Saivory. Founded in 2025 by Stephen Klein, a software investor, and Fajita Pete’s restaurateur Hugh Guill, the Houston-based startup aims to help eateries better understand and activate guest behavior across digital channels as AI increasingly reshapes how consumers discover and engage with brands.

In less than a year, Saivory has partnered with Shipley Do-Nuts and Fajita Pete’s to bring AI-powered ordering to life.

“With Saivory, we were able to answer the question of, ‘what if the ordering process could be reduced to a single step, where customers simply tell us what they want and AI takes care of the rest?’” Klein tells InnovationMap.

The Houston-based startup made such an immediate impact that it was selected as a semi-finalist during Start-Up Alley at MURTEC, the restaurant industry’s leading technology conference, which took place last month in Las Vegas.

“Houston is a great hub for technology innovation, and we were proud to represent the city at MURTEC this year,” says Klein. “We didn’t win, but we were able to talk about some of the work that we have existing in the market for clients right now and a little bit about what we’re working on in the future.”

In the current restaurant technology ecosystem, the third-party aggregators own the customer attention that brings volume to restaurants, while also taking big commissions and having control over the end relationships with the customer.

That can often make it difficult for restaurants to grow loyalty and repeat business from customers. Saivory aims to level the playing field for restaurants, helping them stay more connected to their customers.

Take Saivory’s recent application with Shipley’s Do-Nuts, for example.

Saivory powered the donut giant’s AI-ordering and launched Shipley's website and mobile app to support its over 300 locations in Texas alone.

Shipley’s new AI-powered assistant helps users create personalized order recommendations based on individual or group preferences. And unlike standard chatbox features, the new assistant makes custom recommendations based on multiple customer factors, including budgetary habits, individual flavor preferences and order size. It can also be used for large catering orders.

“They're seeing more traffic to the site and they're seeing when customers use our AI-enabled flows,” Klein says. “And they're seeing higher basket sizes, bigger tickets, by about 25 percent.”

Klein says Saivory’s technology helps strengthen first-party digital relationships, reduce friction and cart abandonment, improve average order value, and delivers personalized, efficient experiences.

“It’s a win-win: the customer gets the right order quickly, while the restaurant gets a bigger margin,” he adds.

Additionally, the technology makes it easier for restaurants to share rewards, loyalty and discounts, ultimately growing more direct traffic and making restaurants less reliant on third-party delivery apps.

Next up for Saivory is adding new components to its platform to enhance the relationship between restaurant and customer, as well as technology around making it easier for restaurants to get found on Google.

“A lot of people are still searching for the best donuts near me,” Klein says. “Or what’s the best Mexican food near me? Customers will increasingly move to AI, where they’re going to ask where they should eat dinner and expect it to just order them dinner. They will eventually expect the technology to know how to do that. So that’s what we’re driving at.”

Shipley Donuts' new AI-powered assistant will learn your doughnut craving. Photo via Facebook.

Shipley Donuts launches AI-powered ordering assistant

fresh tech

Popular Houston-born doughnut chain Shipley Donuts has added a first-of-its-kind AI-powered assistant to its online ordering platform.

The new assistant can create personalized order recommendations based on individual or group preferences, according to a news release from the company. Unlike standard chatbox features, the new assistant makes custom recommendations based on multiple customer factors, including budgetary habits, individual flavor preferences and order size.

"We're not just adding AI for the sake of innovation — we're solving real customer pain points by making ordering more intuitive, personalized and efficient," Kerry Leo, Shipley Vice President of Technology, said in the release.

The system also works for larger events, as it can make individual orders and catering recommendations for corporate events and meetings by suggesting quantities and assortments based on group size, event type and budget.

According to Shipley, nearly 1 in 4 guests have completed orders with the new AI technology since it launched on its website.

“The integration of the AI ordering assistant into our refreshed website represents a significant leap forward in how restaurant brands can leverage technology to enhance the customer experience,” Leo added in the release.

Ad Placement 300x100
Ad Placement 300x600

CultureMap Emails are Awesome

How Houston innovators played a role in the historic Artemis II splashdown

safe landing

Research from Rice University played a critical role in the safe return of U.S. astronauts aboard NASA’s Artemis II mission this month.

Rice mechanical engineer Tayfun E. Tezduyar and longtime collaborator Kenji Takizawa developed a key computational parachute fluid-structure interaction (FSI) analysis system that proved vital in NASA’s Orion capsule’s descent into the Pacific Ocean. The FSI system, originally developed in 2013 alongside NASA Johnson Space Center, was critical in Orion’s three-parachute design, which slowed the capsule as it returned to Earth, according to Rice.

The model helped ensure that the parachute design was large enough to slow the capsule for a safe landing while also being stable enough to prevent the capsule from oscillating as it descended.

“You cannot separate the aerodynamics from the structural dynamics,” Tezduyar said in a news release. “They influence each other continuously and even more so for large spacecraft parachutes, so the analysis must capture that interaction in a robustly coupled way.”

The end result was a final parachute system, refined through NASA drop tests and Rice’s computational FSI analysis, that eliminated fluctuations and produced a stable descent profile.

Apart from the dynamic challenges in design, modeling Orion’s parachutes also required solving complex equations that considered airflow and fabric deformation and accounted for features like ringsail canopy construction and aerodynamic interactions among multiple parachutes in a cluster.

“Essentially, my entire group was dedicated to that work, because I considered it a national priority,” Tezduyar added in the release. “Kenji and I were personally involved in every computer simulation. Some of the best graduate students and research associates I met in my career worked on the project, creating unique, first-of-its-kind parachute computer simulations, one after the other.”

Current Intuitive Machines engineer Mario Romero also worked on Orion during his time at NASA. From 2018 to 2021, Romero was a member of the Orion Crew Capsule Recovery Team, which focused on creating likely scenarios that crewmembers could encounter in Orion.

The team trained in NASA’s 6.2-million-gallon pool, using wave machines to replicate a range of sea conditions. They also simulated worst-case scenarios by cutting the lights, blasting high-powered fans and tipping a mock capsule to mimic distress situations. In some drills, mock crew members were treated as “injured,” requiring the team to practice safe, controlled egress procedures.

“It’s hard to find the appropriate descriptors that can fully encapsulate the feeling of getting to witness all the work we, and everyone else, did being put into action,” Romero tells InnovationMap. “I loved seeing the reactions of everyone, but especially of the Houston communities—that brought me a real sense of gratitude and joy.”

Intuitive Machines was also selected to support the Artemis II mission using its Space Data Network and ground station infrastructure. The company monitored radio signals sent from the Orion spacecraft and used Doppler measurements to help determine the spacecraft's precise position and speed.

Tim Crain, Chief Technology Officer at Intuitive Machines, wrote about the experience last week.

"I specialized in orbital mechanics and deep space navigation in graduate school,” Crain shared. “But seeing the theory behind tracking spacecraft come to life as they thread through planetary gravity fields on ultra-precise trajectories still seems like magic."

UH breakthrough moves superconductivity closer to real-world use

Energy Breakthrough

University of Houston researchers have set a new benchmark in the field of superconductivity.

Researchers from the UH physics department and the Texas Center for Superconductivity (TcSUH) have broken the transition temperature record for superconductivity at ambient pressure. The accomplishment could lead to more efficient ways to generate, transmit and store energy, which researchers believe could improve power grids, medical technologies and energy systems by enabling electricity to flow without resistance, according to a release from UH.

To break the record, UH researchers achieved a transition temperature 151 Kelvin, which is the highest ever recorded at ambient pressure since the discovery of superconductivity in 1911.

The transition temperature represents the point just before a material becomes superconducting, where electricity can flow through it without resistance. Scientists have been working for decades to push transition temperature closer to room temperature, which would make superconducting technologies more practical and affordable.

Currently, most superconductors must be cooled to extremely low temperatures, making them more expensive and difficult to operate.

UH physicists Ching-Wu Chu and Liangzi Deng published the research in the Proceedings of the National Academy of Sciences earlier this month. It was funded by Intellectual Ventures and the state of Texas via TcSUH and other foundations. Chu, founding director and chief scientist at TcSUH, previously made the breakthrough discovery that the material YBCO reaches superconductivity at minus 93 K in 1987. This helped begin a global competition to develop high-temperature superconductors.

“Transmitting electricity in the grid loses about 8% of the electricity,” Chu, who’s also a professor of physics at UH and the paper’s senior author, said in a news release. “If we conserve that energy, that’s billions of dollars of savings and it also saves us lots of effort and reduces environmental impacts.”

Chu and his team used a technique known as pressure quenching, which has been adapted from techniques used to create diamonds. With pressure quenching, researchers first apply intense pressure to the material to enhance its superconducting properties and raise its transition temperature.

Next, researchers are targeting ambient-pressure, room-temperature superconductivity of around 300 K. In a companion PNAS paper, Chu and Deng point to pressure quenching as a promising approach to help bridge the gap between current results and that goal.

“Room-temperature superconductivity has been seen as a ‘holy grail’ by scientists for over a century,” Rohit Prasankumar, director of superconductivity research at Intellectual Ventures, said in the release. “The UH team’s result shows that this goal is closer than ever before. However, the distance between the new record set in this study and room temperature is still about 140 C. Closing this gap will require concerted, intentional efforts by the broader scientific community, including materials scientists, chemists, and engineers, as well as physicists.”

---

This article originally appeared on EnergyCapitalHTX.com.