Nan Feng spent two years of late nights, Python coding classes, and more than 5,000 lines of code on an AI journey that probably sounds familiar to many in hospitality. The SVP of Finance and Accounting at Avion Hospitality was building a corporate forecasting and budgeting model to replace the spreadsheets he’d been working in for more than two decades. 

The further he pushed, the more potential he saw for what AI could do. But he kept running into the same problem. The data he needed was spread across all the brands, systems, and more than 40 properties Avion Hospitality managed throughout the U.S. 

The Plano, TX-based hotel management company was early to embrace AI and Feng was leading the way.

Can I make the process more automated? How do I improve the entire team’s efficiency, and ultimately, how do we serve owners and deliver customer value? That’s where I started. 

Nan Feng, SVP of Finance and Accounting at Avion Hospitality.

Hitting the data wall

Feng tested, validated, and rewrote code with his AI tools. When his models grew too large to handle in a single conversation, he started breaking them down into smaller sections he could validate along the way. The initial models were producing results he could use, but the further he went, the bigger the stakes and the harder the data challenge became.

His audience was CEOs, owners, and the senior leadership team at Avion. “Accuracy was absolutely critical,” Feng said, “because without accurate data, people start not trusting what you’re producing and you lose all credibility.”

But data fidelity in the hotel industry has long been a challenge. Hundreds of brands and technology systems live in a spider web of integrations and data flows that never quite behave the way they’re supposed to.  

Feng knew one place where he did trust the data. Avion had been using ProfitSword, Actabl’s business intelligence platform, for years.

“I love the fact that Actabl can take in all the data from different sources, from brands, the PMS, sales, revenue management, financials, and aggregate it into one box for the operators as well as the corporate teams to use,” he said.

Building a strong foundation

Feng knew Avion needed its own data warehouse, and who to ask for advice.

We didn’t have the internal resources or capability to do it ourselves, Actabl was the first company we reached out to, and at first, it was, ‘Hey, do you have anybody you can recommend.’ And then it soon turned to, ‘Why don’t we just partner on this ourselves? 

Nan Feng, SVP of Finance and Accounting at Avion Hospitality.

Feng’s logic was simple: “You've built your own data warehouse. You have essentially all of our data. So it was a natural partnership.”

The timing couldn’t have been better. In May 2026, Actabl became the first hotel technology company granted a patent for its proprietary approach to hotel data normalization. And the company had just launched a forward-deployed engineering program. The model, first started by Palantir and now adopted by OpenAI, Anthropic and others, embeds an engineer directly inside a company to build custom data, analytics, and AI solutions.

Avion became a founding partner, deploying Actabl’s full Data Services offering. The ProfitSword Managed Data Feed gave the team a clean, normalized stream of data flowing across their portfolio, and a Managed Data Warehouse on their own Snowflake account put that data where Feng could build from. 

Actabl engineer Marc Burton embedded with the Avion finance team and met weekly with Feng and his colleagues to build the infrastructure together. Actabl’s Mike Fatal, a senior data engineer and one of the named inventors on the data patent, ran training with the Avion team.

Together, they rebuilt the reporting infrastructure from the ground up using Power BI dashboards that replaced the manual processes and Excel reports the team had been running for years.

Unlocking what AI can really do

The difference was immediate. Data that once required dedicated scripts and manual exports now flowed directly into an environment Feng could easily query and slice by property, ownership group, or revenue stream. And it fed the AI models he’d been building with clean, governed data reconciled within ProfitSword.

“The data was there before, but at the corporate level it was still a little hard to get to,” Feng said. “Now, with the data warehouse, we can go in, query the information, run different regression analyses, so we can show operators in a meaningful way whether X and Y correlate. Then we can go deeper into where we can make things operationally more efficient, so everything runs on the same connection instead of being fragmented.”

Just ahead of the 2027 budget season, the partnership and data warehouse are already delivering owner reports and USALI (Uniform System of Accounts for the Lodging Industry) analyses previously built by hand.

Across the Avion finance and accounting team, the data foundation is powering treasury agents, bank reconciliation automation, and tools that help staff accountants create journal entries and reconcile balance sheets faster. They’re also getting forecasts to senior leadership on a weekly and even daily basis now. 

The entire team has built out different tools to support different functions, I have a great team that’s hungry for innovation and always looking for something new. But the gatekeeper is human expertise, which is always involved in the middle of every build. We’re not just putting data into (LLMs) and letting it run without testing. There’s always a human element and the experience we all bring to the table as a check and balance.

Nan Feng, SVP of Finance and Accounting at Avion Hospitality. 

The future is now

Feng’s path mirrors what many hospitality teams are now discovering. AI opens a door and ambitious companies walk through it. But the further you go, the more you realize the data underneath is what ultimately determines how far you can go. 

Avion figured that out early and built the foundation before trying to build the future. Feng plans to build an Avion-specific MCP next year, a layer on top of the data warehouse that puts trusted context and answers at the fingertips of anyone in the organization, from corporate finance to regional leaders to on-property staff. 

The next phase is tying all these different tools together. Can we build agents that take all the relevant information together, so not only can the finance and accounting team close month-end faster, but the corporate team can make decisions a lot faster too because the data will be right there?

Nan Feng, SVP of Finance and Accounting at Avion Hospitality.

For Avion, that question already has an answer. Now that the foundation is built, the sky is the limit.