AI needs better plumbing
By Vishnu Rageev R.
Why connected systems decide the payoff
Nearly every hotel owner in a recent industry survey is experimenting with artificial intelligence. Far fewer can show what it has earned them. Wyndham Hotels & Resorts’ second annual Owner Trends Report, released in January, found that 98 percent of the 125 owners and developers surveyed had begun incorporating AI into their businesses, but only 12 percent said it was embedded across most of their operations.
Industry research and technology executives point to a common explanation. The limiting factor is less the AI than what sits underneath it: the accounting, purchasing, labor, room-control and network systems that make up a hotel’s back office. When those systems do not share accurate data, AI tools have little reliable information to work with.
That makes back-office technology a more consequential decision than it may appear, and one where cost, brand requirements and contract terms deserve as much scrutiny as features.
The integration gap
The 2026 Hotel Operations Index, released in January by hotel data platform Otelier and produced with Agilysys and Sage, found that only 11 percent of respondents had a fully integrated technology stack. About 91 percent still relied on some manual reporting and only a small share expressed high confidence in their data.
In practice, that can mean night auditors re-keying figures, managers exporting spreadsheets and finance teams reconciling numbers that disagree from one system to the next. An AI tool layered on that environment can produce faster answers built on flawed information.
Technology executives returned to this point repeatedly. Sunny Kurani, enterprise account manager at Vingcard, said AI “needs reliable information that reflects actual operating conditions.”
“The value will depend on the quality and relevance of the underlying data, along with clear controls and human judgment,” said Lee Bridges, chief product officer at back-office software company Inn-Flow. Frank Pitsikalis, senior vice president of product strategy and chief marketing officer at hospitality software provider Agilysys, described the problem from the employee’s side.
“Employees need accurate information without repeatedly entering data or switching between systems,” he said.
Before evaluating any AI feature, owners can map which systems hold their financial and operational data, which exchange information automatically and where staff still move numbers by hand. Those manual handoffs are often where errors and labor hours accumulate.
“The value will depend on the quality and relevance of the underlying data, along with clear controls and human judgment.”
— Lee Bridges, chief product officer at Inn-Flow
Before the spend
Most hotel financial reporting describes what has already happened. A newer generation of back-office tools aims to put information in front of managers while they can still change the outcome.
“Financial reporting remains essential, but operators also need information while they can still adjust a schedule, reconsider a purchase or respond to a change in demand,” Bridges of Inn-Flow said.
Inn-Flow combines accounting, labor management, payroll, procurement and reporting. The company said its labor tools forecast staffing needs from occupancy, demand and historical performance, allowing managers to adjust schedules as conditions change. In purchasing, its connected accounting and procurement tools display budgets, forecasts, actual spending and outstanding purchase-order commitments as staff build an order.

Lee Bridges, chief product officer at back-office software company Inn-Flow
Inn-Flow also said its newly launched intelligence tool, powered by Inn-Flow Accounting, simplifies financial reporting and lets users explore results interactively, giving owners and managers more time to investigate variances rather than assemble reports. Its AI Cart Optimizer, introduced at the HITEC conference in June after Inn-Flow acquired procurement software provider Lilo, compares prices across vendors and recommends lower-cost alternatives.
Bridges said buyers retain full control over final decisions.

Frank Pitsikalis, Agilysys senior vice president and chief marketing officer
Agilysys applies a similar idea to operational reporting. The company said its Agilysys Analyze platform combines operational data with AI-powered forecasting, anomaly detection and natural-language queries.
“Managers can ask questions in plain language and quickly identify trends that inform staffing, pricing and inventory decisions without relying on manually assembled reports,” Pitsikalis said.
These recommendations are only as reliable as the forecasts and data behind them. Owners should ask vendors how forecasts are generated, how accuracy is measured and how easily a manager can override a suggestion.
Procurement test
Purchasing and accounts payable have become an active area for back-office AI and one where owners can compare competing approaches.
Folio, a Durham, North Carolina-based company that describes itself as AI-native, said its Folio Bills product uses computer vision to read and transcribe invoices, match them to purchase orders and receipts and route payables for approval. Its purchasing tools search approved supplier catalogs, identify brand-approved items and show the budget impact before checkout.

Kate Adamson, Folio’s CEO
and co-founder
Kate Adamson, Folio’s CEO and co-founder, said users care more about outcomes than technology.
“We don’t scream about AI from the rooftops across our product, because we find that our day-to-day users do not really care how Folio works,” she said.
She described the independent owner-operator who handles purchasing, inventory, the front desk and overnight calls.
“They fill all roles and are constantly context-switching,” she said.
Free does not mean effortless. Members should confirm which features are included, which suppliers serve their market, how the platform connects with their accounting system and whether fees apply to payments or financing tools. Folio and Inn-Flow represent different approaches. Folio emphasizes mobile access and fast setup.
Adamson said the company gets “systems hooked up in days not months.” Inn-Flow builds procurement into a broader suite that includes accounting, labor and payroll. The choice mirrors a larger one facing owners: specialized tools that connect to existing systems or fewer vendors covering more functions.
Beyond the software
Not all back-office technology is software. Room controls, door locks and the property network increasingly generate operational data and they determine whether cloud-based tools work reliably.
Energy is a common target. In Wyndham’s survey, 54 percent of owners who had adopted AI said they used it for energy efficiency.

Sunny Kurani, enterprise account
manager al Vingcard
Vingcard, part of ASSA ABLOY, offers occupancy-based controls that adjust heating, cooling and lighting when rooms are empty. Its VConnect integration can pass arrival, departure and room-status information from the property management system to compatible access-management, room-control and workforce tools, reducing manual lock tasks.
Kurani said online locks can also report hardware status, such as battery levels, allowing maintenance before a lock fails. Vingcard offers these technologies as modules. “Operators can begin with the technologies that address their most immediate needs and expand over time,” Kurani said. Results from energy controls depend on climate, utility rates, occupancy patterns and configuration.
Kurani said working with a single provider across access, connectivity, energy and operations can also simplify purchasing, vendor coordination and support, which he described as especially valuable for independent hotels with small teams. The trade-off is greater dependence on one vendor, which makes contract and exit terms more important.
Positron said it plans to add support in early 2027 for the Model Context Protocol, a standard that lets AI agents connect with software systems, to give AI tools visibility into network performance and troubleshooting. Roadmap items like this are plans, not capabilities an owner can buy today.
Every cloud application also depends on the network.
Pierre Trudeau, president and chief technology officer of Positron Access Solutions, said cloud tools, AI and connected building systems can automate routine tasks and monitor equipment, “but they also increase the importance of a stable and secure managed network across the property.”
In older hotels, upgrading that network can mean opening walls. “For most existing hotels, upgrading their Gigabit Ethernet foundation to reach every guest room is costly and disruptive if it is forced to install new cabling,” Trudeau said. Positron’s GAM technology instead runs Gigabit Ethernet over existing coaxial or telephone wiring, which the company says can avoid extensive construction and keep rooms available during upgrades.

Pierre Trudeau, president and chief technology officer of Positron Access Solutions
Hotels can work through a Positron-certified managed service provider that handles network design, installation and monitoring, reducing the need for property staff to become networking specialists.
Positron said it plans to add support in early 2027 for the Model Context Protocol, a standard that lets AI agents connect with software systems, to give AI tools visibility into network performance and troubleshooting. Roadmap items like this are plans, not capabilities an owner can buy today.
Counting the cost
None of the five companies that responded to Asian Hospitality disclosed pricing, leaving owners to build their own cost picture.
The subscription is usually only part of the total. Owners should ask vendors to itemize recurring fees and whether they are charged per property, per room or per user; implementation, data migration and configuration; integration fees, including any charged by the property management system provider or brand; hardware and installation for locks, room controls and network equipment; training for new hires as well as the initial rollout; managed-service and support fees; and renewal terms and annual increases.
Training needs vary. Agilysys and Vingcard said requirements depend on the modules a property selects. Inn-Flow provides virtual sessions with hospitality specialists and a knowledge base to help employees learn specific workflows. Positron’s design shifts most of the technical work to service providers.
Owners with several properties should also ask whether pricing declines per property, whether one hotel’s configuration can be copied to the next and whether reporting can consolidate the portfolio. Folio, for example, said its platform groups purchasing and financial workflows across multiple properties.
Wyndham’s owners ranked the cost of AI tools, cited by 42 percent and staff training, cited by 39 percent, among their barriers to wider adoption. The fact finding that only 13 percent of participating chains report measurable returns suggests documented payback remains elusive, making an owner’s own before-and-after measurement essential.
Brand and contract
For franchised hotels, the brand shapes what is possible. Property management systems are often brand-mandated and third-party tools must connect to them. In Wyndham’s survey, 89 percent of owners said working with a brand is beneficial when incorporating AI, and 54 percent called it essential.
Before signing, owners should ask their brand representative whether a vendor is approved, whether it integrates with the brand’s systems and whether the brand plans to offer a competing tool. Positron, for example, said its solution is approved and deployed by most major brands; owners should confirm approval for their own brand and property type.
Integration promises deserve scrutiny. “The setup depends on the systems a property uses and how those systems make their data available,” Bridges said. Owners should ask which integrations are live today rather than planned, who maintains them and what happens when one fails.
Data privacy and security topped Wyndham’s list of barriers at 46 percent. Contracts should state who owns property data, how it can be exported if the hotel changes vendors, what security certifications the vendor holds, how payment data is protected and how quickly the vendor must report a breach. Contract length and termination terms matter more as tools become embedded in daily work.
People in charge
Technology executives expect AI to move from recommending actions to taking them. “As agentic AI develops, technology will move beyond identifying issues and recommending actions to executing multistep workflows with appropriate controls and human oversight,” Pitsikalis of Agilysys said.
Owners remain cautious. In Wyndham’s survey, 57 percent said they required human oversight to be comfortable with AI making operating decisions; 40 percent were comfortable without it.
Automating invoice matching is different from letting a system approve spending, change schedules or adjust rates. Owners should set approval thresholds, user permissions and audit trails before expanding automation, and decide in advance who is accountable when an automated decision goes wrong.
Where to begin
The research suggests owners should begin with a problem they can measure: invoice-processing time, purchasing variances, hours spent preparing reports, labor forecast accuracy, energy use or network downtime. Establishing a baseline and testing at one property or department shows whether a tool works in a hotel’s actual conditions. Vendor case studies can inform that decision, but they are no substitute for a property’s own numbers.
The next wave of tools will promise to act on their own: reordering supplies, adjusting schedules and troubleshooting networks. Those capabilities will depend on the same foundation the current ones do. The owners best positioned to use them may not be the earliest adopters, but those whose systems already agree on the numbers.
Get Started on The Right Solution Now. Speak to a Positron Representative Today!
Resources
Accelerating Digital Equity Nationwide
High speed broadband access is now more than ever a fundamental need for remote learning and tele-working, video streaming, entertainment and video conferencing especially with multiple users in the same household.
GAM Solution Brief
Positron’s G.hn technology provides managed non-blocking symmetrical Gigabit bandwidth services to subscribers over existing telephone wire or coaxial cable on extended distances, at a lower cost and with reduced complexity compared with G.fast.
Extending Gigabit Fiber services to Every Door without installing fiber to the door
Positron’s G.hn Access Multiplexer (GAM) can extend your fiber services to every door of an MDU/MTU over the existing telephone pairs or coaxial cable to deliver Gigabit broadband. The Positron GAM delivers symmetrical Gigabit broadband without the need to install fiber to every door.
Positron Access Receives Highly Commended Honor for 2021 Global Telecoms Awards
The Positron G.hn Access Multiplexer (GAM) redefines how fiber broadband services are extended over the in-building telephone or coax wiring to deliver Gigabit experience to every door.




