Definitely Doug (09/13/2024): AI and the Last Frontier?

AI and the Last Frontier?

If I were to pick one hotel business process as the least automated in the 21^st century to date, it would probably be procurement. 

While procurement solutions have been around for decades, they have tended to focus on internal processes and controls. The actual selection and ordering of goods and services for the hotels are often still handled by telephone, via traveling sales reps, or by email. Where orders have been digitized, the buyer usually uses the website of a large supplier or General Purchasing Organization (GPO) to place orders. In most cases it is still an unintegrated process, requiring manual entry of orders without easy access to data on budgets, inventory, accounts payable, or payments. Most of these interactions are still paper-based (even if that paper may now be in the form of a digital image) and involve significant manual effort.

Compare this to our personal buying behavior, where we have all become accustomed to, even dependent upon, sites like Amazon that manage the process of locating what we need; comparing prices, features, and packaging; processing payment; and arranging delivery. Not surprisingly, many hotel companies are challenged to keep their colleagues from using Amazon for many hotel supplies, because it is simple, easy to use, and (usually) price competitive.

Yet procurement is one of the most important sources of bottom-line profit. Unlike revenue (which almost always comes with associated costs), each dollar of savings in purchasing falls 100% to the bottom line. With costs of food, supplies, and other essentials rising faster than room rates in recent years, cost control has become even more critical. And labor shortages have led many hotels to use fewer suppliers because of the time involved with manually consulting multiple websites to compare prices, or of managing orders and payments with too many suppliers. But fewer suppliers mean less price comparison and optimization, and higher costs.

Today’s column will explore some new capabilities to optimize procurement, many of them enabled by Artificial Intelligence (AI) technologies that have only matured sufficiently in the past couple of years. I cannot say that the new products are fully mature, but some of them are truly impressive and worth a look. Many hotels who were previously using typical industry approaches have found cost savings on the order of 10%, and some outliers who previously had no procurement technology or GPO have reportedly reduced costs by as much as 25%.

Each of the leading-edge capabilities I will cover below has been implemented by at least one company I spoke with (with the sole exception of ESG reporting, which several are evaluating). Most were implemented by, or on the near- or intermediate-term roadmap for, several companies. I expect rapid evolution of these capabilities among both newer and established players over the next one to two years. This article will give you some ideas of what you can look for as you evaluate solutions.

With a topic like this, I must first narrow the scope, lest what is intended as a blog risks ballooning into a book. I will focus on systems that address selecting, ordering, and receiving goods and services, and passing off order information to be recorded in an accounting system for payment and reporting. I will exclude bid processes (e.g. request for quotation and similar) and reverse auctions. Also, procurement interacts tightly with other hotel processes, such as menu management and costing, inventory, point-of-sale, budgeting, accounting, and supplier payments. Some procurement software packages include modules for one or several of these, while others integrate with selected third-party solutions. However, this article will touch on these capabilities only tangentially.

There is no one-size-fits-all solution for procurement software; much depends on the size and complexity of the hotel or group and what other systems it is using. Smaller operations may opt for simpler, easier to use solutions that provide basic capabilities for inventory management and payment. My objective here is not to tell you what the best solution is, but rather to make you aware of emerging capabilities which, if you have not reviewed this space in the past couple of years, you may not even have imagined.

As always, I am indebted to senior executives at several companies for helping me understand the emerging best practices through in-depth interviews. These included BirchStreet Systems, DiningEdge Technology, Folio, FutureLog, Lilo, and Reeco, and I thank them profusely for their time and wisdom. I also reached out to three other companies that did not respond to requests for interviews.

Common Capabilities for Procurement Systems

For readers who may be unfamiliar with procurement systems, I will outline some of the common capabilities that have (in general) been present for many years and that are widely deployed in most commercial products. For hotels, many of these will be baseline requirements, but not state of the art.

• Hotels buy many different goods and services, often from dozens of suppliers. While many use GPOs for the bulk of their purchasing, there are always other suppliers, whether for local or specialty items, furniture and other long-lived assets, regulated items such as liquor, or services or utilities that GPOs may not cover. A true understanding of spending (and saving potential) requires coverage of as many purchased items as possible.
• While many types of businesses use procurement systems (often embedded within an Enterprise Resource Planning platform), hospitality businesses have specialized needs that generic solutions are often challenged to address. Most hotel companies are small or medium sized businesses that need simplicity, but also have procurement complexities (such as changing menus and food spoilage) that retail, service, and many other businesses do not. The hospitality-focused solutions addressed in this article try to find the sweet spot that provides both necessary functionality and ease of use.
• Support for desktop, web, and mobile devices.
• The ability to normalize quantities to standard units to facilitate price comparisons. This sounds simple but often is not, because suppliers may offer both different package (item) sizes and differing numbers of packages (items) per case, pallet, etc.
• Budget (“checkbook”) tracking, enabling each manager to see how much of their budget they have spent to date, and how much remains. Where authorized, this feature may also allow a manager to shift unspent funds from one account to another.
• Managerial override approvals for out-of-band purchasing requests, such as exceeding budget, buying from non-preferred suppliers, or buying prohibited items.
• Multicurrency support, including the ability to compare items priced in different currencies.
• Roll-up reporting and analytics for each outlet, property, group of properties, or the entire enterprise. For multinational hotel groups, intelligent handling of “outlier” currencies (e.g. those experiencing hyperinflation) to avoid contaminating roll-up results.
• Support for inventory integration and manual inventory counts. Most enable “order guides” that can be customized to each hotel and used to manually scan items in storage rooms, shelf by shelf, and record those requiring replenishment on a mobile device (including while offline, since connectivity may be limited). Hotels can organize order guides in whatever way they find most useful, such as one for each storeroom, with items ordered as they appear on shelves.

Emerging Best Practices 

Several new-entrant software companies are leveraging modern technologies, especially various forms of AI, to overcome key challenges faced by procurement software and processes. Some legacy players have also added one or two of these capabilities or have them under development or on their product roadmaps. 

Supplier Onboarding and Setup: AI-powered systems can now scan months of past invoices to identify the items a hotel purchases and its suppliers, and then use the information to create order guides. They can then use AI (and the hotel’s logon for each supplier site) to find each item in each supplier’s online catalog, and to map the sites so that they can quickly retrieve descriptions, photos, prices, and packaging information in real time when needed. In some cases, they can also learn how to place an order without human interaction. Some of these products reduce setup time for a hotel from weeks or months to just an hour or two (and have thousands of suppliers already set up). Minimal setup costs enable some vendors to offer low-risk trials without big upfront fees.

Real-time Pricing: These systems can fetch real-time product availability, packaging, and pricing from multiple suppliers with no manual effort. Whereas historically this required a buyer to sign into multiple supplier websites, navigate to the item, copy costs and packaging units to a spreadsheet, and calculate the best deal, that can now happen entirely behind the scenes; the buyer simply sees the results. Real-time pricing also eliminates the inevitable price discrepancies that occur when using prices that are updated only periodically. Contracted pricing arrangements are fully reflected.

Price Comparisons: These systems can display on a dashboard the current prices for an item are from every configured supplier who carries it. Approved and preferred suppliers can be identified, and price comparisons can show the final delivered price including shipping (actual if known, estimated if not). Comparisons are converted into standardized units. Minimum orders, delivery timetable, order cutoff times, and out-of-stock indicators may also be included for quick reference.

Product First, Supplier Second: These systems start from the assumption that you need to buy a particular product, not that you want to place an order with a particular supplier. You choose the product and see who has it at what price, and then choose the supplier. You can choose a preferred supplier even if their price is higher if you like, but now that will be a conscious choice, made with full knowledge of the impact on cost. AI may be used to remember what overrides the user makes and suggest the same overrides in future purchases.

Plain Language Search: While frequently purchased items will have been identified during setup based on past invoices, most systems can also handle plain-language search terms (similar to Amazon) for first-time purchases. They can quickly surface matching items from any configured supplier catalog.

Order Quantity Optimization: The best systems can suggest the optimal order quantity based on the historical relationship between orders and other metrics (such as hotel occupancy) and forecast estimates for those metrics. They can also consider (or highlight) savings available through quantity discounts or incentive thresholds that may be within reach.

Consolidated Ordering: Many systems can consolidate orders across multiple suppliers so that the hotel has the illusion of ordering from a single source, while the system is splitting out the orders behind the scenes. Intelligent consolidation can include various elements of optimization such as meeting minimum order requirements or accelerating delivery times. At least one vendor will also (optionally) consolidate payment, so that the hotel has only a single invoice to pay. Most solutions provide at least some visibility of incentives and progress towards rebate thresholds; these can be used to optimize the consolidation or to suggest possible overrides (e.g., choose a different supplier or increase the quantity if you are modestly short of the rebate threshold as month-end approaches).

Real-Time Order Placement: Orders are placed behind the scenes and in real time with suppliers that support order placement via Electronic Data Interchange (EDI) or Application Programming Interface (API). In other cases, orders may be placed by interacting robotically with suppliers’ extranets. For less automated suppliers, orders can be sent by email (often with a link that enables the supplier to view the order in the hotel’s system). 

Electronic Document Ingestion: Invoices and bills of lading can be extracted from emails or scanned, and their contents analyzed and loaded into the procurement system as well as transferred to the accounting system for payment and reporting.

Two- or Three-Way Match: Systems can match invoices and receiving documents to orders (called a three-way match if all three are checked); receiving operations can also record discrepancies such as partial shipments, damaged goods, or substitutions (typically using a handheld device). Smaller hotels often handle receiving less formally, in which case a two-way match (order to invoice) may suffice. In all cases, price discrepancies, substitutions, and quantity changes will typically be flagged. 

Automated GL Coding: General Ledger (GL) account codes must be assigned to each purchase prior to entry into the accounting system. While historically this was often a tedious, manual process, the newer systems quickly learn what GL accounts are used for various purchased items and automatically suggest them for each new purchase. At least one system recognized that the same item might need a different GL code if it is purchased by housekeeping than by the food and beverage department. In any case, the hotel can always override the choice, but this should rarely be necessary.

Receiving: Some systems provide a mobile receiving app for use on the loading dock. It will show suppliers with deliveries expected that day, check for changes made after the order was placed, and allow the receiving clerk to report missing or incorrect or damaged products, or quantity discrepancies. One product supported weight and temperature measuring devices that can further help in identifying issues, especially with food shipments.

Off-Platform Buying: Some of the systems can record purchases made outside the platform (for example, if engineering buys a tool at a local home goods store). This can be important as a metric for management companies that want to enforce procurement policies at their hotels, and also to identify suppliers that should be added to the system. How this information is obtained matters: if it relies on manual entry by hotel staff, it may never get entered and you will never know how much off-platform buying is happening. It is better if it comes from the accounting system, where even off-platform purchases will be recorded somewhere.

ESG Reporting: A few companies have on their roadmap the addition of Environmental, Social, and Governance (ESG) metrics. These are increasingly required or desired for regulatory reporting or third-party certifications for hotels. They include metrics such as carbon footprint of the item, its packaging, and transportation; purchasing from minority-owned businesses; and the like. Many suppliers are already collecting some of this information and looking for ways to highlight it to buyers. Some procurement software companies are interested in helping to fill that gap and to assist hotels in improving their ESG reporting capabilities and footprint. This will be a space to watch in the next 12 to 24 months. 

Conclusion

For all the hype AI has gotten in the hotel industry, some of the most solid innovations it has enabled have been in back-of-house systems, where complex manual processes can often be re-engineered from the ground up. Procurement has been ripe for disruption and is now the target of several new and innovative products.

Is procurement an application where AI is ready for prime time? Based on the market traction some of these companies have gotten in recent months, it seems like the answer is yes, at least for smaller, simpler operations, and quite possibly for many larger ones. There are significant potential savings both from procurement costs and labor savings, that simply are not practical to pursue without AI.

As always, feedback to my articles is welcome. Since the host site does not support discussions, I will post a link to this article on my own LinkedIn page once it has been published, and I invite you to comment, like, or share from there!