This Founder Brought AR to Custom Framing

Ben Koren incorporated Frameology to take his framing business online. According to Koren, the early results appeared to challenge the concept: the company generated $3,000 in revenue in its first month, recorded the same amount in its second, and then experienced essentially flat sales for six years.: $3,000 in revenue the first month, the same the second, then six years of essentially flat sales. The client says that, 13 years later and without having raised venture capital, Frameology is a profitable e-commerce brand generating eight-figure annual revenue. 

The company has developed a 3D configurator that renders a customer’s photograph in different frames, browser-based augmented reality (AR) that displays the proposed piece on the buyer’s wall at scale, and a life-size paper template intended to assist with installation.

As more home decor purchases move online, Frameology has focused on the difficulty of helping customers visualize personalized products before ordering them. Koren’s company, Memori, now licenses the technology developed for Frameology. Through Powered By Frameology, retailers, registry platforms and photo-focused apps can offer the system under their own brands. The platform includes the configurator, tools powered by artificial intelligence (AI), and access to a U.S. print-on-demand supply chain.

From a Manhattan frame shop to an online configurator

Koren’s family has worked in framing since 1986, when his father operated a high-end frame shop in Manhattan. Koren saw an opportunity to bring that experience online at a time when he considered the available digital framing options limited.

Frameology took two years to launch. However, the initial website, which offered expensive frames comparable to those sold through the family’s physical shop, attracted few customers. The company reported $3,000 in revenue during its first month and the same amount during its second.

Koren eventually concluded that the website itself was not the central problem. Instead, the company needed to reconsider its assumptions about its customers and their purchasing decisions. Custom framing presented logistical and visualization challenges that were not immediately apparent from the product itself.

Frameology’s sales remained largely flat for six years as the company tested different approaches. Koren says this process demonstrated that technology could accelerate certain parts of the business but could not replace the time required to understand customer behavior.

Serving different buyers

Frameology’s experience led the company to distinguish between two broad types of customers. One group consists of gift buyers working toward birthdays, anniversaries, or other deadlines. These customers generally want a straightforward purchasing process and clear delivery information. The other group consists of home decorators making longer-term decisions about pieces that will remain on their walls. These buyers may want greater control over individual design details and more time to evaluate their options.

The preferences of the two groups can conflict. Features that make the process feel faster for gift buyers may provide insufficient detail for decorators, while additional customization options may create friction for customers seeking a quicker purchase. According to Koren, many of Frameology’s product decisions have been shaped by efforts to accommodate these differing needs.

Using AR and 3D to address design concerns

Frameology identifies concern about the final appearance as one potential obstacle in the custom-framing process. Because a frame is produced for a particular photograph, set of dimensions, and intended space, customers cannot evaluate it in the same way they might assess a standard product in a physical store. This may cause some buyers to hesitate before completing an order.

The company’s configurator allows users to upload a photograph and view it in different frames through a 3D preview. Using a mobile browser, customers can also display the framed piece on their wall at approximately its intended scale. The feature does not require a separate app download. According to Frameology, these tools are intended to give customers a clearer impression of the finished product before purchasing it.

Gallery walls presented an additional challenge because customers needed to plan the placement of several frames and then reproduce the arrangement during installation. Frameology responded by developing a life-size paper hanging template.

The company views the template and its AR tool as two versions of the same concept. The AR preview helps customers assess a proposed arrangement before placing an order, while the physical template is intended to support installation afterward. Both tools respond to questions about how the completed arrangement will appear on the customer’s wall.

Where AI fits in the process

Frameology uses AI for selected predictive tasks while retaining human review for decisions requiring contextual judgment. One example is ImageIntent, a system that analyzes an uploaded photograph and identifies what it appears to depict, such as a wedding, a newborn, or a new dog.

The company treats the photograph selected by a customer as a source of information about the occasion or subject behind the purchase. This categorization can then inform how the platform presents relevant products or options.

According to the client, developing the system took close to a year. A member of the Frameology team continues to manually review 1,000 categorized images each month and record errors. The company reports that the system’s error rate declined from approximately 30% during its early development to 3%. These figures have not been independently verified.

Koren attributes the reported improvement primarily to repeatedly revising the company’s prompts using the errors documented through human review. Frameology’s approach is to use AI for classification while leaving decisions about what should be detected and how the resulting information should be applied to people.

Licensing the stack through Powered By Frameology

Powered By Frameology is a white-label commerce platform for personalized physical products. Retailers, registry platforms and photo apps can integrate components of the system into their own branded customer experiences. These components include the 3D and AR configurator, gallery-wall builder, ImageIntent system and order-processing infrastructure.

The back-end system routes orders to a network of U.S. print-on-demand manufacturers. According to the company, products can be shipped and customer service provided under the partner’s name, without displaying Frameology’s brand to the end customer.

Frameology says partners can implement the platform through embedded code, an API, or a managed storefront operated on their behalf. The company positions the managed option as one way for retailers to introduce a personalized product category without assigning the entire project to their internal technology teams.

The platform is also designed to reduce the need for partners to hold inventory or make an upfront investment in manufacturing infrastructure. However, the precise technical, financial, and operational requirements may vary according to each implementation.

Frameology tests the features offered to partners through its own consumer website. Koren describes Frameology.com as the platform’s first customer, with new functionality evaluated through the company’s existing retail operation before being made available to licensees.

Turning experience into a licensed product

Koren’s decision to license the system reflects his view that software alone does not account for Frameology’s development. Although AI and other tools can accelerate software production, he believes they cannot independently reproduce the operational knowledge accumulated through years of customer testing.

That knowledge includes decisions about which customization choices to present, which ones to remove, and where customers are most likely to hesitate. It also includes considerations specific to different purchasing situations, such as the needs of a gift buyer placing an order shortly before an occasion. Powered By Frameology incorporates these decisions into the product offered to partners.

Frameology’s experience suggests that adopting technology in an established retail category involves an ongoing process of testing and learning rather than simply purchasing new tools. Its AR features, AI system and paper installation template were each developed in response to customer concerns the company had identified through its own operations.

Whether licensing the platform will become the company’s next successful line of business remains uncertain. Its performance will ultimately depend on how retailers, platforms and their customers respond to the offering.

Entrepreneur Media was not involved in the creation of this content.