Streamlining Product Communication with Knowledge Graphs

Specifiers, installers, and users of product systems often face the challenge of finding specific information tailored to their unique use cases. The available information is typically general and requires sifting through multiple PDFs and web pages to find relevant details. For example, specifiers often receive general specification information in siloed documents that cannot anticipate every possible configuration. Similarly, installers deal with multiple PDFs that cannot describe how to make connections between all compatible products. Research into product development in the construction industry has shown that simplifying product communication for existing products yields a much higher R&D ROI than developing new products. This realization led to the development of a systematic approach using knowledge graphs described here.

Utilizing Knowledge Graphs

A knowledge graph can serve as a semantic structure to organize product communication content in an intuitive way for users, rather than following a manufacturer's business architecture and conventions conceived in print carried over into digital format (i.e. letter format pdf's stored in an online document library). Here's how a knowledge graph, paired with componentized content, can be beneficial for product communication:

‍

• Create a Product Communication Strategy: Prioritize high-reward content to ensure that users get the most critical information efficiently. While it's conceivable to create an exhaustive library, it isn't practical. A knowledge graph allows you to prioritize the highest reward content and ensure it is more complete and useful than your competitors.
• Manage and Update Content at Scale: Use automation to create, manage, and update product communication content, ensuring efficiency through asset reuse.
• Optimized Content Delivery: Deliver step-by-step, illustrated information for specific queries such as "What products do I need for an air purification system in a New Delhi university lecture hall?" or "How do I install air filter product X with control product Y and air quality sensor Z in a commercial building in California, ensuring it is up to code?" The knowledge graph can aggregate dispersed information from multiple sources into a coherent, easily navigable format.

‍

Examples of the Challenges and Solutions Offered by Knowledge Graphs

Specifiers, Architects & Engineers

• Challenge: Specifiers often receive general specification information in siloed PDFs, web pages, and product trainings. For product systems, these siloed documents cannot anticipate every possible configuration without being excessively long and unmanageable. Moreover, specifiers frequently need to learn the rules and parameters of the system they are specifying, which can be time-consuming and error-prone. Maintaining detailed PDFs for each product's interaction with every other related product would require constant updates across multiple languages, posing a significant logistical challenge.
• Solution: Componentized content held together by a knowledge graph. For example, a specifier looking for an air purification system for a university campus in New Delhi can query the system for region-specific products that meet necessary specifications. The knowledge graph retrieves and assembles the relevant content, including design considerations and drawings, and even contact information for the local technical sales agent, or pings the agent with the request so they can follow up. Specifiers are paid for their time to create accurate designs, so making their job easier by specifying your products is a proven win.

Installers & Contractors

• Challenge: A contractor setting up a system of connected products typically deals with multiple PDFs containing general information. These siloed documents cannot anticipate every possible configuration without being excessively detailed and, as a result, can be vague. Furthermore, any one of these product-specific documents cannot feasibly describe how to make the connections between all past, present, and future products it is compatible with, leaving contractors without the specific guidance they need.

• Solution: Componentized content managed by a knowledge graph can deliver precise, step-by-step instructions, complete with diagrams, tailored to the unique configuration of products for the specific use case, even if it has never been done before.

End Users

• Challenge: When troubleshooting, users often find answers buried in lengthy PDFs, available only in specific languages.
• Solution: A knowledge graph can provide specific troubleshooting steps and deliver them in a user-friendly manner, potentially preventing unnecessary product returns.

The System at Doran Design Studio

At Doran Design Studio, we utilize a knowledge graph that includes nodes for products, use cases, regions, and design considerations. The data can be stored in and delivered from a client's PIM/DAM/CCMS system[s]. The service Doran Design Studio offers involves slicing up existing content and creating new content at scale using the knowledge graph as a scaffold. Even if the client does not have their own knowledge graph, Doran Design Studio's use of one significantly increases the efficiency of content creation, management, and prioritization, even if the output is still siloed PDFs (although we recommend it is not).

If the client does not have an existing knowledge graph, we can help deploy one using standard industry-recognized ontologies and easy to use software systems.

‍

If the client has an existing knowledge graph, then a user's query can traverse the graph to find all relevant product communication content in the PIM/DAM/CCMS systems and deliver it without extra noise. An LLM could feasibly generate an adequate specification, or step by step installation guide for the user if the content is structured in this way.

‍

Large consultancy firms offer similar services but they will cost exponentially more, and the incentive for these companies is not to finish the projects but to continue to profit from its upkeep indefinitely. Doran Design Studio has already implemented these systems for communicating product systems and handed over the maintenance to clients.

Knowledge Graphs and LLMs

Componentized content, when structured within a knowledge graph, transforms your product information into highly readable and machine-friendly data. While a knowledge graph is not artificial intelligence itself, it organizes your product information semantically, making it more accessible to LLMs (Large Language Models). This structured approach overcomes the challenges posed by non-machine-readable content and complex formatting in PDFs. By componentizing your content within a knowledge graph, machines can efficiently understand and deliver relevant information to users, thereby enhancing the overall product experience.

‍

Key Benefits

• Strategic Content Prioritization: Deliver high-reward content efficiently, ensuring critical information is accessible and competitive.
• Scalable Content Management: Automate the creation, management, and updating of content, improving efficiency and asset reuse.
• Enhanced Content Delivery: Provide tailored, step-by-step information, making product systems easier to understand and implement.
• Customer and Specifier Retention: Offer precise, user-friendly support, enhancing customer satisfaction and loyalty, leading to increased market share.