AI Versus Expert-curated Supplier Directories: How to Reliably Find Photonics Suppliers?

When using a general-purpose AI tool for supplier search, it helps to understand how such systems obtain and process information. Modern tools based on large language models (LLMs) essentially operate through two distinct mechanisms, each with its own limitations:

Finding suitable suppliers in photonics is substantially more complex than asking for a general technical explanation. It requires precise classification and reasonably complete coverage of the market — tasks for which general-purpose AI tools are not optimized. Common problems are:

• Misclassification: Closely related product categories are easily mixed up, leading to wrong supplier placement or missing them entirely.
• Ambiguity & terminology: AI struggles with ambiguous website content (production vs. mention) and may overlook suppliers using uncommon terminology.
• Incomplete coverage: Like a general web search, AI can easily miss some of the best suppliers, which is critical for high-value, specialized purchases.

These issues — in addition to outdated data and hallucinations — demonstrate that while fluent text generation may produce a great impression, it does not automatically translate into dependable supplier discovery, especially in a specialized field like photonics.

Specialized supplier directories are built with the specific needs of photonics users in mind, and with specific expertise. Their structure and maintenance methods directly address problems like those explained above:

• Accurate classification of products and consistent terminology are (in case of a high-quality resource) based on expert knowledge and obtained data on real product offerings. Suppliers are assigned to product categories only when they actually offer those products, avoiding misclassifications.
• Comprehensive and continuously updated coverage results from continuous work on data updates, improving classifications, etc. — not just a quick effort in the moment someone asks a question.
• Integration with technical background information is a special quality of the RP Photonics Buyer’s Guide: Supplier listings are closely integrated with encyclopedia articles and other technical resources. Users therefore have immediate access to authoritative explanations that help in evaluating product types and supplier capabilities.

Overall, expert-curated high-quality directories offer the completeness, precision, and technical consistency that engineers need when making sourcing decisions — qualities that general-purpose AI tools currently cannot guarantee.

By the way, we also substantially use AI for our resources — we don't present just old-fashioned static lists, exact-match searches and the like: There are search filters, semantic search, etc. For example, if you are not sure about the proper name of the category you search for, just enter what you think, and it will present a list of most closely related categories. (We have a list of AI-based features for users and one for advertisers). And in the background, we also use AI a lot — for example, for extensive quality checks. So you don't choose between AI or RP Photonics: We give you both expert content and AI-powered improvements.

Of course, human-curated directories are not automatically of high quality. For example, there are attempts to generate easy income without significant investment — for example, not engaging photonic experts but only cheap employees who have little understanding of the field.

Indeed, they do. In practice, however, this happens only to a limited extent and in ways that are not guaranteed to produce reliable results. They may start using only some more or less outdated training data before doing additional web searches when recognizing that the user needs more. If you send the same prompt multiple times, the AI may use a directory in one case but rely on different sources or even on other search and filter strategies in another. In that sense, the results can be even more variable than with a Google search, where at least the current rankings provide some limited consistency.

Moreover, retrieval-based AIs typically do not explore directories systematically. They may look at only a few pages suggested by a search engine and can easily misinterpret the content, for example by overlooking the structure of product categories or by drawing incorrect conclusions from short descriptions.

Therefore, it may well be that the results presented by an AI tool are influenced by what it found in the RP Photonics Buyer’s Guide, but there is no guarantee of consistency or completeness.

AI tools have amazing strengths — particularly in producing high-quality, fluent language (that's what LLMs were made for!) and in helping to explain or summarize complex topics. However, they can hardly replace an expert-curated directory when it comes to critical tasks such as systematically finding suitable suppliers in a specialized field like photonics.

The main reasons are structural: General-purpose AIs do not control their data sources, do not systematically use directory structures, and cannot guarantee completeness, correctness and consistency of supplier listings. Those are precisely the aspects where curated resources excel.

It is therefore likely that this situation will persist. AI tools will increasingly become useful companions for understanding technologies and formulating requirements, while expert-curated directories such as the RP Photonics Buyer’s Guide will remain the primary, reliable basis for supplier discovery. Professional work requires reliable data sources and consistently applied procedures, not some quasi-random bits and pieces.

Combining both approaches can also be effective — for example, checking with AI whether you can find more suppliers, or clarifying certain details. But for dependable sourcing decisions, the curated resource should come first. (Unless, of course, the AI tool is programmed to consistently use a specific high-quality data source.) Power users of AI understand how AI works, what its limitations are, and where to use other tools for best results.

Of course, AI will continue to improve in many respects. However, some limitations are fundamental. In particular, the quality and reliability of well-curated specialist directories are unlikely to be matched by general-purpose AI systems (which is also not really their goal). In fact, improvement of an AI system may well mean a better understanding of its own limits and a greater tendency to refer users to more appropriate tools rather than attempting tasks for which its results are unlikely to be reliable. After all, their core task is to serve users well.

By the way, as it is certainly in the interest of the photonics community that high-quality resources like those offered by RP Photonics will remain usable for many years to come, it is a good idea to use them consistently not only for your immediate benefit. How about making a bookmark in your browser's toolbar now?

You may also want to read the RP Photonics Spotlight article of 2025-02-27 titled Tools from RP Photonics for Responsible Purchasing. Responsible purchasing is more than just starting a search tool; it must involve more steps, e.g. getting clarity on the requirements and checking all relevant suppliers against those requirements.