Situation Breakdown
As AI Agents continue to expand their capabilities, a critical challenge is emerging around user preferences, identity authenticity, and data fluidity. With the dynamic nature of data, information, and knowledge, the necessity of loosely coupling AI Agents with data spaces—such as databases, knowledge graphs, and document collections—remains central to their effectiveness. At the same time, user identity authenticity is becoming increasingly vital, especially given the growing scrutiny by governments worldwide on the practice of surreptitious user profiling by applications and service providers.
Quick Recap
In previous newsletter editions, we explored the creation of AI Agents using natural language to handle tasks such as:
Declarative Querying: Enabling SQL, SPARQL, and GraphQL queries for boosting LLM response quality through Retrieval-Augmented Generation (RAG) and GraphRAG operations.
Feed Reading: Automating the consumption of content from preferred OPML, RSS, and Atom feeds.
Product Support: Offering AI-powered assistance for handling visitor queries on websites.
We also discussed how these AI Agents are deployable via the OpenAI Custom GPT store. Now, you can enhance this further by associating these agents—or custom derivatives—with a publicly accessible profile document, yielding two major benefits:
Digital Assistants that are loosely coupled with data spaces curated by you.
Digital Assistants linked to verifiable credentials for identity authenticity.
Here’s an example of one of my profile pages: Kingsley Idehen's Link-in-Bio.
Page with AI Agent Opened
Note the following:
Inclusion of Plain Old Semantic HTML (POSH) based metadata used by user-controlled identity authentication protocols such as RelMeAuth and IndieAuth.
X.509 certificates in both .crt and .pem formats usable by PKI-oriented protocols such as S/MIME for digitally signed and encrypted messaging.
AI Agent functioning as my Digital Assistant capable of handling conversations about topics of interest to me, culled from my various data spaces.
Demonstrating How It All Comes Together
Setting Up Your Own Profile Document with AI Agent Integration
To get started, follow these straightforward steps:
Create an HTML document that includes metadata necessary for establishing identity authenticity.
Embed your AI Agent using a chatbot widget.
Populate underlying data spaces linked to your AI Agent progressively as new information and knowledge are discovered.
These steps can be either seamless or challenging, depending on your toolset. In my case, I rely on the following from our vertically integrated product portfolio:
YouID Browser Extension (or App): For generating identity documents with hyperlinks as entity names.
OpenLink AI Layer: For creating and embedding AI Agents declaratively into HTML documents.
OpenLink Structured Data Sniffer (OSDS) Browser Extension: For progressively populating data spaces with new discoveries.
OpenLink Virtuoso: As the underlying platform for managing and integrating data spaces, with support for HTTP, ODBC, and JDBC protocols, alongside fine-grained attribute-based access controls.
OpenLink ODBC & JDBC Connectors (Drivers): For seamless, "no-copy" external database integration.
Conclusion:
AI Agents, when integrated with personal profile documents, offer a promising solution to the longstanding challenge of identity authenticity on the web. This approach not only honors user preferences but also introduces innovative opportunities for data monetization. It enables creators and curators to negotiate with LLM providers by safeguarding their data spaces with fine-grained access controls. This framework sets a solid foundation for authentic and secure AI experiences at scale.