Embedded Files
I build AI systems that interact with people.
My current work explores agentic voice AI and multimodal tools — systems that can call sources, conduct conversations, and transform information into structured outputs.
Projects include Aila Media, an AI voice agent for interviews and reporting, and Vegan Converter, a camera-based AI that turns photographed recipes into plant-based versions.
Aila calls. You report.
Aila calls. You report.
Aila Media is an exploration of agentic AI applied to voice-based reporting and research. The project develops systems where AI agents can independently conduct structured conversations with humans—placing calls, asking questions, listening to answers, and producing usable outputs such as transcripts, summaries, and draft material.
At its core, Aila experiments with how voice agents can extend human investigative capacity. Instead of replacing journalists or communicators, the system is designed to handle the most time-consuming layer of work: reaching sources, conducting structured interviews, capturing accurate speech-to-text, and organizing the material into formats that humans can review, verify, and edit.
Technically, Aila operates as an orchestrated stack of AI components. A journalist or researcher defines a hypothesis, topic, or interview brief. From that prompt, an AI agent prepares a conversational structure and initiates a phone call. During the call, real-time speech recognition converts audio into text, language models interpret the meaning of the answers, and the system dynamically generates follow-up questions. After the conversation, the full interaction is delivered as audio, transcript, and structured written material.
The research focus lies in agentic behavior: systems that can act autonomously within defined boundaries. Rather than a static chatbot, the agent can conduct many conversations in parallel, adapt to responses in real time, and gather primary information directly from human sources.
Because the system interacts with people directly, ethics and editorial responsibility are central to the design.
Aila is built around principles such as:
Transparency that the caller is an AI system.
Human editorial oversight of all published material.
Consent and clear communication with interview subjects.
Preservation of source integrity and traceable audio records.
Careful handling of bias, hallucinations, and misinterpretation.
The ambition is to explore how AI-native voice systems can responsibly support journalism, research, and communication, allowing professionals to spend more time on analysis, verification, and storytelling—while AI handles the operational work of gathering information.
Aila Media is currently being developed as a prototype and research project focused on the intersection of AI agents, voice technology, and responsible information gathering.
Currently building ailamedia.com
AI Vegan Converter Camera
Vegan Converter is a browser-based AI application that turns a photo of a dish or a printed recipe into a vegan version in seconds. The app activates your phone camera directly in the browser, so there is nothing to download. You simply photograph a plate, a cookbook page, or a printed recipe, and the system analyzes what it sees.
The technical core combines computer vision and language models. First, the app uses image understanding to identify either the dish itself or the recipe text captured by the camera. If the user photographs a finished dish, the model infers the likely recipe category, key ingredients, and cooking method. If the user photographs a recipe from a book or magazine, the system extracts and interprets the text and structure of the recipe. In both cases, the output is a reconstructed vegan version.
The conversion layer is not based on crude ingredient swapping. The AI evaluates the function of each non-vegan ingredient in the recipe. It asks what the egg, butter, cream, cheese, or meat is actually doing in the dish: binding, emulsifying, adding fat, creating texture, building umami, or contributing protein structure. It then generates replacements that match that function as closely as possible in a plant-based version.
The project builds on my earlier work with plant-based cooking, including the book How to Be Vegan in 28 Days, published internationally by Headline Home.
Journalist. Founder. AI Builder.
Journalist. Founder. AI Builder.
I build AI tools that interact with the real world — voice agents that conduct interviews and multimodal systems that transform information into practical outputs.
After 20 years in journalism and publishing (TV2, Aller Media, Norad, Gyldendal), I am now building AI-native tools for reporting, research and everyday decision making.
Founder of Aila Media, exploring agentic voice AI for journalism.
Photo: Jim Hensley and Nina Dreyer Hensley
Page updated
Google Sites
Report abuse