We previously explained how generative AI can improve customer care services
Hereafter, we bring an example of how generative AI-based query systems can simplify access to highly specialized documentation. We achieve this through discussions on mixology and Cocktail Engineering—like a Daiqueri, or rather, a D-AI Query!
What’s Cocktail Engineering?
Cocktail Engineering (https://www.cocktailengineering.it/) is a cutting-edge initiative in the world of mixology, established in 2016 by Giovanni Ceccarelli. Originally launched as a blog focused on the art of cocktail-making, it has evolved into a full-fledged company and a thriving community of professionals.
Why mixology?
Mixology is both an art and a science, focused on crafting cocktails and mixed drinks. It involves skillfully blending various ingredients such as liqueurs, syrups, juices, spices, and other elements to create drinks that are both balanced in flavor and visually captivating.
Unlike a bartender, a mixologist is a specialist dedicated to innovation and research in the world of cocktails. A mixologist needs a diverse skill set, including a deep understanding of ingredients, mastery of mixing techniques, and knowledge of cocktail history and cultural traditions related to beverages. And, of course, creativity plays a vital role in pushing the boundaries of this craft!
- Accuracy in Recipe Presentation. There’s no room for alterations when it comes to ingredients and dosages! Sequences must also be strictly followed, especially for more technical preparations such as syrups. The equipment suggested must align with the specific requirements of each recipe.
- Avoiding Errors and Maintaining Context. A mixologist wants the proposed recipes to reflect their unique style and expertise. The credibility of an expert depends on this precision and authenticity.
How Does DAI Query Work?
At the core of DAI-Query is a Retrieval Augmented Generation. This technique leverages textual content from the “Cocktail Engineering” community blog, along with published videos and webinars.
When a user submits a request, the engine identifies the most relevant content through a scoring system, which filters out less relevant materials
- This approach ensures that responses are generated solely based on extensively covered topics from the site, maintaining the accuracy and consistency of results.
- Users can also access key articles, enabling them to explore topics of interest in greater depth.
- To ensure procedural accuracy, Knowledge Graph-based techniques are employed to maintain adherence to dosages, pairings, and preparation methods.
- Tone and Prompt Engineering
The tone of responses has been carefully calibrated through prompt engineering to provide the right level of summary. And why not add a touch of creativity?
What Are the Advantages of DAI Query?
At the core of DAI-Query is a Retrieval Augmented Generation. This technique leverages textual content from the “Cocktail Engineering” community blog, along with published videos and webinars.
- Additionally, user requests drive the creation of new content, tailored to the community’s curiosity and evolving tastes.
- The technology behind DAI Query is highly adaptable to other knowledge bases.
- For example, it can be applied to tasks such as consulting ingredient lists and blacklists in cosmetics (as seen in MATILDE) or reviewing tender documents and maintenance procedures.
