How it works
Imagine you're in a store, and you ask a new sales assistant for a "four-person tent". They take your request quite literally and guide you straight to the tents tagged as "four-person tents". This is like the traditional chatbot or keyword-based search – it gives you exactly what you asked for, no more, no less.
Now, imagine our service, Gigachat, as a seasoned camping gear sales expert. When you ask this expert for a "four-person tent", they don't just think about the number of people. They recall your past preferences, understand your typical camping conditions, and consider your overall needs.
For instance, you've mentioned before that you frequently camp in rainy areas. The Gigachat expert might recommend a "six-person, water-resistant tent". Even though you didn't ask for a six-person tent or mention the weather, Gigachat factors these in. The bigger tent offers more space when you're stuck inside due to rain, and the water-resistance is perfect for your usual camping spots. Gigachat can then explain why this option might be a better fit for your needs, essentially offering you a more personalized and relevant solution, like an expert upselling based on your unique needs.
This is how Gigachat works: it converts both your request and all the items you've fed into it (like the tents) into mathematical vectors that capture the essence of their features. It then uses these vectors to find the items that share the most similarities with your request. Once it has these options, Gigachat, like the expert sales assistant, provides a clear, human-like explanation of why these options are the best fit for you, enabling an effective, contextually-aware 'upsell'.
Explore our use cases to see how Gigachat can elevate your business.