No. The easiest way to test is to try the SMART from the references, so you can quickly see how it works for those merchants. If you still want to test it before implementation, please contact our sales team.

The cost of one AI response using gpt-3.5-turbo (SMART Neo) is 1 message credit. The cost of one AI response using gpt-4 (SMART Quantum + Pulsar) is 20 message credits. The difference is due to the fact that the costs of gpt-4 are multiple times higher compared to gpt-3.5-turbo at OpenAI.

One SMART refers to a single chatbot that contains the company's own data and has a customized persona.

Your credits, also known as conversations, are renewed at the beginning of each month, regardless of when you subscribed to the service. So if you subscribe on March 15th, your credits will be renewed on April 1st.

No worries. Everything will continue as usual, and there won't be any unexpected additional costs. If this happens repeatedly over several months, then we will negotiate a solution with an expert.

If you choose to make it public, anyone can interact with your AI. You can also keep SMART AI private, meaning only you and your team have access to it. This is well suited for internal company support functions, e.g. as support for technical sales.

When we attach your material for download, we tell you the character count of all attached documents.

11,000,000 Characters (~140MB)

The data can be entered from several different sources, such as URL, PDF, FAQ, sitemap, text, Notion...). And if necessary, e.g. real-time data via API.

SMART can be custom integrated into any system.

There are several ready-made connections for popular applications in the world, for which SMART has ready-made connections. These are: Zapier, Wordpress, Whatsupp, Slack, Shopify and Messenger. These are easy to install and are part of the SMART Pulsar version.

SMART's data is stored in AWS services in the USA.

It is true that the GPT3.5 version sometimes invents too many things. Our experts can take different measures for this matter. But if nothing about SMART Neo helps and the hallucination continues, we can upgrade to SMART QUANTUM, which uses a more advanced gpt-4. It follows the basic prompt much better and doesn't hallucinate, but it's slower and more expensive than gpt-3.5 turbo (1 message using gpt-3.5 turbo costs 1 message balance. 1 message using gpt-4 costs 20 message credits)