Power your business with AI

With the rapid development of AI, the answer to this question changes every day, but it can be split into ‘how’ and ‘what’. We use AI-based tools to speed up the way we design and develop applications (using AI-enabled code assistants and design tools), and it also allows us to expand what web applications can do (see example cases).

With the advent of AI, especially large language models, we have got access to a gigantic factual database about almost everything in the world. That database can be queried in a natural language, understands the context, can combine facts, is capable of rudimentary reasoning, and is able to return answers in a structured form (e.g. JSON or XML). That opens up a completely new set of opportunities of what’s possible but requires a new set of skills, i.e. how to ask the right questions to get desired answers in the right format (so-called “prompt engineering”). Because large language models capture the state of the world from the past (September of 2021 in the case of ChatGPT), these answers need to be combined with the knowledge from specific up-to-date data sources. Of course, traditional mobile and web application development skills are even more required to present the results in interesting ways (see our case of Gaide.me application).

Integrating AI in your business allows you to achieve several benefits:

In a simple scenario, ChatGPT stores the content of the prompts for up to 30 days for business purposes, so one should not directly use protected information in the prompts. There are however ways to make it secure:

OpenAI's API conscientiously safeguards data privacy with robust security protocols designed to shield all data processed via its interface. While utilizing the ChatGPT Enterprise, access to conversation histories is exclusively reserved for organization-end users. OpenAI personnel are only authorized to interact with the organization’s data if incident resolution is necessary, explicit consent is offered for recovering user conversations, or as mandated by legal requirements.

In terms of data retention, ChatGPT's privacy policy states that ChatGPT Enterprise retains data securely, enabling features like conversation history. Organizations dictate retention durations, while deleted conversations are diligently eradicated from OpenAI systems within a 30-day timeframe.

One needs to be extra cautious and not use the responses from ChatGPT directly without double-checking, especially in high-stakes scenarios. The correct approach to maximize quality of responses and reduce so-called ‘hallucinations’ is a combination of:

Yes, that can be achieved in many ways. The two most important ones are:

In simple cases, it can take as little as 2-4 weeks. Please contact us to discuss the details of your specific case.

There are many options, but assuming the solution will be using currently the most popular AI model - ChatGPT, the cost would be as follows:

ChatGPT operates on a token-based pricing system, where each conversation's cost is derived from the quantity of tokens used. A single token roughly counts as 0.75 words. Current rates per 1,000 tokens are:

To obtain specifics about GPT-3.5 Turbo and GPT-4 pricing, kindly refer to the following link.

For an illustration, let's assume a monthly usage of 10,000 interactions with each using approximately 1,000 words for both input and output. The corresponding cost impact would be:

Please note, these are approximate costs and can vary based on actual usage.
Contact us to learn about other possibilities.