What we do
We create really good conversations between humans and machines across the entire design and implementation process.
Throughout the entire process, we work user-centered and data-driven. And from the very beginning, we think about how we can imporve the application, step by step.
Design & Project Briefing
What is it actually about? We find that out at the start of every new project. Only when we really understand what the purpose of the application is can we think about it. We gather information and decide where to go into depth via ideation, user and stakeholder interviews. Even in this first step, we consider the technical possibilities and evaluate various implementations in order to find the ideal fit for the project.
This refers not only to the selection of the channel, but also to possible integrations of generative AI to further individualize and automate parts of the communication.
UX Research & UX Testing
When UX research and UX testing? Always! At every stage of the design process, we look at what users want and whether the conversation is flowing. We use qualitative and quantitative methods and work with all the data we have available. Even at a very early stage, we try to obtain as many insights as possible and, above all, real data.
We also rely on the power of Generative AI to optimize our processes and testing.
Concept & Design
Do conversations always follow certain patterns? Yes, but no. In our design process, we give the systems character, appearance and voice. To do this, we have developed tools that make the process and communication with stakeholders easier and bring the bots to life even before the first prototype. We also use the same methods for our generative AI branding - so that dialogs that are adopted by LLMs (Large Language Models) also have a distinctive language.
Based on this, we define the individual interaction steps. In order to design coherent conversations, we keep the user in mind as well as the technical limits and possibilities.
Implementation & Optimization
Which framework is the right one? There is no universal answer to this question. We decide individually for each project which platforms and technologies are the best fit. This starts with selecting the right framework for the bot logic and extends to setting up various prompt chaining and RAG (Retrieval Automated Generation) pipelines in order to seamlessly integrate ChatGPT or other LLMs (Large Language Models).
We ensure that both automated and non-automated testing methods are connected and that the success of the application can be measured for KPIs and OKRs through the right monitoring.
Architecture & Hosting
NLU pipeline, bot logic, TTS integration, backend and API integration? Exactly. We need these to make conversational interfaces more than just FAQ pages in a different format. The whole thing becomes even more challenging if we want to connect different LLMs (Large Language Models) and other options for Generative AI.
We work both locally and via various cloud services to meet all project and data protection requirements.
You want to talk about a project?