Centigrade Academy
From initial orientation through human-controlled agents to self-directed AI automation: our formats systematically build on each other. Find your station - and grow from there.
Station 1
You want to understand what human-centered AI means and be inspired by concrete use cases. Doing so allows you to speak confidently in discussions with Product Owners, Requirements Engineers, UX experts and AI developers.
Participants will learn how human-centered methods and AI can complement each other positively by automating or assisting tedious or overly demanding tasks with AI. In addition, participants will discover potentials as to where AI also opens up new, previously unknown possibilities (e.g., interviews with AI personas or generation of visual diagrams without graphical tools).
From the correct derivation of requirements from given use cases in the form of scenarios, user needs, epics, and user stories, to the creation of AI assistants to support requirements engineering. This includes, among other things, the creation of AI persona agents to validate requirements using Computer Using Agents (CUA).
The training offers a wide spectrum, from the basics of prompt engineering to insider tips and tricks for professional prompts in the own domain. Prompting frameworks such as C.R.I.S.P.E., A.P.E., or R.A.C.E. and their use in daily UX work, combined with their application in various models like ChatGPT, Claude, or MS Copilot, will be covered. Participants will gain confidence in prompting and also learn how to create, reuse, and share prompt templates with colleagues.
Whether it is AI usage, vibe-coding, automation, or the technically-feasible specification of requirements: nowadays, one repeatedly reaches a point where a basic understanding of coding is crucial to effectively achieve one's goals within a team. This training therefore introduces the basics of low-code and specifically addresses people who have had little to no programming knowledge so far.
As early as 2025, the year of "AI agents" was proclaimed as the next evolutionary stage of GenAI. Agents possess a memory, can perform tasks autonomously, formulate plans, write code, and collaborate with one another. This training demonstrates the opportunities and risks of using AI agents to increase efficiency and automate complex problem-solving, and prepares participants for the skillset required to successfully keep AI agents and their token hunger in check.
Participants will learn how to make successes and failures measurable and visible for their stakeholders. An overview of common success metrics and KPIs will be provided, and the advantages of usage data tracking and evaluation will be highlighted.
Station 2
You already use AI but want to integrate it more deeply into your daily professional work and be able to show AI basics to your colleagues by yourself.
The training presents an approach to describing requirements and background knowledge that is understood and picked up in the language of software engineers and AI coding agents. Instead of describing concepts "intuitively," it demonstrates how to define specifications and documentation using established methods and terms, so that they can be more effectively incorporated into the final product.
In this training, participants will build upon their beginner knowledge to learn new coding approaches (Workflow Automation vs. Vibe Coding) and their advantages and disadvantages, in order to reach their goals faster together with AI. For example, it will be shown how certain processes can be automated with n8n using Small Language Models, targeted Vibe Coding, or even entirely without AI.
In this training, participants will learn how to use any documents and texts (e.g., interview transcripts or manuals) as data sources thanks to Retrieval Augmented Generation (RAG) to answer natural language questions. Additionally, participants will learn how to structure the AI pipeline and source documents to ensure the RAG approach delivers the most accurate results possible. The limitations of RAG and suitable alternatives such as knowledge graphs and LLM wikis will also be examined with their respective advantages and disadvantages.
This training deepens prompt engineering practices. It teaches how to use JSON schemas to force specific output structures to be reproducible and machine-readable. Additionally, it demonstrates how to formulate metaprompts and prompts that check themselves for quality.
In a dialogue with humans, an AI can reach its full potential by understanding what users want and responding in an easy-to-understand manner. However, "out-of-the-box" AIs are trained to follow any path a user might take. Creating a guided communication path, or one that scales according to the user's abilities and prior experiences, is not easy. This training reveals how to design conversations with AI so that they are guided toward a specific user goal.
Station 3
You want to independently develop AI-powered products and processes and make them available to your colleagues - from the agent harness to the automated AI toolchain to the architecture.
Participants will learn how to derive user needs from usage scenarios and to verify them in the field successfully and convincingly according to the INVEST principle. Additionally, the training demonstrates how verified user needs are converted into robust user requirements in the form of agilely plannable user stories. It will also be shown how user stories can be split into self-explanatory and independently developable units, thereby reducing project risks and prioritization efforts.
Intensive course with content from the focus areas of UX, Scrum, and AI, including: Automatic generation of user stories and epics based on user research results; brainstorming and ideation with virtual user personas; use of Retrieval Augmented Generation (RAG) methods to increase understandability and provide automatic summaries of existing sprint and milestone planning, and prompt engineering methods for the quality assurance of existing user requirements.
Jeff Bezos once said: "Amazon must be an API-first company!" and implemented this vision more than successfully. But what is an API anyway, and how do you design or develop an API within your own corporate context? This training provides answers and is aimed at participants who want to make their solutions maximally usable for all colleagues, including software engineers and low-coders.
Although GenAI suites like OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini, or others are becoming increasingly powerful, no single model ultimately solves the underlying problem reliably, cost-effectively, in compliance with data protection regulations, and deterministically. As always, it depends on the combination of the right tools at the right time; otherwise, only fragmented results or random products are produced that do not contribute to a common goal. The training teaches how to link complex AI toolchains along the development process for various problem statements using a low-code environment, enabling them to solve complex problems reproducibly and make them accessible to other colleagues or AI agents.
Agents prefer to communicate via MCP servers. But what exactly is an MCP server and how does it work at its core? Why can they solve complex problems across application boundaries and how do you control their workflows and orchestration through clever handoffs and boundaries of responsibility? Answers to these questions will be provided in this deep-dive training.
Every department has hundreds of different use cases that can be partially automated or qualitatively improved. However, many of these use cases require advanced skills in prompt engineering and low-code automation. This training is therefore aimed at participants who have already developed these skills and now wish to apply them effectively within their own departmental context.
We also assemble the human-centered AI learning path as an in-house programme - tailored to your team's roles, tools and maturity. From a single workshop to a multi-month curriculum.
So let's explore the possibilities for your digital product development together. We look forward to your inquiry.
Thank you! Your submission has been received!
