yannick-design.de – yannick-design.de

CV

yannick-design.de — Fri, 18 Jul 2025 18:36:12 +0000

Chat

yannick-design.de — Mon, 14 Jul 2025 19:59:59 +0000

Will lead you to my GPT

David Oelssner

yannick-design.de — Fri, 04 Jul 2025 22:07:53 +0000

Mit Yannick konnte ich meine berufliche Rolle klar positionieren, Skills schärfen und mein fachliches Wissen ausbauen.

Working with Yannick helped me clearly define my professional role, sharpen my skills, and expand my knowledge in his areas of expertise.

A Step-by-Step Guide to Gen AI Workflows 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 12:50:28 +0000

A Step-by-Step Guide to Gen AI Workflows

Watch the full video here

Quick background about me:
I recently tackled a common challenge in our organization – creating comprehensive surveys that actually generate actionable insights. What traditionally took days of expert work can now be accomplished in 20-30 minutes with a well-designed Gen AI workflow, I’d say.

In this post, I’ll walk you through how to build effective AI workflows step by step, using my survey creation workflow as a practical example. Whether you’re new to AI workflows or looking to improve your approach – these principles will help you design processes that deliver consistent, high-quality results.

Watch the full video here

Why Build AI Workflows?

Before diving into the „how,“ let’s talk about the „why.“ This is actually easier than you might think. A well-designed AI workflow can actually save significant time on repetitive tasks, ensures consistency across outputs, and above all makes expertise more accessible to your entire team.

What this means is, it breaks complex tasks into manageable steps and creates reusable processes that improve over time. In my survey example, we transformed what was once a specialized skill – requiring deep knowledge of survey design principles – into a streamlined process anyone on the team can follow.

Step 1:
Interview an Expert

The first step in creating any effective workflow is understanding what experts currently do to accomplish the task. This has to do with the fact that you need to capture their thought process, not just their output.

Here’s a list of excellent questions to ask an expert – I personally use these relatively frequently:

What exactly is the goal of the workflow?
What does the result/target artifact(s) need to include?
What information did you need in order to start the task?
What was your thought process from beginning to end?
How did you evaluate the quality of your sources, data and intermediary steps?

Also, let me say upfront: Even a quick 15-30 minute conversation with a knowledgeable colleague can save you hours of frustration later and help align your workflow with what’s actually needed.

For my survey workflow, I spoke with our research team to understand their process for designing surveys that yield meaningful insights. This conversation revealed crucial requirements: surveys need to be engaging but brief, questions must connect to clear hypotheses, and above all – analysis needs to be planned before deployment.

Step 2:
Define the Goal

Every effective workflow needs a concrete goal focused on creating a specific artifact that solves a problem. This means you can’t just say „help teams understand their customers better“ – that’s too abstract.

For the survey workflow, I defined two clear artifacts as goals:

A finished survey with all sections, questions, and explanatory text
An analysis and interpretation guide to help make sense of the collected data

Notice how these are tangible outputs rather than abstract goals. This is particularly important because it makes it easier to evaluate success and iterate on your workflow.

Step 3:
Craft Your Workflow

With your goal defined, break down the process into logical steps. Think about what information is needed at each stage (inputs), what is produced at each stage (outputs), and what tools will be used.

For my survey workflow, I identified these key steps:

Define the research question
Expand the main question into sub-questions
Draft hypotheses for each sub-question
Select the most relevant hypotheses to test
Draft survey questions based on selected hypotheses
Generate the complete survey with all sections
Critique the survey and improve it
Create an analysis and interpretation guide

Each step has clear inputs and outputs. What this means is, the „Expand Research Question“ step takes the main research question and research context as inputs and produces expanded sub-questions as output.

Step 4:
Design Your Prompts

This is where the magic happens, I’d say. Well-crafted prompts are above all the building blocks of effective AI workflows.

An effective prompt should follow this structure:

Introduction: Tell the AI its role and main task
Context Section: Provide necessary background information
Detailing Section: Reinforce and detail the task after providing context
Closing: Include language constraints and final guidance

Let’s examine the prompt from the „Expand Research Question“ step:

You are a Research Assistant and need to expand the Main Research Question into Sub-Questions that help explore smaller nuances of the initial question.

Context of Research:
„““{{Research Context}}“““

Main Research Question:
„““{{Research Question}}“““

Make a List starting with the Main Research Question and generated Sub-Questions. Ensure to mark each clearly.

Make it {{Language}}. Let’s Think Step by Step:

This prompt defines the AI’s role as a Research Assistant, provides context through variables that can be easily replaced, gives clear instructions on the expected output format, and ends with language specification and the „Let’s think step by step“ technique to improve quality.

I personally recommend: Always use the largest models available first – for LLMs this means Reasoning Models like Claude 3.5 Sonnet or GPT-4. These models „think“ before answering and are able to produce more complex and more reliable output in one single step than other types of LLMs.

Let’s examine the prompt from the „Expand Research Question“ step:

You are a Research Assistant and need to expand the Main Research Question into Sub-Questions that help explore smaller nuances of the initial question.

Context of Research:
„““{{Research Context}}“““

Main Research Question:
„““{{Research Question}}“““

Make a List starting with the Main Research Question and generated Sub-Questions. Ensure to mark each clearly.

Make it {{Language}}. Let’s Think Step by Step:

Step 5:
Choose a Test Case

Before implementing your entire workflow, select a realistic test case. This is actually more important than you might think.

For my survey workflow, I chose to create an employee happiness survey for our digital agency, which has a high proportion of senior roles and few junior positions. I also made a little documentation of all available questions and fields in Microsoft Forms to use this documentation as input for the survey drafting step.

This specific context gave me the constraints needed to properly test the workflow.

Step 6:
Start Testing

When testing your workflow, follow these principles:

Test early and focus on one prompt at a time
Iterate on each prompt until you’re satisfied before moving to the next
Identify and fix issues immediately rather than testing the entire workflow at once

This has particularly to do with the fact that nothing is more frustrating than a workflow that doesn’t work and you have no idea why it doesn’t work.

In my testing, I discovered that the prompt for drafting hypotheses needed refinement to generate better connections to the research questions. By fixing this single prompt before moving on, I avoided compounding errors further down the workflow.

Step 7:
Document Steps with Clear Input-Output Relationships

For each step in your workflow, clearly document:

Description: A brief explanation of what the step accomplishes
Team or Person responsible: Who executes this step
Time: Estimated duration to complete
Inputs: What information is needed
Outputs: What this step produces
Blockers & Bottlenecks: Potential challenges
Tools: AI models or platforms used
Prompt: The exact prompt text with variables clearly marked
Cost: Resource requirements (time, subscription fees, etc.)

This documentation makes your workflow transparent, shareable, and improvable. What this means is, team members can understand how the pieces fit together and can actually contribute to improving the process.

Step 8:
Refine Based on Results

Once you’ve tested individual prompts, run through the complete workflow and evaluate the final output against your goal.

For my survey workflow, I compared the generated survey with professional surveys our research team had created previously. I looked for question clarity, logical flow between sections, and above all whether the analysis guide would produce actionable insights.

Step 9:
Organize for Future Use

Once your workflow delivers satisfactory results, organize it for easy access and future use:

Keep your documentation clean and updated
Mark which steps have been tested successfully
Consider taking breaks between workflow development sessions
Resist the urge to automate too early – ensure the workflow is stable first

Let me say upfront: If you keep your board organized and clean, you will be able to keep going way faster when you come back to it.

Common Challenges
and Solutions

Challenge: Prompts aren’t producing consistent results
Solution: Add more structure to your prompts and be specific about the format you want. Use variables for customization while keeping the core instruction consistent.

Challenge: Workflow steps don’t connect smoothly
Solution: Ensure outputs from one step match the expected inputs for the next. What this means is, sometimes you’ll need to add intermediate steps or refine prompts.

Challenge: The final output doesn’t meet quality standards
Solution: Work backward to identify which step is introducing issues. Sometimes adding a critique step can help improve quality.

Real-World Impact

The Survey Workflow in Action

Let’s see how this all works in practice with our survey workflow example:

We begin with a clear research question: „How happy are people working at the company?“

This expands into sub-questions covering areas like work environment, project allocation, leadership, career development, and more. Hypotheses are drafted for each area, such as „Senior employees feel their specialized skills are underutilized.“

The most relevant hypotheses are selected based on the research goals, survey questions are drafted to test these hypotheses, and a complete survey is generated with proper sections, including demographics, overall happiness assessment, workload questions, leadership evaluations, etc.

The result: A comprehensive employee happiness survey that would typically take days to create but now takes under 30 minutes.

Ready to build your own AI workflow?

Start small:

Choose a repetitive task you know well
Talk to others who perform this task regularly
Define a clear artifact as your goal
Map out the logical steps
Create prompts for each step, testing each step with a specific example
Refine until satisfied
Document for future use

Remember that workflow design is a skill that improves with practice. Some people are really good at workflow design, but with any task – some people like it and some don’t. It is perfectly fine if you struggle with workflow design. You don’t have to construct workflows – this skill is not inherently more valuable than any other skill

Conclusion

Building effective Gen AI workflows isn’t about complex technical skills – it’s about clearly defining goals, breaking down processes into logical steps, and writing prompts that guide AI to produce what you need.

What this means is, by following the steps outlined in this post, you can transform time-intensive tasks – like creating comprehensive surveys – into efficient, repeatable processes. Your workflows might start simple, but they’ll grow in sophistication as you learn what works.

Remember to:

Start with expert input
Define concrete goals
Break processes into discrete steps
Test early and often
Document thoroughly & organize for reuse

Above all:
Resist the urge to automate too quickly. Try to resist automating the workflow as much as possible – naturally it’s going to change over time and with automation there comes a whole layer of complexity that just makes it harder.

What task will you transform with your first Gen AI workflow? If you have ideas, problems, fears, thoughts, or use cases you want to solve concerning the use of GenAI tools, please feel free to reach out.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

Implementing AI without methodology is like still fiddling with VHS tapes in the Netflix era 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:55:52 +0000

A Guide for Implementing AI with Methodology

Implementing AI Without Methodology is like Still Tinkering with VHS Tapes in Netflix Times

Look, I’ll be straight with you here – if you’re thinking about implementing generative AI in your company without a proper methodology, you’re basically still trying to figure out VHS players while everyone else is streaming on Netflix. That’s just not going to work in today’s world, and I’ll tell you why.

Why This Actually Matters

So, generative AI has been absolutely exploding since early 2023, right? And I’ve been working with companies – a lot of them actually – helping them figure out how to use these LLMs and generative AI systems effectively. What I’ve noticed is that the companies that just jump in without thinking it through, they’re the ones that struggle the most.

The thing is, and this is something I always tell my clients, knowledge work has always been really hard to scale. You can’t just throw more people at a problem and expect it to work better – actually, that often makes things worse. I’ve seen this happen so many times in agile software development, where adding more team members just creates more communication overhead and actually slows things down.

The Data is Pretty Clear on This

Now, before you think I’m just talking theory here, let me share some actual numbers with you. There are these studies – and I’m talking about proper MIT research here – that show LLMs can reduce the time needed for tasks by about 27% while improving quality by 45%. That’s not just incremental improvement, that’s transformational.

What I find fascinating is that across different studies, you consistently see these patterns: knowledge workers become 10-15% more effective overall, they complete tasks 25-30% faster, and the quality of their work improves by 40-45%. That’s the kind of improvement that can completely change how a business operates.

Why Consulting Models
Are Critical

Here’s the thing – and this is something I’ve learned from working with dozens of companies – you need proper frameworks to make sense of complex stuff like AI implementation. I personally use what I call consulting models because they help create a common language within organizations.

Without these models, you get everyone talking past each other. The tech team is talking about LLMs and RAG systems, the business folks are worried about ROI, and management is trying to figure out what any of this means for their strategy. It’s a mess, frankly.

My Approach

Four Models That Actually Work

So I’ve developed this framework – actually four different models – that help companies navigate AI implementation systematically. Let me break this down for you:

Level 1: Identifying Measures

Strategy Model: This helps you visualize different layers of your AI strategy. I use this to show companies how their AI initiatives connect to their overall business goals.
Maturity Model: This is where you assess where your company actually stands with AI adoption. Are you just starting out? Do you have pockets of excellence but no coordination? This model helps you figure that out.

Level 2: Evaluating Measures

Evaluation Model: This brings together the classic desirability, viability, and feasibility framework but applies it specifically to AI initiatives. It’s a visual way to assess which AI projects are actually worth pursuing.
Stage-Gate Model: This helps you understand how mature your AI rollout is – from initial idea all the way to scaled implementation across the organization.

What Makes This Different

The thing is, these models don’t just work for AI implementation. I’ve used similar frameworks for digital transformation, innovation management, data-driven work processes, even sustainability initiatives. The key is having a structured but flexible approach that can adapt to different contexts.

I always tell my clients – and this is something I believe strongly – that successful transformation requires both courage and reflection. You can’t just copy what worked for another company; you need to understand your specific context and challenges.

Why Most Companies
Get This Wrong

Here’s what I see happening all the time: companies get excited about AI, they start a pilot project, maybe they get some initial results, and then they try to scale it without really thinking through the implications.

They don’t consider things like:

How does this fit with our existing processes?
What training do our people need?
How do we measure success?
What are the risks we need to manage?

That’s where having a proper methodology becomes absolutely essential. Without it, you’re just hoping things work out, which is not a business strategy.

The Reality Check You Need

Look, I get it – there’s a lot of hype around AI right now. But the data is pretty clear that this technology is going to transform how knowledge work gets done. The question isn’t whether AI will impact your business, it’s whether you’ll be ready for it.

Companies that approach AI implementation methodically, with proper frameworks and clear processes, they’re the ones that are going to succeed. The ones that just wing it? They’re going to struggle.

What I'd Recommend

If you’re serious about implementing AI in your organization, here’s what I’d suggest:

Start with a proper assessment of where you are now
Define clear objectives for what you want to achieve
Use structured frameworks to evaluate and implement initiatives
Focus on building capabilities, not just deploying tools
Measure and iterate based on real results

The companies I work with that follow this approach consistently see better outcomes. They avoid the common pitfalls, they get buy-in from stakeholders, and they build sustainable AI capabilities rather than just implementing point solutions.

The Bottom Line

Implementing AI without proper methodology is like trying to navigate without a map. Sure, you might eventually get somewhere, but it’s going to take longer, cost more, and you’ll probably end up somewhere you didn’t intend to go.

The companies that are going to thrive in the AI era are the ones that approach it strategically, methodically, and with proper frameworks. That’s not just my opinion – that’s what I’ve seen work in practice across dozens of implementations.

So the question is: are you ready to do this right, or are you still going to be tinkering with VHS tapes while your competitors are streaming in 4K?

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

This strategy model makes AI management a breeze 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:55:04 +0000

A Strategy Model Makes AI Management a Breeze

So here’s the thing – I’ve been working with companies on AI adoption for a while now, and I keep seeing the same pattern. Everyone wants to jump into AI, but nobody really knows where to start or how to actually manage it properly within their organization.

That’s why I want to talk about this strategy model today. It’s basically a framework that helps you think systematically about AI implementation – and trust me, you’re going to need this kind of structure.

Quick Background
on Why This Matters

The thing is, most companies I work with initially think about AI in completely the wrong way. They immediately jump to „Oh, we need a chatbot on our website“ or „Let’s build some AI feature for our customers.“ But actually, that’s putting the cart before the horse.

What I’ve learned – and this is something I always tell my clients – is that AI adoption needs to be tied directly to your business goals and organizational structure. You can’t just throw AI at random problems and hope something sticks.

How It Actually Works

So let me walk you through this model. It’s built on two axes, each with two areas:

Value Creation
vs. Value Proposition

Value Creation (Internal Stuff) vs. Value Proposition (What You Offer Customers)

This distinction is super important, and here’s why: Companies often think exclusively about AI in the context of new website features or products. But meanwhile, their internal knowledge work processes are running incredibly inefficiently and slowly.

The differentiation between internal employee use cases and external customer-facing applications is crucial because you need completely different approaches for implementation. These two extremes are related, sure – a changed portfolio also means changed internal processes – but they come with totally different argumentation strategies and implementation considerations.

Operations
vs. Management

Operations (Process & Implementation) vs. Management (Decision & Direction)

This helps you differentiate between strategic, directional decisions and the actual implementation of those decisions into concrete results. In many companies I consult with, these levels get mixed up, which can lead to unintended changes in company direction that can quickly put the business in economic trouble.

The Four Quadrants

From these two axes, we get four quadrants:

Organizational Management (Value Creation x Management)
This is about deriving AI measures for IT infrastructure and organizational structure based on your existing organizational vision. We’re talking about changed hierarchies, adjusting business units, building new teams or roles, and changes to IT architecture and hardware.
Internal Process Implementation (Value Creation x Operations)
This focuses on translating strategic decisions into AI tool selection and AI measures to change and improve existing processes or build new ones. This includes targeted pilot projects, employee workshops, procuring AI tools, training programs for staff, and similar initiatives.
Business Model Management (Value Proposition x Management)
This deals with AI measures that involve validating or establishing new or changed business models derived from your business vision. Think competitive analyses, business ventures, testing new offering formats, and much more.
Portfolio Implementation (Value Proposition x Operations)
This is about deriving AI measures to implement your business model and strategy, integrating necessary AI offerings, features, or services firmly into your portfolio to improve or underpin your market position.

My Recommendation:
Start Internal

Here’s what I typically recommend to most companies I work with – and this might surprise you – start with internal AI applications.

Why? – Well, first off, using AI in internal processes carries fewer risks because the applications are tested by a limited user group and you get quick feedback. This enables precise use cases, clear boundaries, and comparison possibilities.

But more importantly, companies and employees gain deep intuitive understanding of the capabilities, challenges, and limitations of current AI models and tools through direct use. Plus, the impact on the business is easier to measure with internal activities – whether in terms of speed, quality, or cost reduction.

In contrast, the impact of AI features in external products is often hard to assess and requires high initial investments. That’s why I advocate for focusing on the „Internal Process Implementation“ quadrant in most cases.

The Bottom Line

Look, this model isn’t just some theoretical framework. It’s a practical tool that helps you avoid the typical mistakes I see companies make when they’re trying to figure out AI adoption.

The key thing is that it forces you to think systematically about where AI fits in your organization and what you’re actually trying to achieve. Too many companies just start experimenting without any clear direction, and then they wonder why their AI initiatives don’t deliver results.

By using this model, you can focus your resources, measure your impact more effectively, and actually build sustainable AI capabilities that align with your business goals. And that’s exactly where AI management belongs – not as some separate initiative, but as an integral part of your business strategy.

That’s the approach I take with all my clients, and it’s consistently delivered better results than the scattered, unfocused approach most companies start with.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

This maturity model is your cheat sheet for Gen AI adoption 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:54:19 +0000

A GenAI Maturity Model

The Cheat Sheet to Not Fall Behind in AI Adoption

Let me give you a quick background on this – I’ve been working with companies on GenAI adoption for quite some time now, and what I’ve realized is that most organizations actually struggle with the same fundamental challenge: they don’t know where they stand in their AI journey, and more importantly, they don’t know what steps to take next.

The maturity model I’m presenting here is basically our roadmap – or let’s say, a cheat sheet – to help companies navigate this complex process of GenAI adoption without losing track of where they are and where they need to go.

Why You Actually
Need This Model

Now, here’s the thing – and this is something I see repeatedly in my work with clients – companies often think that just implementing new AI tools is enough. That’s actually completely wrong, especially when we’re talking about GenAI.

The reason is relatively simple: GenAI tools are characterized by their flexibility and individuality. That means you can’t just roll them out company-wide and expect everyone to magically become productive. You need to slowly introduce employees to working with GenAI tools, giving them sufficient space to understand the basic concepts and experiment with the technology.

There are some parallels here to when Google emerged as a search engine or at the beginning of the internet – people needed time to develop what I call „digital literacy.“

The Core Concept

Understanding Where People Are

When you look at GenAI adoption from this perspective, it becomes pretty clear that you need to take a more differentiated approach to how individuals within a company engage with this new topic. The measures you implement should differ based on how much someone has already engaged with GenAI.

This understanding and previous experience with the thinking patterns that characterize GenAI – that’s what we divide into maturity levels. And let me tell you, this is actually the key to successful AI transformation.

5 Levels

This GenAI maturity model is divided into 5 levels, and each level is described in terms of cultural characteristics and impact on business processes – because these two levels are what fundamentally make up companies.

Level 1: Interested
This is where most companies start, to be honest. The company and employees show interest in GenAI. Employees and/or management inform themselves about the impacts of GenAI on various industries, get competitive analyses, assessments, or feedback, and learn about potential GenAI measures.

Level 2: Piloted
Now we’re getting into the action phase. The implementation of GenAI experiments and pilot projects within established business processes by a task force is underway. The measurability of GenAI impact in use cases is actively promoted to better assess deployment possibilities of GenAI tools and to set up investment planning for further rollout.

Level 3: Applied
At this level, GenAI tools are used for various everyday business processes. Comparisons of different processes and different GenAI tools have become standard. There are cross-functional responsible persons with significant decision-making authority who are known and recognized.

Level 4: Embedded
The integration of GenAI into core business strategies is complete, and all strategic decisions consider GenAI. It’s ensured that GenAI is considered in all (new) projects and workflows.

Level 5: Transformational
GenAI continuously drives the transformation of business processes, the alignment of directional decisions, and strategic planning. Customized GenAI tools and GenAI measures significantly improve customer satisfaction and company performance.

Investment
vs. ROI

Here’s something that’s really important to understand – and this is where I see many companies get confused: I distinguish between two overarching phases within the model: Investment and ROI.

The first three levels typically see higher investments than savings. But from Level 3 onward, when GenAI is deeply integrated into the company structure, the funds generated through savings and revenue increases exceed the expenses for GenAI measures.

This requires forward-thinking leadership that recognizes and utilizes AI’s potential – and that’s not always easy to find, I have to say.

Setting the Right Maturity Goal

Now, here’s where strategy comes in. Besides locating the company within the maturity model, it’s crucial to set a maturity goal – meaning which maturity level a company wants to reach long-term, ideally combined with a time-based goal.

In many companies, reaching Level 5 won’t be desirable because identifying GenAI as the primary driver for company development in all areas doesn’t have to make sense. The maturity goal depends on how high the GenAI maturity of the competition is and how strong an influence GenAI has on the company’s business model in the short or long term.

I personally think most companies should aim for a maturity goal of Level 3-4. That’s usually the sweet spot where you get maximum benefit without over-investing in areas that don’t drive your core business.

How to Assess Your Current Level

The assessment of a company’s maturity level on a topic like GenAI can’t be done through pure facts alone – you need to survey employees because of the cultural dimension of GenAI adoption.

This makes the assessment both simple, because it can be standardized, but also time-intensive, especially in large companies.

We generally distinguish between questions for managers and questions for employees. Both questionnaires consist of topic areas with standardized questions and answer options that each enable evaluation and classification into the GenAI maturity model.

Manager Assessment

For leadership, we focus on four main areas:

GenAI Potential & Usage:
How familiar is your department with generative AI potentials? To what extent are generative AI tools and technologies used in your company?

GenAI Management:
How do you measure the success and impact of generative AI in your company? What resources are available for AI projects?

GenAI Strategy:
Do you have a strategy for deploying and scaling generative AI technologies? How does your company plan to adapt to future developments and innovations in generative AI?

GenAI Rollout:
How is knowledge about generative AI spread in your company? To what extent do your legal and technical guidelines support agile and broad application of generative AI in the company?

Employee Assessment

For employees, we look at three core areas:

GenAI in Your Work:
Your understanding of generative AI, use of AI tools in daily work, your involvement in AI projects, and the influence of AI on your own role.

GenAI Training:
Your perception of AI & data protection in your company, training and education on generative AI, knowledge dissemination and education in your company about AI.

GenAI in Your Company:
Your knowledge of interdisciplinary collaboration in the AI context, your knowledge of your company’s readiness for disruptive AI changes, your knowledge of future orientation and adaptability to AI progress.

The Bottom Line

Look, the key takeaway here is that GenAI adoption is not just about technology – it’s fundamentally about organizational change and cultural transformation. The maturity model gives us a framework to navigate this change systematically rather than just hoping for the best.

What I’ve learned from working with companies at different maturity levels is that the ones that succeed are those that take a structured approach, set realistic goals, and invest in both technology and people development simultaneously.

The future belongs to organizations that can effectively integrate GenAI into their operations while maintaining their human-centered values and strategic focus. This model is your roadmap to get there.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

Without this evaluation model, you’re flying blind with Gen AI 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:53:39 +0000

An Evaluation Model for GenAI Initiatives

Without this Evaluation Model, You're Flying Blind When Choosing Gen-AI Measures

So, let me give you a short background on what you’re dealing with here. My evaluation model – and I think this is actually quite important – represents a methodical approach to assess whether specific measures and pilot projects in the Generative AI space make sense or not. What I’m trying to do here is find a balance between, let’s say, emerging ideas for using Generative AI, the current capabilities of the models, the business impact of these ideas, and – thisis crucial – how well they fit culturally within the organization.

Through my model, I can quickly identify which measures should get priority and which ones – and this is where it gets interesting – need to be put on the back burner for now. And more importantly, I explain the reasons behind these decisions.

A Tool for Selecting
Gen-AI Measures

Look, I’ll be honest with you – in a world that’s characterized by complexity and, let’s face it, complete chaos, the tangible adoption of generative AI in companies remains a challenge. That’s just the reality I’m dealing with.

But here’s the thing:
keeping track of everything and aligning all factors toward a common course requires wisdom and foresight. That’s why I’ve reduced the various perspectives to the three well-known aspects of „Desirability,“ „Feasibility,“ and „Viability.“ This reduction – and I think this is really smart – should enable clearer discourse about the suitability of specific measures.

My goal is to create deep understanding for the organization. And what’s absolutely essential here is that technical feasibility, symbolized by the intelligence of the models, is just as important as business relevance and employee support. Only when these three levels are united in a measure does it lead to success.

A Concrete Example

In many companies, ideas that come from high-status individuals or were copied from competitors get implemented. But this approach can be misleading because – and this is something I always tell my clients – companies are unique, and an idea might not fit existing processes.

Consider a company that sells a complex technical product about which there’s little publicly available information and no internal documents or resources exist. In this case, developing intelligent chatbots isn’t a sensible measure because the technical feasibility simply isn’t there. No LLM can meaningfully advise on specific topics without extensive fine-tuning, and customers often don’t have the necessary background knowledge to use such tools effectively.

Here's Where It Gets Interesting

If you flip this consideration around, it shows which measures can improve company performance in the short term. Is the product consultation difficult and the sales team overloaded while marketing lacks detailed knowledge? Then targeted Gen-AI applications can help. Internal chatbots could transcribe meetings and make knowledge accessible. Fact-checker tools could review marketing content and point out technical accuracy issues. Tools for simplifying technical jargon could serve as learning aids for employees.

These measures directly address the company’s bottlenecks, require little to no model expertise, and support employees in their daily work. They’re probably faster to implement, will be more easily accepted by employees, and thus lead to positive economic effects for the company more quickly.

The Evaluation Model in Detail

The evaluation model consists of three overlapping circles – what I call a Venn diagram. These three circles symbolically represent the three aspects: Desirability, Viability, and Feasibility. Areas where these three aspects overlap naturally lead to adoption. So measures that fulfill all three aspects simultaneously will have a very high success rate.

Areas where only two of the three circles overlap represent a sub-optimal compromise. This is where adoption difficulties occur with measures.

The 3 Aspects

Feasibility
This aspect deals with the implementability of specific use cases. I’m talking about questions like: „Are there models that can cover the work steps?“, „Are there ready-made solutions for this, or can you build the infrastructure?“, or „Is it legally permitted to implement this idea?“ Feasibility primarily relates to content-related questions that become important during solution implementation. If this aspect isn’t fulfilled, implementation becomes lengthy, difficult, or even impossible.

Desirability
This aspect focuses on the softer, more cultural parts of measure evaluation and puts people’s needs at the center. It’s about questions like: „Do your customers want to use this and does it solve their problems?“, „Do your employees want to use this and does it make their work easier?“, or „Is it ethically and morally acceptable if you implement this idea?“ If this aspect isn’t fulfilled, implementation will encounter push-back in the company, be ignored by end customers, or even be publicly criticized.

Viability
This aspect puts economic viability at the center and deals with the benefit of measures. It’s primarily about having an economically positive effect for the company overall. I’m talking about questions like: „Can the operational costs be sustained?“, „How much does this change the Customer Lifetime Value?“, „Is the use case significant enough compared to other processes?“, or „Does it solve one of your growth problems?“ If this aspect isn’t fulfilled, implementation won’t „survive“ long after completion or will be abandoned during implementation.

My personal Take

The evaluation model – and I think this is actually quite profound – is a tool for reflection and self-awareness that critically questions existing measures for Gen-AI adoption. Companies must have the courage to rethink these approaches and expand their capacity for action. This enables not only conscious and efficient resource allocation but also focusing on the really effective generative AI measures.

Only through this differentiated yet simple perspective can you master the challenges of digital transformation and look toward an inspiring future. It’s time to become active and shape change pragmatically and solution-oriented.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

Stage-Gate Model: Keep Your Gen AI Initiatives on Track 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:53:03 +0000

The Stage-Gate Model

Keeping Gen-AI Initiatives in Focus

Look, the Stage-Gate Model doesn’t just offer a structured approach for evaluating AI initiatives and their progress – it’s actually, I’d say, a philosophical way of looking at how growth happens in companies. What I mean is, by mapping out the journey from that first spark of an idea toa scalable initiative, it basically forces us to question traditional approaches in a critical way.

Instead of jumping into initiatives with their full complexity right away – which, let’s be honest, most companies tend to do – this approach enables a step-by-step introduction. What this does, essentially, is inspire companies to tackle change in a positive and pragmatic way, while keeping that long-term vision of successful, company-wide rollout always in sight. I personally think leadership teams should actively participate in this process and question the status quo to drive innovative solutions forward. You can find an overview of all models here.

The Stage-Gate Model helps us evaluate the progress of any initiative and have meaningfulconversations about the next steps in the rollout process

A Tool for Targeted Initiative Development

In the realm of generative AI, many companies are still at the beginning of their journey, if I’m being honest.

This first phase is characterized by uncertainties and, quite frankly, a lack of proven methods. This forces companies – and I mean really forces them – to question traditional approaches and explore new paths. To meet this change with clarity and structure, I’ve developed the Stage-Gate Model. A model that, through methodical procedures, minimizes the risk of misallocating resources to unsuitable initiatives.

An initiative begins as a germinating idea. In the initial phase, it’s crucial to test its direction and feasibility through experiments and feedback loops. This could be a trial phase with generative AI tools or, for example, an employee survey to evaluate practical utility in daily work. Based on the insights gained from the Minimum Loveable Idea, the next step involves developing the idea into a full-fledged initiative with allocated budget. From this point on, comprehensive consideration of the idea and measurability of future experiment results move to the foreground.

This iterative process demands not only vision and inspiration but also pragmatic implementation of new technologies. It’s time for leaders to question the status quo and actively pursue the path of digital transformation. Only this way can they shape progress and secure a sustainable future.

An Example

Gen-AI offers a revolutionary opportunity:
With minimal risk, it relieves marketing resources and significantly increases performance. For most companies, this is a groundbreaking idea. Initial experiments with prompts and company content quickly show that current LLMs don’t just theoretically support companies – they practically reduce the time and resources for content creation drastically. The Problem-Solution-Fit is proven – now it’s about formalization.

When expanding this idea, new processes are defined, prompt templates and system prompts are created. This formalized approach is tested and optimized in numerous concrete use cases, leading to a measurable reduction in required resources and a significant increase in online visibility.

The next steps involve firmly integrating these new processes into the daily routine of multiple departments and proving that this new approach can become the standard. This often requires mentors and initial guidelines to transfer the pilot project team’s experiences to broader applications.

If this approach works in a broader target group, the last step can be tackled: Expansion to a scalable process whose rollout can be carried out with minimal resources. This typically includes new company guidelines, training sessions, video tutorials, FAQs, leadership training, and much more.

After completing the planning and incorporating costs into company expenses, the Stage-Gate Model can be left behind and the complete rollout can begin.

The Model in Detail

My Gen-AI Stage-Gate Model is divided into 5 steps.

Each step represents a milestone within the rollout of this initiative and is characterized by certain properties that serve as a checklist for completing this phase.

Stage I: Minimum Loveable Idea

Desirability: The initiative addresses a critical business problem that’s highly relevant to the company’s operations and values
Viability: The initiative opens clear business opportunities that contribute to the company’s long-term business goals
Feasibility: The initiative aligns with the company’s overall strategy and utilizes existing resources and capabilities available to the company

Stage II: Problem-Solution-Fit

Desirability: The initiative aims to solve a significant business problem that’s worth addressing and improves the speed and scalability of processes
Viability: A preliminary budget plan has been developed to ensure the economic feasibility of the initiative
Feasibility: The initiative is compatible with the company strategy and actively supports the company’s business goals

Stage III: Minimum Viable Measure

Desirability: A basic outline for the initiative has been created, best practices have been integrated into the approach, and the procedure is ready for initial broad application testing
Viability: Possible weaknesses and risks have been analyzed and corresponding strategies formulated
Feasibility: Necessary technologies and approaches have been conceptualized, internally tested, and show promising results. Top management supports further expansion of the initiative

Stage IV: Measure-Operations-Fit

Desirability: The initiative has evolved beyond the prototype stage, with positive response from users of the previous phase. Since the beginning, the initiative has developed in iterative steps
Viability: Broad tests support ROI projections and indicate positive results with long-term value
Feasibility: The company supports further rollout of the initiative and there’s collaborative cooperation between departments or external partners regarding the initiative. The rollout financial plan has been updated

Stage V: Minimum Scalable Measure

Desirability: A strategy for company-wide rollout has been defined
Viability: The rollout financial plan considers scaling costs and costs for all necessary training measures
Feasibility: Company plans include continuous definition of milestones, strategies, and management approaches for Gen-AI initiative expansion

Initiative Status
and Next Steps

Determining the stage of a Gen-AI initiative is, I have to say, remarkably simple. The steps can be viewed as a checklist for each stage. If all prerequisites for completing a step are met, the initiative progresses to the next stage.

However, if one or more aspects of the defined thresholds aren’t reached, the initiative remains in this step until all requirements are fulfilled. Only after reaching the next Stage-Gate can the next step begin.

The unfulfilled criteria up to the next Stage-Gate act as a checklist and guideline for the current work focus. Decisions for transitioning to the next step are always made at the Stage-Gate. The first decision for advancing a „Minimum Loveable Idea“ therefore only occurs when all defined criteria are met.

Retrospective evaluations of initiatives are also possible. However, the failure to meet earlier criteria should lead to immediate pause. In case of delayed evaluation, it’s essential to fulfill all previous expansion criteria of the respective Stage-Gates before further expansion and rollout is considered.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

The Future of the Web: The Impact of Generative AI 🇬🇧

yannick-design.de — Fri, 04 Jul 2025 11:51:54 +0000

The Future of Web

How Generative AI is Actually Reshaping Everything

Watch the hole webinar here

A quick background on me – I’ve been working with AI systems for quite some time now, and I have to say, what we’re seeing with generative AI is something that’s fundamentally changing how we think about web interaction. Let me break this down for you in a way that’s actually practical.

Long-term Changes

The AI Assistant Era is Coming – Whether We Like It or Not

So here’s the thing – and I say this based on what I’m seeing in the industry – AI assistants are going to become the primary interface for user interaction within the next three to five years. That’s not just speculation, that’s what the data is telling us.

I mean, think about it: 62 percent of users already prefer chatbots over traditional support. And here’s what’s really interesting – 71 percent of Gen Z users are using chatbots for product research. That’s not a small shift, that’s a fundamental change in user behavior.

What does this mean practically? Well, I think we’re looking at a situation where websites become essentially databases that AI systems query, rather than interactive user interfaces. Microsoft, Google, Apple – they’re all integrating these assistants deep into their devices, using both local and cloud-based models to provide personalized, contextual support.

The key thing here is that these aren’t just fancy chatbots anymore. We’re talking about sophisticated RAG systems – that’s Retrieval-Augmented Generation for those who aren’t familiar – that can actually understand context and deliver precise information.

Watch the hole webinar here

Medium-term Perspective

The Hybrid Era – And Why It’s More Complex Than You Think

Now, in the medium term – and this is where it gets interesting – I believe we’re going to see a hybrid usage pattern. Some users will still prefer traditional browsing, while others will rely primarily on AI assistants to get specific information.

Here’s what I’m seeing: user expectations are changing rapidly. They’re becoming more impatient and demanding immediate, precise answers. That means companies need to respond by personalizing their content more heavily and understanding their target audiences better.

But here’s the challenge – and this is something I tell my clients all the time – you need to make your content machine-readable and semantically rich. That’s not just about SEO anymore, that’s about being discoverable by AI systems. And yes, you can automate a lot of this with LLMs, but you still need the strategic thinking behind it.

The user intent becomes crucial here. We’re not just talking about keywords anymore, we’re talking about understanding the underlying motivations and contexts of user queries.

Short-term Actions

What You Should Actually Be Doing Right Now

Okay, so practically speaking – and this is where I always get concrete with my recommendations – companies should start implementing chatbots on their websites immediately. Not because it’s trendy, but because it gives you qualitative insights into user preferences and optimizes your content production.

Here’s what I recommend: Start with prototypes. Run design sprints. Test the functionality and value of AI assistants with your actual users. You need an interdisciplinary team – content experts, product specialists, support staff – because this isn’t just a tech project.

And here’s the thing that many companies struggle with: you need to be willing to give up control and let the systems operate independently. That requires courage and the acceptance that mistakes will be made, but those mistakes are valuable learning opportunities.

A Pragmatic Approach

From Theory to Implementation

Look, I’ve seen too many companies get stuck in the theoretical phase. What I always tell my clients is: start with comprehensive knowledge about your products and services, then structure that data for AI systems. This enables precise and contextually relevant answers.

The key is understanding that we’re not just building chatbots, we’re building intelligent systems that can actually understand and respond to complex user needs. That’s a fundamental shift in how we think about user experience.

My Take on the Future

What This Actually Means

Here’s my honest assessment: The future of the web lies in the harmonious integration of technology and human interaction. Generative AI isn’t just changing user behavior – it’s fundamentally altering how companies communicate with their customers.

But here’s what I think is most important: success lies in the ability to adapt and use technological possibilities creatively and thoughtfully. We’re not just implementing new tools, we’re reimagining the entire interaction paradigm.

As someone who’s been working in this space for a while, I can tell you that the companies that succeed will be those that embrace this change proactively rather than reactively. The technology is there, the user behavior is shifting – the question is whether businesses are ready to meet their customers where they’re going, not where they’ve been.

This analysis is based on current industry trends and practical implementation experience. The timeline and specific impacts may vary depending on adoption rates and technological developments.

Wanna talk about what that means for your company?

Let’s discuss how this proven framework can address your organization’s specific challenges and opportunities.

yannick-design.de – yannick-design.de

CV

Chat

David Oelssner

A Step-by-Step Guide to Gen AI Workflows​ 🇬🇧

A Step-by-Step Guide to Gen AI Workflows

Why Build AI Workflows?

Step 1: Interview an Expert

Step 2: Define the Goal

Step 3: Craft Your Workflow

Step 4: Design Your Prompts

Step 5: Choose a Test Case

Step 6: Start Testing

Step 7: Document Steps with Clear Input-Output Relationships

Step 8: Refine Based on Results

Step 9: Organize for Future Use

Common Challenges and Solutions

Real-World Impact

Ready to build your own AI workflow?

Conclusion

Wanna talk about what that means for your company?

More stuff to read

Implementing AI without methodology is like still fiddling with VHS tapes in the Netflix era 🇬🇧

A Guide for Implementing AI with Methodology

Why This Actually Matters

The Data is Pretty Clear on This

Why Consulting Models Are Critical

My Approach

What Makes This Different

Why Most Companies Get This Wrong

The Reality Check You Need

What I'd Recommend

The Bottom Line

Wanna talk about what that means for your company?

More stuff to read

This strategy model makes AI management a breeze 🇬🇧

A Strategy Model Makes AI Management a Breeze

Quick Background on Why This Matters

How It Actually Works

Value Creation vs. Value Proposition

Operations vs. Management

The Four Quadrants

My Recommendation: Start Internal

The Bottom Line

Wanna talk about what that means for your company?

More stuff to read

This maturity model is your cheat sheet for Gen AI adoption 🇬🇧

A GenAI Maturity Model

Why You Actually Need This Model

The Core Concept

5 Levels

Investment vs. ROI

Setting the Right Maturity Goal

How to Assess Your Current Level

Manager Assessment

Employee Assessment

The Bottom Line

Wanna talk about what that means for your company?

More stuff to read

Without this evaluation model, you’re flying blind with Gen AI 🇬🇧

An Evaluation Model for GenAI Initiatives

A Tool for Selecting Gen-AI Measures

A Concrete Example

Here's Where It Gets Interesting

The Evaluation Model in Detail

The 3 Aspects

My personal Take

Wanna talk about what that means for your company?

More stuff to read

Stage-Gate Model: Keep Your Gen AI Initiatives on Track 🇬🇧

The Stage-Gate Model

A Tool for Targeted Initiative Development

An Example

The Model in Detail

Initiative Status and Next Steps

Wanna talk about what that means for your company?

More stuff to read

The Future of the Web: The Impact of Generative AI 🇬🇧

The Future of Web

Long-term Changes

A Step-by-Step Guide to Gen AI Workflows 🇬🇧

Step 1:
Interview an Expert

Step 2:
Define the Goal

Step 3:
Craft Your Workflow

Step 4:
Design Your Prompts

Step 5:
Choose a Test Case

Step 6:
Start Testing

Step 7:
Document Steps with Clear Input-Output Relationships

Step 8:
Refine Based on Results

Step 9:
Organize for Future Use

Common Challenges
and Solutions

Why Consulting Models
Are Critical

Why Most Companies
Get This Wrong

Quick Background
on Why This Matters

Value Creation
vs. Value Proposition

Operations
vs. Management

My Recommendation:
Start Internal

Why You Actually
Need This Model

Investment
vs. ROI

A Tool for Selecting
Gen-AI Measures

Initiative Status
and Next Steps