R&D: Learn Faster, Experiment Faster
The great thing about AI for R&D, for example, is that you can experiment much faster with technology you've never used before. Take Python. I had almost no experience with that, but suddenly had to do a lot with it last year. This used to mean weeks of familiarization: taking courses, reading documentation, slowly finding your way around...
Now I'm asking an AI model: “I'm an experienced .net developer, but I've never done Python before. Make me a one-hour training with the most important differences.” That works perfectly. As a developer, you know how concepts work, but the syntax of a new language suddenly becomes much more accessible.
Presales: from google marathon to sparring with deep research
AI saves me the most time preparing for customer meetings. I used to spend hours Googling when we went to a customer in a new industry where we didn't have much experience: reading through all reports, collecting trends, summarizing that into something useful. Now I'm asking for a research model: “Act like a consultant and give me the latest trends in this market segment over the past five years.” After half an hour, I have an extensive report about what's going on in that market. Then, “Help me create a structure for a pitch deck based on these trends and this customer challenge.”
And then there's the fun part. When my presentation is ready, I give it to AI with the question: “Take a critical look at this and ask five difficult questions.” Or: “Will I achieve my goal with this presentation, or can I do it better?” I don't have an economic background, so I can just ask, “Are there things I'm overlooking? You are a buyer: what criticism would you have?”
Sometimes useful feedback comes out, sometimes not. But it takes me five minutes to ask, I'll get a cup of coffee, come back with input that I can do something with or not. That risk is minimal.
From idea to working prototype
With customers, you see that AI mainly helps to make things visible. It used to be difficult to get something out of your head on paper. A customer has an idea, has to explain it to me, I have to interpret it and a lot is lost along the way. Now you can have mockups and prototypes generated very quickly. There are tools where you type in a question and you immediately get a working screen. It is often still proof of concept, but you can create something you can talk about right away. I can even ask AI to chart a process. I used to have to sit down and draw myself after a meeting, back to my notes. That took an extra 2-3 hours. Now, during the conversation, I already have the diagram ready. “Is this what you mean? Do we see risks here?”
This visual component makes communication with customers much better. Instead of abstract conversations about what we're going to make, you look at something concrete together.
Not everything is going well with AI
Not everything is going well with AI. The biggest problem is that models sometimes hallucinate. Then content comes out that is incorrect or the model simply does not answer your question. Especially when it comes to automated processes, you have to be careful. If you have a process of 20 steps and things go wrong in step 2, but there is no person in the loop, you have a big problem in step 20.
The speed of development is also sometimes tiring. Every two weeks, a new model with new best practices, new prompt templates. Do I need to do something about that again? It's not like you're getting a new version of Word that you can calmly evaluate. And let's be honest: AI isn't the silver bullet that solves all problems. Some things are just better done manually. It helps you in the process, but it does not replace you.
Experiment or fall behind
So it's a bad idea to let everything fall out of your hands and outsource it to AI. On the other hand, if you're not doing anything with AI after 4 years of ChatGPT, you've got a problem. You have to experiment. Look around you! What do other organizations do? What do they use AI for? Make sure you have a ChatGPT subscription, that's really the minimum, and, above all, experiment with research. I'm amazed how few people use it, even though it's so powerful. You get really good output about market trends, technical developments, whatever.
Want to know more?
Want to learn more about how AI is changing developers' work? At Blis Digital, we use AI to learn faster, build better, and make smarter decisions without losing expertise.
Read here how we do this.
