How to get started with AI in 2025

Welcome back to another year on the AI rollercoaster ??

2024 was a rollercoaster—we tested, learned, and built so much. Now, I couldn’t be more excited to scale things up in 2025. ??

This year, I’m making a few changes. Writing newsletters takes a lot of time (and, let’s be honest, LinkedIn is leaning more towards a Facebook-style feed these days). So, we’ve decided to move our educational AI newsletter to Connect AI.

Here on my personal LinkedIn blog, I’ll shift focus to more less frequent technical deep dives—perfect for when I fall into the rabbit hole on specific AI topics.

I’ll keep you in the loop at the start, but if you’re into the content, be sure to subscribe to our newsletter (link in the footer). That way, you won’t miss a single thing.

Now let's get started ??

The buzz around AI can be overwhelming. From ghost writer to chatbots, it feels like there’s an AI solution for everything. But for many businesses, diving headfirst into the vast ocean of AI can seem scary. Here's the good news: you don’t have to do everything at once.?

Starting small and focused can deliver quick wins without unnecessary risks.

How can you get your AI projects started?

These steps will change the way you look at AI and help you understand that AI isn’t about doing everything—it’s about doing something smarter.

1. Start simple

Starting your AI journey doesn’t mean tackling the hardest challenges. Begin with low-hanging fruits—those repetitive, time-consuming tasks that slow your team down.

Examples of simple AI implementations:

• Customer service automation:?Use AI chatbots to handle FAQs, like password resets or return policies.
• Data entry and sorting:?Employ AI tools to process forms or classify emails.
• Content summarization:?Save time by using AI to extract key insights from lengthy reports.

By automating these tasks, you free up your team to focus on high-value work. These small successes build confidence and create momentum for adopting AI further.

2. Focus on non-sensitive areas

When starting out, it’s wise to focus on areas that don't involve sensitive or critical data. These carry higher risks in terms of compliance, security, and potential errors.

Why start with non-sensitive tasks?

• It minimizes regulatory concerns. Think EU AI Act and other regulations.
• It avoids steep costs for advanced security features.
• It allows teams to experiment with fewer repercussions.

3. Test, learn, refine

Adopting AI isn’t a one-and-done process. It’s about testing solutions, learning from outcomes, and refining your approach. Think of it like conducting small experiments.

How to implement the test and learn approach effectively:

1. Make it concrete:?What task are you solving exactly? Are you reducing response times, improving accuracy, or cutting costs?
2. Build Feedback Loops:?Analyse how customers or employees interact. This helps you to understand areas of improvement and excel on that. Use key performance indicators (KPIs) to assess impact. For example, track customer satisfaction (CSAT) after deploying an AI chatbot.
3. Iterate:?If something doesn’t work as expected, adjust or focus on another area. AI thrives on continuous improvement.

It's time to get started!

How can you identify the best starting point?

Not all tasks are created equal when it comes to AI automation. To choose the right place to begin, ask yourself:

... continue reading on our Connect AI Blog

Great start into 2025

Matthias