Hi there,

Growing your business with AI can be hard to figure out - especially in the beginning.

To get you on the right track, I've created an AI Adoption Checklist based on my experience in the field.

By the end of this series, you will have a 12-point plan for successfully launching an AI project and making the most of its impact on your business.

In today's edition, we will cover the first 6 important steps of this guide.

Let's get started!

Why you need a checklist

Many AI projects fail. Depending on who you ask, estimates suggest that 70% up to 90% of all AI projects don’t show the expected results.

Having worked on numerous AI/ML projects over the past few years, I have noticed that each one fails for a different reason. AI is a tricky beast, and there are many things that can go wrong.

However, successful AI projects followed a similar pattern.

I have tried to capture these patterns in a simple checklist of "best practices".

The checklist has 12 points, and for the sake of brevity I will cover the first 6 in this newsletter (the remaining 6 will be covered next week).

So here's Part 1:

The Ultimate AI Adoption Checklist (Pt. 1)

[ ] List current business challenges

If your goal is to have business impact, then everything should start with the business. Write down some key challenges your company is currently facing. This can be super high-level (org chart) or more operational (department), depending on your role. Think about either improving existing processes or creating new product/service experiences.

Example: If you run an e-com store, a key challenge might be cart abandonment.

[ ] Prioritize challenges based on impact

Once you've listed all the challenges, rank them according to their impact on your business success. Using the e-com example, if cart abandonment results in significant lost sales, it should be a top priority.

At this point, prioritization is more important than actual monetary value - rough estimates are fine!

[ ] Select one significant process or product for AI intervention

Don’t try to boil the ocean. Instead, focus on one specific pain point where AI can make a meaningful difference.

Example: AI could help you analyze why shopping carts are abandoned, or suggest countermeasures to prevent it.

[ ] Attend a beginner's workshop or webinar on AI

You don't have to become an AI expert, but you should know some essential basics. A good way to do this is to take a beginner-friendly course or webinar –there are many great resources out there. Here are two great examples:

• AI for Everyone by Andrew Ng - Very basic, easy to digest course

• AI Bootcamp by Armand Ruiz - More hands-on, in-depth course

[ ] Read introductory AI articles or books

If you prefer to read, you can pick up a good textbook or articles for a quick introduction. Just make sure the book covers current trends such as GPT models. Ideally, you can find something that explains how AI has changed industries similar to yours. You can also browse through the archive of this newsletter for some inspiration.

[ ] Discuss with an AI expert

At some point, you should bring in an expert, even if it's just for a short time. An expert will help you understand how exactly AI can benefit your business, and help you assess the feasibility and impact of potential use cases. Of course, feel free to reach out any time.

[ ] Brainstorm non-AI alternatives

The best AI solution is the one you never needed. Every AI-powered tool, solution, or process adds a layer of complexity to what you're doing now – at least at first. So before you commit to AI, think of simpler solutions.

Example: If cart abandonment is the issue, perhaps a straightforward email reminder to customers might be enough.

[ ] Compare non-AI vs. AI solutions

Once you have a list of potential solutions, weigh the pros and cons of each. Consider factors like implementation time, cost, required resources, and potential benefits.

Example: You may find that sending an automated card abandonment email to every customer would result in a flood of emails and opt-outs. Instead, you want to send it only to customers with a high likelihood of buying - something you can use AI for.

[ ] Define clear success metrics

To gauge the success of your AI project, you need measurable outcomes. Define metrics that align with your business objectives. These could be quantitative, like a percentage increase in sales, or qualitative, like improved customer or employee satisfaction.

Example: If you implement an AI tool for marketing analytics, a potential metric could be a 10% increase in lead conversions within six months.

[ ] Predict short-term and long-term benefits

Start small, think big. Consider both the immediate gains (such as cost savings) and the potential long-term benefits of the AI solution (such as business growth or competitive advantage in your niche).

Example: The marketing analytics tool could lead to higher lead conversion rates short term, but also provide valuable insights to better understand your customers long term.

[ ] Validate estimates with other people

We often fall in love with our own solutions. That's why it's always good to cross-check your assumptions with other people who are not so involved in the process. Consider building a small committee of experts from different departments, or at least get a second set of eyes to challenge your estimates.

[ ] Consider initial vs. ongoing costs

Every AI project has both upfront and ongoing costs. Upfront costs include things like purchasing software, acquiring hardware, and initial training. Ongoing costs, on the other hand, can include maintenance, software updates, ongoing training, and potential scalability requirements.

It's important to budget for both to ensure the sustainability and effectiveness of the project over time.

[ ] Make or Buy

The entire field of AI is maturing rapidly. Many things you had to build in-house at great expense are now available as off-the-shelf services. Depending on your use case, make a deliberate decision whether you build, buy, or adopt a hybrid solution.

Example: You can use ready-made AI services for basic tasks like extracting text from documents. But for more specific tasks like predicting customer purchase probabilities, you need a customized AI model trained on your own data.

[ ] Gather data requirements

Not every AI use case requires a lot of high-quality data to get started.

For example, if your first AI use case is about creating better marketing content, you only need basic information about your target audience, their problems, and your product. This data is usually available in customer personas or other customer research materials.

However, if your use case involves building a predictive AI model for a complicated demand forecasting situation, you will need a large amount of high-quality data from both internal and external sources.

[ ] Ensure data privacy and compliance standards are met

You may have the data, but can you use it? This is especially critical when dealing with personally identifiable information (PII), and when operating in areas with privacy laws such as GDPR. Make sure your use case addresses these concerns, and that you have the necessary permissions.

[ ] Assess data quality

Data quality means fit for purpose. So check if your data is ready for your use case along the following dimensions:

• Veracity - is your data correct?

• Volume - do you have enough data?

• Velocity - how quickly is your data updating?

• Variety - does your data represent enough sub-groups or samples?

Example: If you want to build an AI-powered lead scoring system, you may find that most leads are actually low quality and there are not enough "high quality lead" samples in your training data, in which case you would need to collect more of them.

Conclusion

This concludes Part 1 of the Ultimate AI Adoption Checklist. I hope you found it useful. If so, then please leave a feedback! It would mean so much to me! ❤️

Stay tuned for next week, where I’ll dive in the second part of it.

In the meantime, feel free to the download the full checklist here:

See you next week!

Tobias

Want to grow your business with AI? Here are 3 ways I can help:

1. Book a meeting: Let's find out how I can help you over a coffee chat (use code FREEFLOW to book the call for free).

2. Read my book: Improve your AI/ML skills and apply them to real-world use cases with AI-Powered Business Intelligence (O'Reilly).

3. Follow me: I regularly share free content on LinkedIn and X.