Your Data Is Broken... But Sure, Let’s ‘Just Add AI’

Read time: 2.5 minutes

The 5 Recommended Statements for ‘Just Add AI’

1. Present The Data Basic Facts First

The first thing before talking about AI is to realize the reality of what kind of data you’ll have to provide:

• No values

• No duplicates

• No different data types

• No dates that are out of date

Without having a good understanding of the data you’re going to work with, all your discussion of AI is just speculation.

2. Ask Yourself What You Are Going To Do With This Decision.

Rather than discussing what features you want, simply ask yourself: "What will this improve my decision-making process?"

If no one gives you a great answer to that question, then you’re still not ready to consider adding AI because the request is not yet mature.

3. Use the Rule of Clean First.

When communicating to your audience, simply communicate: Clean Data First, Intelligence Second.

No model will exceed the design of the underlying framework.

4. Convert the Hype to Requirements.

When someone says "add AI," they usually mean:

• Faster answers

• Better predictions

• Fewer manual analyses

Turn those vague comments into concrete technical requirements because that’s the starting point for doing the real work associated with the AI project.

5. Simply Let Data Speak For Itself.

If you want to illustrate your point, simply run a quick data-quality test. Then present your findings.

There are a few things that will help reset unrealistic expectations faster than having actual data-quality issues that can be measured.

💡Key Takeaway: 

AI doesn't make bad data accurate. It produces more accurate results faster and with greater confidence.