Using AI on My Own Data

By now you have probably tinkered around with some AI chats and maybe even thought it was intriguing. Have you wondered how you can ask it about your business data? You can train AI with your data and the process works similarly with households, clubs or any group that has information about itself.

There 3 big steps prepare your data for AI:

1. Identify – What data do you want to ask questions of?

2. Organize – How do you access and make sense of that data?

3. Clean – Make sure everything is right and safe.

Quick note about privacy first (and last, when we anonymize the data). Never put any secrets or personally identifiable information in this data. Don’t load your data into any untrusted AI provider, it’s best to use your own AI.

This time we are talking about Identifying your data, the next 2 steps will be in upcoming articles.

You probably have an idea of what kind of questions you want to start with. Let’s use Operational Data as an example.  You might have this data in several places:

1. Standard Operating Procedure Docs – Most businesses have folders full of how-to’s and system manuals, new-hire handbooks are great too.

2. Tracking spreadsheets – Maybe you have used spreadsheets to track internal projects, tools, assets or even service calls

3. Inventory – Most inventory systems will let you export the data or have other access to that data

4. Supply chain data – Vendor information, how you deal with them and what they provide

5. Sales – Your point of sale system has mountains of information hiding in it.

6. Correspondence – Email might present too many issues, but other forms of communication, like message boards and chats can be used

7. CRM system – Keeping track of your clients and what you do for them, if it has a ticketing system there will be excellent operational data

8. Human Resources system – people and their job titles, be sure to leave the sensitive info behind.

Any internal system you use on a regular basis is a candidate. Some systems might take more effort in the later stages if they are hard to access or a bit messy.

While internal data is essential, external data can significantly enhance your AI project. Explore possibilities like:

1. Industry benchmarks - Compare your performance to industry standards.

2. Economic indicators - Understand market trends and fluctuations.

3. Weather data - Analyze how weather affects your business.

4. Social media data - Gain insights into customer sentiment and behavior.

5. Newsfeeds – News that could impact your industry

6. Third-party data providers - Access specialized datasets for specific needs.

So much data… Don’t panic! We are just identifying right now. You might find that some of these sources lead to other sources of data, some might be split and some might be combined. This discovery process can be enlightening and you will probably find some issues with accessing the data and how clean the data is. Keep track of those issues and get ready for the Organizing and Cleaning.

Tech Tip: It’s a good practice to export your data from online services periodically, that way you are protected from service outages, account deletions, and service cancelations. You can also be ready to use a better service if one comes along.

Jim Hundley

ASCENDalyst.com - 727-346-6020