Bringing in Outside Data for AI Models

Bringing in external data to help artificial intelligence (AI) and machine learning (ML) applications is among the first topics client executives want to talk about when scoping consulting projects. It then becomes a topic which ISVs have to deal with very quickly when marketing and selling an AI module or application to their customers.

I consider the question quite reasonable.

You cannot dismiss the questions as unnecessary even if the client doesn’t understand the AI and ML processes. Often, the client will have a couple of extra data sources in mind. If not, I usually suggest a few examples. Industry trade association projections or general economic growth and inflation are common examples. Sometimes clients want to use forecasts they get from the suppliers. These and others are all potentially useful.

(Side note – rarely do the external factors change model predictions or increase the accuracy to any substantial degree.)

As an ISV developing an AI and ML module or incorporating those functions, you should seriously consider the ability to include external data. However, unless you are charging ongoing maintenance to cover pulling in the data sources, make the client maintain the connections to external data. Otherwise, your automated data pulls will break eventually as the external source changes its format or location.

Most of the time for AI and ML, possessing the ability to use external data constitutes having a few extra data tables which the client will need to populate repetitively. Don’t be surprised if they stop updating their input data tables regularly.

Having the AI and ML models use the external data sources, requires a little more validation code to make sure the data is good during schedule model refreshes. That is not hard but does require foresight.

Designing to include external data source options to the minimally viable product early makes customizing for individual clients later easier as well.

Talking about the ability to incorporate external data is a marketing plus too.

Creating a standard response for salespeople to address this common question is needed early in commercializing your AI and ML product or strategy. For you, the ISV, this is probably the most significant benefit to using external data – it helps you sell the AI and ML module or application. It can be your competitive advantage.

We help ISVs and systems integrators create a minimally viable AI and ML strategy and product, which should incorporate external data sources outside of the standard application databases. Creating the needed marketing and sales support materials to answer this common question is what we do exceptionally well too. Give us a call if you would like to know more.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s