Telco Role in AI or Data Monetization Seems Limited, Really

One learns over time to be skeptical about some claims repeatedly made by leaders in many industries. Consider the claim by retail telco executives that they possess all sorts of behavioral data that can be monetized. What, exactly, are those sorts of data?

Some skeptics might point out that the data is mostly about use of communication services and devices; as well as service plan preferences. Telcos sometimes claim to have demographic data, but most of that is indirect. Location data is possible, but many app providers seem to be able to generate that themselves. 

Compare that to what a Meta or Alphabet might know: browsing histories; search queries; app usage; clicks; location; social graph. 

Feature	Telco Behavioral Data	Facebook/Alphabet Behavioral Data
Data Types	Service usage, support, demographics, purchases, devices	Browsing history, search queries, app usage, clicks, social interactions, location, interests
Granularity	Often aggregated or anonymized for privacy	Highly granular, often linked to individual profiles
Personalization Potential	Limited, especially with anonymization	Very high, enables highly targeted ads and content
Use Cases	Service improvement, pricing, churn reduction, marketing	Targeted advertising, content recommendation, user profiling
Regulatory Constraints	Strict privacy regulations, especially for sensitive data	Privacy regulations, but often more flexible with consent
Data Utility	Good for trends, limited for individual insights	Excellent for both trend and individual-level insights

Skeptics might argue that telco data is rather limited as a source of value or monetization. But telco execs keep insisting what they have is valuable. Some of us would say we haven’t seen it. 

Telco data might be viewed as substantial for internal operational and strategic planning, but it is generally less powerful than the behavioral data available to digital platforms for marketing and user engagement purposes. 

By analyzing support interactions, complaint logs, and feedback, telcos might be able to identify recurring issues such as network reliability problems, billing confusion, or service dissatisfaction. Also, analytics can detect subtle patterns such as a drop in usage, frequent complaints, or late payments that can signal a customer is at risk of churn.

But even there, the analytics might only be used proactively for business accounts. I see little evidence telcos use that data to do something about potential consumer account churn. 

In some ways, the issue is similar to the potential value of artificial intelligence, where telcos are going to be users of AI, but probably not in any particularly advantaged position where it comes to being a supplier of AI products and services.