The non-emergency call market

I believe I have stumbled upon a growing, untapped market shaped by labor trends and emerging technologies: the market of non-emergency calls directed to government. Here, I share my thoughts and a possible business opportunity sprouting from it. I am looking for advice on whether or not to pursue it and corrections on the logic I am about to show you.

Every year, citizens call the government millions of times for non-emergency situations. Situations that don't require an emergency response. Those calls start from all sorts of phones and end in government call-centers, who solve the situation by providing information, dispatching units, or adding a service request to a city department. Here is the problem: call-centers are losing people. The US is currently transforming in such a way, that call-center staff is becoming rarer and more specialized. And those who survive the environment are becoming a proud few. Already, bot assistants are being used to respond to non-emergency calls. In an effort to reduce the burden on staff. 

The quality of voice bots just went through a breakthrough. Never has the technology been so developed and so human like. Deployment is trivial. And companies are competing for who can provide the best bot call-center experience. Add to that, we could be looking at millions of minutes of non-emergency conversation. Bots seem to be on route to further expand and answer more non-emergency calls. The market might be big enough to sustain a community of developers, who look to serve at least two clients: citizens and government. Citizens looking for city services and information. And government looking to respond to citizen petitions (using the most diverse software). Each one with it's own challenges. My question is, is it wise to develop non-emergency bot technology and community? Is it worth it?  

And here is a breakdown of the possible size of the market. I try to provide two extremes. 

-Annually, around 240 million calls are made to 911. Of these, I'll assume (rather conservatively) that 40% to 60% are non-emergency, translating to approximately 96 to 144 million non-emergency calls each year. 

-311 is smaller, and has not been strictly implemented nationally. Low usage areas can experience 0.2 calls per person per year. And high usage systems like NYC311 can experience up to 4 calls per person per year. I'll assume that 66 to 660 million 311 non-emergency calls happen annually.

-Combining 911 and 311 estimates, the total number of non-emergency calls ranges from 162 to 804 million annually.

-I'll assume the average call duration for both 911 and 311 is between 2 to 5 minutes.

-Resulting in the large range of 324 to 4,020 million minutes of non-emergency calls every year.

-Modern voice bot development ranges from $0.08 to $0.3 per minute. This is not considering your own infrastructure. Nor operational expenses.

The market could be large. The retiring workforce is leaving a large blank, already being filled by bots. It seems like the US will trend towards non-emergency bots.