BFSSSS
BF Strategic Site Selection Services

Bob's Blog

Introduction

In my non-professional travel and everyday life I often view things through the prism of a site selector. In this blog I share some observations and thoughts that might be of mild interest to the larger community of site selectors, economic developers, and corporate managers. I am not a frequent blogger (more often doing client work), and when I do so I may add a post with a link in LinkedIn. I use my twitter account, @BFSSSS, more as a rant about poor customer service or what seem to me to be non sequiturs. Actually, I have found twitter to sometimes yield quicker responses than via customer service kiosk queues or phone calls when those media are backed up or unavailable, for example after a flight cancellation, while overseas, or after one of several hurricane evacuations from my summer home on the Outer Banks.

A First Look at Using AI in Site Selection

Artificial Intelligence has been prominent in the news lately. Students are using it as an aide (if not to plagiarize) term papers and admission essays to the chagrin of educators. IBM is halting backoffice hiring while it determines which of an estimated 30% of backoffice jobs AI can replace. Hollywood screen writers are striking partially to obtain guarantees that AI won’t be used in their stead. Wall Street quants are using AI to convert plain speech into computer code which can be input into computer trading programs.  

As a novice in AI, I took ChatGPT AI, an iPad app, for a ride to see how it responded to some broad site selection issues:

·         I first asked it for the best locations for environmentally sustainable growth. Well, its answer was just about everywhere – Europe, North America, Asia, Australia, Latin America,

·         Well then what are the worst locations? ChatGPT returned a fairly good list of qualities that should be avoided – including pollution, political instability, and poor infrastructure – but did not tie them to any areas, countries, or locations.

·         Switching tact, I asked it to compare New York City and Washington, D.C. as a second headquarters location for Amazon – a site selection issue which had drawn considerable press a few years ago. Well ChatGPT had a lot of good things to say about each location that were relevant to the decisions, but drew no conclusion only going so far as to call each location fantastic. It had nothing to say about the negatives of each location missing the political and union opposition Amazon encountered in New York.

·         Well maybe I should be a little more specific, so I asked it if it would be easier to hire customer service representatives in El Paso or Jackson, MS. A list of the factors that would help answer that question was returned, and I was told to do my own research. OK then, can you tell me the starting salaries for customer service representatives are in each city? At last definitive answers, followed by a number of caveats. Interestingly the source used was Glassdoor’ which I had never used relying on the Bureau of Labor Statistics or commercial databases such as the Economic Research Institute.

·         I then wondered what role AI could play in reducing fieldwork workload where site selectors interview local employers about the business environment and competition.  GP Chat let me know there were more than 30 call centers in the area but that it had no ability to gain market insights on its own.

·         Lastly, I asked the specific question on what incentives the Reno area offers to attract investment in battery manufacturing. I was told I may want to consult the official website of the Reno government.

While my first foray into AI was mainly amusing, it and subsequent research (e.g., MIT Technology Review – AI is Coming to the Classroom Meet the teachers who believe it could improve education) it did offer some insights on how AI could be useful in site selection. For one.  corporate officers involved in site selection and economic developers might want to chat with AI before hiring expensive consultants to economically develop a better understanding of underlying issues and questions to be asked and discussed. In education it is being found that rather than as a vehicle to escape work, AI can give students some one on one face time that professors don’t have the time to give and help them develop their thought processes. Just as IBM believes AI will fill many back office roles, I think it will be used effectively in creating client presentations and in the Request for Information (RFI) process both for requesting site selectors and responding economic developers.

The programmers of ChatGPT AI, universally available to the masses, have seemed to have strived to make it politically correct and in my test it made no negative commentary, was not decisive, and was almost self-deprecating in its replies. This may a reaction to the current scrutiny of AI be the Senate and others for its ability when programmed by bad actors to effect elections and even mankind. I did not delve too deeply into a specific project for confidentiality reasons in this exercise, however, I believe experience and knowledge by the user will still be required by the user to deploy AI effectively in site selection. With savvy user AI will be a powerful tool in framing and refining discussions, researching data, and reporting results.

(Given enough interest I will add a link to the transcript of my chats with ChatGPT AI)

Bob Frederickson