I have sworn off starting another company. I don’t want to spend another 4 - 10 years working 14/7. But I still get ideas. So here’s one for any of you interested.
I just completed a 5 day class on the process utilities go through to set rates. It’s complex. And they kept saying parts are more art than science which means parts are judgement calls.
There’s a ton of input to this, at several stages. And there’s a very specific desired outcome. The rates should deliver to the utility a set profit, on the order of 9%, with every ratepayer paying their fair share.
And every intervener1 is trying to gain maximum advantage at the expense of others. For example:
The utility wants all expenses counted, and have calculations that underestimate income from rate payers, so they have a higher profit.
The Consumer Advocate wants higher rates for large commercial customers and therefore lower rates for homes and small businesses.
Large commercial companies want lower rates for large commercial customers and higher rates for homes and small businesses.
Environmental groups want zero carbon and are fine with higher rates for everyone to obtain that.
The list goes on…
I asked at the conference if there were any AI tools to help on this. I received a blanket no. And if there were, the instructors at this class would know as they’re heavily involved in setting rates.
There is an abundance of documentation available for each rate setting proceeding. From that you could have an AI work through each step and provide a strong argument for what should be set as the rate base, allocation into each rate payer bucket, rate tariffs, etc.
You also need to take as inputs the goal of the user. A utility company should get different numbers from the consumer advocate, the commission staff, etc. You also want as input whether it should generate very conservative numbers, fair numbers, or aggressive numbers.
The output should be not just the numbers and the calculations around them but also the legal wording for the documentation that intervener will present to the commission.
And this is critical - you need at least two experts on this process, either as partners or advisors. An individual who has a 5+ year track record of multiple times they were an expert witness in proceedings. And an individual with a 5+ year track record as an intervener. Best is an intervener who worked for either a commission or the consumer advocate.
Like so many start-ups, who ever creates a product that is good enough and gets a critical mass using it first - wins. Good enough is many interveners see it as providing them a significant advantage.
An Intervener is an organization accepted by the commission to provide testimony on a proceeding. A rate setting proceeding commonly has 15 - 30 interveners.
Ratemaking is a possible area for AI improvement.... though a challenge would be getting your expert testimony, both oral and written, to match up with the AI slop.
A better area of AI implementation is in the the interconnection study process. This is less art and more repetitive model simulation where AI can have a huge impact on a process that is supposed to take less than a year by commission approved tariff, but is easily taking 3-8 years.
According to most environmentalists, rates are set by "greedy" capitalists. ;)