4 ms·

The day-to-day is basically a combination of: * Moving data around - mostly using Excel + SQL, though R is growing in popularity * Using that data to calculat

The day-to-day is basically a combination of: * Moving data around - mostly using Excel + SQL, though R is growing in popularity * Using that data to calculate figures that are significant to the business - including premiums, IRRs, profit margins, and the value of future claims * Presenting the results of those calculations to management and recommending a course of action I'm not sure what "old-school original data scientists" do/did or how similar this is.

sizzle 8y ago
How is this not automated by some type of QuickBooks/Intuit type bookkeeping plug and chug numbers software? Startup opportunity?
- tfehring 8y ago
  At least at my company, the vast majority of bullets 1 and 2 is automated already. Financial information is calculated and fed to the financial statements automatically every reporting period, all of the recurring analysis supporting those results is generated with one button click, etc. The manual parts of 1 and 2 are centered around ad-hoc analysis when results aren't as expected, updating our automated systems when we add new products or product features, refining and adding granularity to those calculations as needed, etc. But actuaries' main value-adds are (1) building and pricing new products and product features and (2) helping management understand the financial results and their business implications. If you're just re-pricing a commodity product like term life or home/auto insurance, a lot of the pricing can be (and is) automated. But that's where companies have started building out more complex ML-based pricing, and actuaries are heavily involved in building and steering that functionality and explaining the results. If you can come up with a compelling way to automate a significant portion of that work while still providing results that insurers' CFOs and Chief Actuaries will sign off on, sign me up. But in general, if you're looking for automation opportunities in insurance, there's far more low-hanging fruit on the operations side than in actuarial. Same goes for finance and accounting, though to a lesser extent than operations.
- zv 8y ago
  Look at something like Tower Watson. I would say high barrier of entry.

sizzle 8y ago

How is this not automated by some type of QuickBooks/Intuit type bookkeeping plug and chug numbers software? Startup opportunity?

tfehring 8y ago
At least at my company, the vast majority of bullets 1 and 2 is automated already. Financial information is calculated and fed to the financial statements automatically every reporting period, all of the recurring analysis supporting those results is generated with one button click, etc. The manual parts of 1 and 2 are centered around ad-hoc analysis when results aren't as expected, updating our automated systems when we add new products or product features, refining and adding granularity to those calculations as needed, etc. But actuaries' main value-adds are (1) building and pricing new products and product features and (2) helping management understand the financial results and their business implications. If you're just re-pricing a commodity product like term life or home/auto insurance, a lot of the pricing can be (and is) automated. But that's where companies have started building out more complex ML-based pricing, and actuaries are heavily involved in building and steering that functionality and explaining the results. If you can come up with a compelling way to automate a significant portion of that work while still providing results that insurers' CFOs and Chief Actuaries will sign off on, sign me up. But in general, if you're looking for automation opportunities in insurance, there's far more low-hanging fruit on the operations side than in actuarial. Same goes for finance and accounting, though to a lesser extent than operations.
zv 8y ago
Look at something like Tower Watson. I would say high barrier of entry.