Gen AI - Applications for an Actuarial Entrepreneur
- Logenthiran Kanisan
- Jun 20
- 4 min read

As the dust settles, the hype and fluff around Gen AI becomes much more visible, and we now also see a little more clearly where this technology adds value for us. Being an actuarial entrepreneur, there are basically a few key questions to answer, as with any tool:
Can Gen AI generate more revenue for me?
Can Gen AI help to reduce my costs?
Can Gen AI make me more efficient?
An important disclaimer, which you are probably well aware of - be careful of inputting any confidential information into any Gen AI tool that is not a proprietary/enterprise tool.
Can Generative AI generate more revenue for me?
Clearly the largest consulting firms are doubling down into the Gen AI segment, building agentic AI, tools, services etc to extract as much revenue as they can from this vertical and its hype. They are pouring lots of investments into them, but it remains to be seen if they will get a decent ROI from it.
However, the reality for a small boutique consultant (who will still be a slightly larger boutique consultant after scaling for many years), is that our clients come to us because of our personal touch and service. Business comes to us from word of mouth, from friends who have worked with us, from our ex-bosses, from people who trust us. The human(s), the soul(s), the personal magic is irreplaceable.
Yes there are other boutique consultants who have built specialised software, RPA (Robotic Process Automation) tools, and soon (or already have?) agentic AI tools as well, but the reality is that we are consultants and not software builders. It takes a very different skillset and knowledge to build robust & production-ready software and tools, complete with the support services required.
We can also expect that clients will want to pay significantly less for subscription to an AI tool as opposed to a pure consulting service, so the market size needs to be significantly large to reap the rewards targeted.
There are verticals outside of the traditional actuarial practice areas where clients are not willing or able to pay the usual fees required for actuarial services. A software based & automated solution can enable us to enter those segments - a Gen AI enabled chatbot backed by an actuarial computation engine can potentially allow us to do that.
Can Gen AI help me to reduce costs?
This is a “polite, entrepreneurial” way of basically asking can I use Gen AI instead of hiring junior team members. My opinion is a resounding NO for the reasons I list below. This is contrary to what we hear about out there with large consultants and firms replacing staff or reducing hiring due to Gen AI.
Maybe it’s due to my own philosophy & world view, maybe it’s due to the profession and industry segment I am in.
Why is it a NO for me:
a. Gen AI has no consciousness
Excuse me for leaning into my mystical side here - a person is a person, because of that spark of life, the spirit within/pervading. That cannot be replaced by a soul-less machine.
I am building a team, mentoring and enabling my team to grow and thrive, and have joy in that process. And of course build my own financial freedom and personal freedom while doing that.
Our clients come to us because of the touch of that spirit, hence I have to nurture and grow it.
b. Wide diversity of practice areas and methods
Despite our small size, we operate in a wide area of domains, and gradually in different geographies as well. Typically automation can be built for specific repeatable tasks, with Gen AI providing flexibility to understand natural language.
It makes a lot more sense to invest in training and developing juniors who are able to work on these wide domains of practice, gradually being able to handle client interactions as well.
c. The importance of Actuarial Judgment
Just earlier today, a senior actuary mentioned in a conversation that Actuarial Science is more of an art rather than a science. Yes it is built on a foundation of mathematics and statistics, but the interpretations and applications are really up to the skills, creativity and personality of the wielder of the craft.
As a boutique consultant, I’m building a guild of artisans, not a production machine.
Can Gen AI make me more efficient?
Here it is a resounding YES, no doubt about it.
These are the simple, yet powerful applications I have been repeatedly using Gen AI for:
a. Coding & de-bugging
We must always be the master of our tools, and not let our tools be our masters.
Coming from a non-programming background, I often struggle with coding. I know what I want to get done but spend hours looking up the syntax to do that and after that spend more hours debugging syntax errors.
Gen AI is an amazing tool for this purpose undoubtedly. The learning curve to code has become incredibly shorter, so the focus is now on understanding and knowing what to do, and implementing it is a lot easier.
b. Desktop Research
Desktop research has become a lot more quicker and efficient, compared to the days of Google searches alone.
No doubt checking the sources is imperative & the tendency of Gen AI to hallucinate is something to be wary of.
This is an extremely important benefit for a boutique consultant.
c. Taking minutes
Taking minutes was a rather annoying task, that takes away our focus from paying attention to the process of recording. The numerous tools that we have to do this now are very useful of keeping track.
Again close checking of the outputs are needed due to the tendency to hallucinate, but checking work is a lot easier than doing it ourselves.
d. Image recognition of tables
The most frustrating form of data storage are scanned documents with tables of values, making it impossible to easily copy them out.
Gen AI tools have been so useful for this purpose now, enabling us to generate usable tables from images. Again please take care to not put in confidential values in non-proprietary/non-enterprise tools.
My current view on Gen AI
Clearly I’m more of a skeptic rather than a staunch believer, but certain use cases and benefits of Gen AI are undeniable.
Comments