Real AI Use Cases in Finance & Accounting
Here are 5 real (and useful) AI use-cases for finance and accounting folks.
Today’s Sponsor: Tropic
Q4 is almost here. And so are the renewals.
Over 30% of SaaS contracts renew in Q4, right in the middle of budgeting season. Tropic’s free Slack, Gong, and HubSpot playbooks give you insider benchmarks, sharp negotiation tips, and contract gotchas from our procurement pros. Two more guides (1Password and ZoomInfo) are on the way. One sign-up gets them all. We’ve already helped customers save $56M in H1. Grab your playbooks and breathe easier this Q4.
Using AI in Finance & Accounting
I keep hearing finance folks say there are no “real” or “useful” AI use-cases for finance/accounting yet. In today’s post I am going to prove those folks wrong with a few that my team is doing right now….
AI isn’t going to do everything for you (yet), but there is plenty you can do right now with just ChatGPT (or other similar models).
1. Drafting Policies & Memos
NOBODY likes writing the initial accounting policy/memo draft. Often 80%+ of these are just to make the documentation look nice, cite resources for auditors, etc. The remaining 20% is the actual detail of how the accounting should work.
My mind is still blown by how good ChatGPT did on a first draft for an “Internal-Use Software Capitalization” policy. Here is my prompt:
I read through the first draft (which was already quite good) and then asked for some fine-tuning of the policy and some additional things it should consider. The first time might not be perfect so just fine-tune and provide additional relevant context.
And then boom! ChatGPT spit out a policy that was 80%+ of the way there. I can then easily take that draft, add in some company-specific context/data and it will be great.
Here is a link to the full draft that ChatGPT created. I didn’t change anything…
Now apply this to every policy, memo and updates you have to make and you have some really meaningful time savings!
2. Build Useful Stuff
There are A LOT of “simple” things that you used to need to go to the engineering/IT teams for, but now finance folks can just do themselves. Play with SQL, python scripts, Excel macros, building small apps, etc.
But…do NOT try to build complex stuff. It’s almost never worth a finance/accounting person’s time and they usually lack the skills.
In less than 1 hour I built a web app that you can share with candidates/employees to help them think through the potential value of their stock options! It’s not super complicated (just need to know the logic) so even someone like me can “vibe code” it.
»Check it out and feel free to share it with your people!
3. Create Excel Templates and Perform Research
IPO Prep - Comparable Company Analysis
Below I asked ChatGPT to build a table that shows key metrics of comparable companies that recently went public so you can compare to your company and the metrics you need to IPO.
It built all the summary data into one tab with key metrics to compare to your company. Also it had more detail (including all earlier years presented on the S-1).
I did have to re-prompt it a couple of times because it screwed a couple of things up but it still saved a lot of time. Obviously you should still check it because it can hallucinate, but it’s getting REALLY good pulling financial information.
Prepaid Reconciliation Workpaper
AI can also do a pretty incredible job building Excel month-end close workbooks with the right prompting.
Let’s say accounting didn’t have a prepaid amortization workbook. I can just ask ChatGPT and it can spin one up with formulas based on my policy and everything.
It won’t be perfect, but it does a PRETTY good draft. It created 5 tabs with instructions, formulas, an automatic summary, and JE upload template!
Or maybe you have a decent starting point, but you need help with an Excel formula to capture the amortization properly. ChatGPT is VERY good at that!
4. Analyzing Data
There are lots of low-hanging fruit in data analysis that AI can handle — trend reviews, anomalies, out-of-policy checks, data matching, etc. Some of this may be integrated in tools like spend management tools, but it’s also easy to dump into AI and ask questions.
I asked ChatGPT to create a standard T&E policy and then review all my T&E transactions…
What did AI find for me?
It found all three potential fraud transactions that I added to the data:
There was a transaction for an iPhone purchase that was clearly out of policy
Someone bought a gift card which is not allowed
Flagged an unusually high hotel booking
I also asked ChatGPT to review the past 12 months of data to find the people who are most frequently out of policy and submit reports late. I then got a detailed report of who submitted reports late and by how much and the number of out of policy transactions.
Once you feed AI your policy and transaction data then it is VERY good at spotting issues.
5. Variance Explanations (BvA & Accounting)
If you have organized data (big IF) then you can use AI to find significant month over month changes in your accounts or where significant differences are between budget versus actual.
If you are dealing with a lot of data then these variance explanations can take a lot of time every month.
A lot of accounting and FP&A tools are building this directly into their platform so you don’t have to take the data out to get the AI variance analysis.
Security / Privacy with AI
A lot of folks (especially in finance) are concerned about privacy given they work with confidential data. Rightfully so…
Here is what to do:
NEVER put company or confidential information in your personal ChatGPT (or other model)
Ask what your IT/security team has approved and how you can use it (they should have guidelines)
If you don’t have a policy yet, scrub the data (or use dummy data) so there is nothing confidential and then play with it in AI
If your IT/security team still hasn’t put together a policy then you might need a new IT/security team….Everyone should have access to AI today.
Footnotes:
Get the playbook on how to save on popular vendors like: Slack, Hubspot, 1Password, ZoomInfo, and Gong (from Tropic)
I recently partnered with one of the best in-house tech lawyers I know to launch OnlyLawyer. Subscribe and send to your legal team!
The limitation I see is that companies do not let employees upload real company data to AI tools due to confidentiality. This limits a lot what workers could do with this tools…