How Generative AI can help with Monthly Reporting

This data was originally featured in the October 4, 2023 newsletter found here: https://www.trustinsights.ai/blog/2023/10/inbox-insights-october-4-2023-technology-and-connections-generative-ai-for-tasks/.

It’s monthly reporting time, and monthly reporting is one of the greatest opportunities for us to find efficiencies and use Generative AI to expedite the process. How long does monthly reporting take you?

For me, doing reporting for our clients takes about a day, sometimes two days to process all the data and turn them into reports. Every month, I’m always on the lookout for ways to save additional time, to shave time off the process. Why? The less time I spend making the report, the more time I can spend thinking about the report, its implications, and what I can recommend for our clients.

Here’s an example of a simple report optimization. We have a series of templates we use for each month’s reports. In those templates, there are placeholders for the month, and typically I copy the report from one month to the next to preserve some historical data, plus any upgrades I’ve made over time.

Yet for dozens of websites and projects, even this can be a bit laborious. Finding the instances of the previous month and updating them to the next month takes time – not a ton of time per report, but it adds up. I could have someone else do that part, but that’s not a good use of their time either. No, the best choice here is to automate this process – it’s important work, but it’s low value work.

In the days before generative AI, I would have spent a few hours writing this code. Today, I simply log into ChatGPT, fire up the GPT-4 model (the current best model for general coding), and ask it how to find and replace text in a directory full of PowerPoint presentations:

Boom. The Python script is ready to use. I copy and paste it to my computer, and now I’ve saved about 12 hours per year. Is 12 hours a year a huge deal? No, not really. In the grand scheme of things, it’s relatively small. It’s not going to transform the company and make us instantly more profitable…

… by itself. Here’s the thing about process optimization with AI. It’s not sexy. You’re not going to win any awards for its innovative use. But as with all incremental things, it adds up. Shave 15 minutes here, 30 minutes there, and the savings add up quickly, to the point where suddenly you’re looking at free hours, not free minutes, in a day.

When we first started Trust Insights five years ago, the reporting process took a couple of full days for reports that weren’t super insightful at the time. They were significantly better than what other agencies were producing, to be sure, but looking back on them is like looking back on your Halloween costumes as a kid. They’re cute, you might cringe a little, and you know that what you can do today is so much better than what it was like a little while ago.

You can’t deliver those optimizations if you’re mired in processes that should be done by machines. You can’t improve if you never have time to focus on improvement. Today, the reporting process takes about the same amount of time, but much less time is devoted towards those manual, repetitive, relatively low value tasks and more time is devoted towards the insights, the recommendations, the strategies that good reporting should beget.

The key takeaway here is that AI can be used to save time and dramatically improve the quality of your work product incrementally. You don’t have to – and probably shouldn’t – wholesale replace giant processes with AI-driven processes, especially at the beginning of your AI journey. Instead, look at piloting the use of AI for frequent small wins, small tasks that can add up to big efficiencies in the long run.