Surprising AI Findings So Far

A recent study that found that over $100M in investment in AI by mid-sized companies generated the following result:

Just 3% added productivity. 

Not nearly the kind of results that are being tweeted about or shared on LinkedIn by marketing “influencers” everyday. 

But there are ways to get more done with AI, if you’re doing it right.

I’m Nick Avaria, agency owner, and founder of Agency Acquisitions. Every week I share a client story, insights, and tips into scaling your agency faster. For exclusive video content, make sure you hit Subscribe on my YouTube channel.

What’s a realistic timeframe for seeing a return on AI?

This study went on to say that the time horizon to get ROI on that $100M initial investment would be many years.

2 things:

1. This was for mid-sized companies that are generally more efficient and effective than small businesses

2. It was believed that that the efficiency gains were kept by the employees

   1. Translation: there were gains but the company saw very little of them. The employees absorbed the gains by having less work time or taking longer on other tasks.

Caveat - this study was done 6 months ago and the findings were released a couple of weeks back.

AI’s capability has already changed so much in that time that I personally think the gains would be more if re-studied today.

But there are key insights we can draw from this:

It's still early days and I think for any marketing agency sub 100 employees the current gains from AI are much more significant.

Simply put, small agencies are already not efficient and do not have maximum effectiveness even in a pre-AI world.

So this is an opportunity to get both efficiencies at once. The operational and the AI at the same time.

The operational one has to be first then the AI one because AI can only automate/do the things that have clear processes and parameters.

Here are the steps that I’ve seen work inside of agencies to get the most out of any AI implementation:

1. Identify the specific thing you are automating or getting done through AI

2. Write an SOP on how to do the task without AI

3. Clearly document what:

   1. The inputs are

   2. The process used to process the inputs (what are the principles, methods, etc.)

   3. Examples of what the output looks like

4. Do a time study of how long the task(s) take to do

5. Re-write the SOP with AI included

6. Add-in a quality assurance step to review the work done by AI to the SOP

7. The most important step: Add work (increase output expectations) to the individuals that will save time due to this specific AI deployment. 

   1. If you are an agency that already measures time this is a lot easier to do as this is how you should always do efficiency/ops work. 

Step 7 is the key - if you don't intentionally capture the efficiency/time gain it will be lost/absorbed by people meaning no productivity gain. 

Next week I’ll share my predictions on AI: what will get replaced and what will weather the storm.

Is your agency using AI to gain productivity?

Hit reply with your favourite tools or systems for AI so far. Until next week,

Nick