Who gets the AI dividend?
A few months and blog posts on how AI is affecting the advertising agency model has made one thing clear: there is an efficiency boost in using AI. It could be big or small, depending on what agencies are using AI for. But it is definitely there. Since agencies are a service business, though, it is time to ask who gets the benefit of that efficiency?
Last week, I made the distinction between Efficiency Agencies and Ambition Agencies. The main difference between the two is who gets the AI Dividend. Efficiency Agencies keep it to make their business model more profitable. Ambition Agencies use it to grow their client’s campaigns, through more tests, exploration and learning. Both use AI. Efficiency Agencies use it to grow margins. Ambition Agencies use it grow campaigns (and, hopefully, clients).
Today’s stories all focus on the AI Dividend. How it is created. Who keeps it. And How it gets used.
RPA is an agency that is rapidly adopting AI into its workflows. The interesting thing, though, is what they say this allows them to do:
“As an agency, we are spending more time strategically collaborating with clients rather than spend time doing clunky tasks, AI plus automation is allowing us to do that."
- Lisa Herdman, RPA
This is a great example (of a great agency) taking the benefit from AI to better use its people to work more closely with clients. Interestingly, they are also bringing NewtonX consumer data into the mix, creating a richer dataset for their team to use when advising clients.
Amnet, Magnet and an unnamed auto brand pioneered AI agents to set up and launch a CTV campaign. They are reporting not only that it worked, but the process worked 70% faster than traditional methods.
If nothing else, this gives us a data point into the size of the AI Dividend and maybe even its value. It’s not hard to take that number, get a total number of hours and an hourly rate to get the financial benefit of using AI.
Which then lets you ask: where did that money go? Ambition Agencies would use it to do more.
This one is not agency side, but sell side. Operative is the platform for a number of big publishers and they rolled out a new version that goes big on agents.
The thing I like about this is the inclusion of “audit trails”. As efficiency improves, we expect more tests. Maintaining the reliability of those tests is critical. The idea of audits should let humans feel better about the actions and moves the agents are making.
Mover of the Week: Economic Model of AI.
There is a clear and simple value chain occurring:
AI creates efficiency in the ad agency model, resulting in (what we can call) the AI Dividend.
Efficiency Agencies capture that value on their balance sheet
That approach lacks agency differentiation and generates slop. The returns will begin to diminish.
Ambition Agencies pour the AI Dividend back into the campaign, through more tests, exploration and learning.
That approach also requires a good lab, though, so expect more investment in creating an infrastructure that can be scaled and trusted.
We should see that in…the numbers!
Thursdai index: this week’s numbers.
Experimentation Capability: 6.7
Learning Quality: 6.0
Laboratory Reliability: 7.0
Market Traction: 6.3
More de-hype. Lowered the scores across the board because, while we are seeing a lot of development, we need to see some real results. Closest we have is “time saved” which is awesome. But if we don’t have quality, then time doesn’t matter.
The Thursdai Index and report is the work of 61 Hawk founder Gary Stein with assistance from an OpenAI research agent. The agent sources, sorts and scores news stories. Everything you read was written by Gary, a real human being.
