Are generative agents, which are AI algorithms, powerful enough to excel marketing strategy? It could be, but at the moment it's just frustrating to watch.
Godmode boasts the ability to assist users in generating unique and creative outputs in various contexts, such as determining the best market to launch a product, writing a resignation letter, or exploring hypothetical scenarios, such as an advanced pre-ice age civilisation.
When faced with the question "what do you want to do with AI" the limitless possibilities cause us to struggle to find problems that AI can solve.
The interface is beautiful, but the starting page is too simple. Similarly to Google, you're faced with an empty field to generate results. Where Google works best with questions that results in the best web pages to provide answers, Godmode claims to do the heavy lifting of exploring a new idea or fleshing out a concept.
The issue with this, and with AI at the moment, is that when faced with the question "what do you want to do with AI" the limitless possibilities cause us to struggle to find problems that AI can solve. We are limited by our own prompts, and the likelihood is that we already know the answer in some way, so we're simply asking to find out if AI can replicate our thinking. This empty input box approach doesn't help us to discover things we haven't asked about.
Entering in "improve digital marketing with AI" felt far too macro to gain an output, so I focused my task.
"Develop an AI-powered recommendation engine for personalised content targeting."
Using AI typically requires some knowledge of the process to solving this task, however Godmode provides each step of the plan, adds the reasoning for the step, then asks you to approve the plan or give it a suggestion for a different direction. It's nice, but if you have some knowledge of what you're trying to get out of Godmode already, it's pretty slow and frustrating to approve each task when they already feel logical. Where ChatGPT will provide you with a list of logical answers without much context, Godmode's reasoning can be useful.
As you may know, my exploration of AI is cost-free, so I'm using all free versions of AI tools that can potentially improve digital experience generation and marketing. So after a few approvals of next steps, Godmode asked me to pay, so I had to wait until traffic died down, and at that point I was approving plans without reading them to see where they would go.
The process was fairly logical, but each step was tedious. Although the plan was completed by AI, I had to manually approve each step and ask Godmode to share it's findings as high level points so I could check it was headed in the right direction. Even then, the plan and results were obvious:
I've found articles, now I need to read them. Approve?
I've read the articles, now I need to analyse them. Approve?
I've analysed the articles, now I need to summarise them. Approve?
I've summarised them, now I need to put them to use. Approve?
Apparently this has been a common issue with AutoGPT, which powers Godmode. The results send you round in circles without giving you a concrete answer.
While the loading text was delightful (reticulating splines), ultimately I haven't had a great result from using the free version, but perhaps this would be more useful for smaller tasks.
Comments