Marketing Activation & Campaigns with GenAI

Is there any existing GenAI model can be refer or apply to customer experience, productivity, and growth?

  • Effective, tailored lead-activation campaigns
  • leveraging predictive analytics
  • data-driven recommendations

ref: https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/ai-powered-marketing-and-sales-reach-new-heights-with-generative-ai

I think that you might want to look at a combination of components, and not just a GenAI model. For example, for the data-driven recommendations, you will need a dataset of items and/or user-item history to build a recommender system, and then couple it with an LLM (GenAI) to produce a nice recommendation. For predictive analytics, you may also want to look at some ML models that execute the predictions, and again couple them with an LLM. And for the lead-activation campaigns, sometimes it works to fine-tune a model with instructions that appeal to the specific user-case.

As you can see, it is not just GenAI - it is a suite of components, and GenAI can be the ‘brain’ that helps orchestrate all the pieces of the system.

Great comments! Found something interesting with OpenAI for a recommendation engine, but it’s costly way. will try to find an open source approach.

For recommendations engines, you may also look at other technologies/libraries.

First, define if you want a content-based or a collab-based recommender system (or a hybrid). There may be other categories that these are the main one.

It is common to have both, specially when there’s a cold-start problem.

Once you define what you want, then just search for the many libraries that exist to create rec systems. Surprise is a well known, but there are many.

Finally, build your own rec system “for free” :slight_smile:

Of course you can use LLMs for this too, I am not sure if for the entire solution, but certainly for parts of it.

But my main point is: you don’t need LLMs to do a recommendation system.

I’ve seen the video. They discuss several topics. Any particular topic you’d like to discuss?