How to Grow AI/Generative AI Inference?
This week, I read about Chet Holmes’s Buyer’s Pyramid. It was the first time I saw this pyramid, and the source of this is ‘The Larger Market Formula.‘
Why this got my attention, I am linking this to the situation of Generative AI Landscape, which is going on the market where we all started realizing sooner enough that whatever investment is
Going on the AI Infrastructure is about what & How we will use, and we need to have an AI Infrastructure in place to do the Inference.‘
In the AI field, what does inference mean? Let’s get clarity in simple terms.
‘Inference is the process of running live data through a trained AI model to make a prediction or solve a task.’
Definition of Inference:
In artificial intelligence, especially machine learning and Generative AI, “inference” refers to the process by which a trained model makes predictions or decisions based on new, unseen data. After an AI model has been trained on a large set of data during the training phase, it reaches a state where it can apply what it has learned to actual tasks – this application phase is inference.
In generative AI, inference is the stage where the model generates coherent, contextually relevant outputs tailored to specific prompts or queries after being trained on vast datasets.
Few examples:
- Customer Support Chat Bots use inference to understand and respond to customer queries effectively, providing relevant solutions based on the conversation’s context
- Medical Diagnosis AI Assistance helps diagnose diseases by interpreting medical images, where inference allows the model to provide potential diagnoses based on visual inputs
- Personalized Recommendations in e-commerce, Generative AI can suggest products personalized to the user’s tastes and previous interactions, achieved through inference.
Linking to the Buyer’s Pyramid:
Let’s consider the increasing levels of interaction and commitment to Generative AI within an organization or industry, which mirrors the structure and intent of Chet Holmes’s Buyer Pyramid.
60% Not Problem Aware:
At the bottom of this pyramid, 60% are unaware of what purpose they will use AI Infrastructure for; in other words, what use cases or areas in the organizations can lead to using Gen AI / AI.
Vendors are releasing the latest models and capabilities and investing billions in AI Infrastructure, but most organizations don’t know for what purpose they will use this infrastructure.
Action for this: You and me, we all, need to invest time, resources, and talent to find out the target ‘ AI/GPU Inference’ where to use this AI. Massive works need to be done.
This leads us to read, observe, and hunt for use cases in any business process where AI/Gen AI can be utilized.
This could lead to an entirely new operating model or a wholly new enterprise architecture for the organization, which would help us infuse the AI.
20% Problem Aware:
This twenty percent has identified a few areas and is waiting to see if someone else will do it. Then, we will take action. This group is looking for others’ success stories or lessons learned.
However, they also need to be moved from looking into it to let them know what Gen AI could do in their business; in other words, they need to be aware of the Inference Use cases or how to use it.
This is the most challenging part as of now.
17% Information Gathering:
I am categorizing these seventeen percent into those who have selected a few pilot use cases for Generative AI and are going for POCs/ MVPs; they are investing resources and money to move in the direction where they can get some outcome.
3% Buying Now:
The three percent group has actually bought or is buying it, has already implemented the use cases of Generative AI into business functions, or may have started some business solely focused on AI.
Call for Action:
There is a long way to go, and massive efforts need to be diverted to how to apply Gen AI and AI to business. When we start looking at this one, how to use it, it brings a million-dollar question:
Why use it, and what business benefit will I get from it?
Therefore, I believe each Generative AI Use case needs to be divided into these 7 areas to show benefits to the organization.
This is how I think, and that’s how I have started documenting/ curating several AI Use cases as a repository. You are welcome to contribute and share. By sharing your insights and experiences, we can learn from each other.
