The endgame of Data Literacy - Agentic AI and its Applications
Cuspide Data’s Approach:
We start with data literacy and guide your workforce through a structured, practical upskilling journey—from analysis to AI agents—empowering them to build, govern, and innovate with next-gen AI solutions.
Introduction: Generative AI’s Evolution
The release of ChatGPT in November 2022 started the race to develop dynamic and sophisticated AI products. During the 2010’s, we witnessed the development of what I call the first generation of AI products such as chatbots (i.e. Amazon Alexa and Apple’s Siri) and the development of Computer Vision which is based on identifying, understanding, and analyzing videos and images of people or objects. ChatGPT was the first product of a new line of AI products known as generative AI. It is also what I call the second generation of AI products which centres on productivity, efficiency, and creativity.
Automation’s rise during the 2010’s, while beneficial for businesses, still had a number of limitations which included a strict logic of ‘if-then’ and worked for structured and repetitive tasks. However, as we all know, the vast majority of data consists of either unstructured (i.e. photos) or semi-structured (PDF documents). Part of the reason generative AI caught the attention of everyone was its ability to analyze and generate information from ALL types of data.
In a little over 2 years, generative AI has undergone a rapid evolution of product design and development. We first learned about prompt engineering, the ability to define what we are seeking to then receive an appropriate response. The second phase saw the release of multimodal generative AI products such as DALL-E and Midjourney. During this period, the development of large language models (LLMs), the brain behind generative AI reasoning, continued to evolve. This led to the development of retrieval augmented generation AI products which helps us acquire insights to any data (i.e. spreadsheets, e-books, videos, etc.) that we attach.
The year of Agentic AI
2025 is now the year of agentic AI. Businesses, economists, and investors are all talking about the potential. AI agents is a program that can work autonomously by performing tasks for a user or another software system. AI agents can perform a variety of responsibilities such as decision making and problem solving. Like sensors in your vehicle, it can interact with its environment to acquire context and understanding. Like a computer’s storage, it can store data of all types, and like your computer’s RAM, it can process information. AI agents work by integrating the development of advanced natural language processing within large language models and respond to what the user asks. In other words, it is an advanced and promising form of generative AI and will herald the next (third) generation of AI products which will now centre on personalization, customization, and professionalization.
How AI Agents Work
The centre of AI agents remains the well known LLMs. You still need to understand the technique of prompt engineering to generate a proper response. You also have to ensure you are utilizing the right type of LLM. This is due to the type of data the LLM is trained under. But this is where the similarities end and the additional features begin.
AI agents will obtain up to date information, optimize the flow of content, and create tasks on its own to achieve whatever goal you provide. Thus, AI agents will learn and adapt to the expectations of the user over time. As I mentioned in the above section, AI agents have built-in memory to store any prior requests and conversations. This will increase the personalized experience and customize your expectations. There are three general steps that AI agents will follow.
Goals: Whenever an AI agent is designed, specific goals or objectives are provided along with the means to create tasks to accomplish them. If the prompts are simple, task breakdown is not necessary.
Reasoning: Like any generative AI product, AI agents will base their actions on the data or information you provide. If you design it properly, AI agents will use the tools either within your own business, the internet, or other sources. It can then access the data or information required, update its own knowledge base, and then it will reassess its plan. It will continue to do this until it has the data or information it requires. From this point, it will undergo its decision making tasks and then provide you, the user, a response.
Learning: Once the AI agent provides you with a response, you can provide feedback or follow up. The AI agent will store this information and improve its performance and adjust the preferences of the user for any future goals.
There are many standards and types of AI agents and we will revisit this in a future article.
The Foundation: Data and AI Literacy
Despite the advancements in AI products, the benefits (and disruptions) it provides, as well as the risks and limitations, it all comes down to this statement: Every advancement in AI originated from a foundation in data literacy.
Data literacy is a cornerstone of any major business or economic transformation that occurred over the past 30 years. Data literacy can be best described as the ability to learn, understand and utilize data, just like how we learn how to drive a car, learn a new skill, or learn a new language. The exponential growth of data itself was a key part in the development and widespread use of generative AI. The same can be said for the third generation of AI products, agentic AI.
Why is this the case? While AI agents are built on architecture (i.e. software), function (i.e. the goals), and program (i.e. logic and requirements), it is powered by one item: Data. The saying of ‘garbage in, garbage out’ still rings true today as it did back in the 1960’s when the term was first coined by George Fuechsel. Without quality data, you cannot effectively utilize AI products, not even AI agents. There is a reason why generative AI has a variety of scores around accuracy.
Cuspide Data’s Approach
Data literacy is the start and will lead you to the mastery and ability to effectively develop and use AI agents. Learning how to collect, process, and prepare your data will lead to its effective use under AI products such as generative AI. It will also help develop effective management, governance, ethical, privacy, and security frameworks to protect your data and information. Cuspide Data has a ready to use upskilling program where workforces can transition from positions such as the following.
Data Analyst to Business Analyst: Transitioning from data analysis to applying insights within a business.
Business Analyst to Machine Learning Modeler: Learn how to develop different models to help predict business outcomes.
Business Analyst to AI Specialist: Develop expertise in AI technologies such as generative AI and develop AI products such as retrieval augmented generation products.
AI Specialist to Agentic Solutions Specialist: Master the design and use of AI agents for your own customized and professional solutions.
Finally, once you reach the innovation stage, you can apply your prior knowledge, skills development, and professional experience to help create and use the next generation of AI products.
Conclusion
The evolution of AI from chatbots and automation to generative AI and now Agentic AI signifies a transformative shift in how businesses utilize AI products. As AI agents become more advanced, their ability to provide personalized, customized, and professionalized solutions will become the cornerstone of successful business strategies. However, achieving this vision requires a strong foundation in data literacy, which serves as the starting point for developing and implementing effective AI solutions.
The journey towards mastering Agentic AI begins with data literacy, building a data culture, and setting the path towards innovation. Any employee can progress through the customized learning journey from data analysts to Agentic Solutions Specialists. This will become essential for harnessing the full potential of AI agents.
Cuspide Data’s structured approach to AI upskilling is designed to guide individuals and organizations through this transformative process, ensuring they remain at the forefront of AI-driven innovation. With the proper training, frameworks, and practical application, the vision of Agentic AI is within reach, paving the way for a productive, efficient, and beneficial future.
