Generative AI: Unleashing the Power of Pre-Trained Models

Akash Jattan
31 Jan 2024

It’s been a couple of years since I’ve blogged, but part of my New Year’s resolution has been to start sharing my views about data and technology. I think everyone’s been talking about generative AI, ChatGPT and LLMs, so I thought I would start there.  

The generative AI wave is a fundamental change, and its impact will be wider and greater than expected. It will accelerate the 4th industrial revolution from an adoption perspective. In my opinion, three key fundamental aspects are:

  • Generative AI, spearheaded by tools like ChatGPT, is empowering businesses to develop intelligent applications faster than ever before without starting at 0%. Traditionally, most organisations invest $100s – $m’s of investment in training models to get the basics right. Now you can arguably start at 60 – 80% with pre-trained models and focus on the unique experience with custom data. 
  • This wave of AI is promising to revolutionise how we interact with machines and create content. This is driven by how easy it is to adopt the solution. It engages, contextualises, and delights you with no training required as it humanises the experiences very easily.
  • Its impact extends beyond the technology realm to the boardroom, raising profound questions about the future of work, creativity, and ownership in a world where machines can generate remarkably human-like text, images, and more. 

Understanding Generative AI Architecture

The generative AI domain is a new Machine Learning Model (MLM), or some would agree it’s part of a Natural Language Processing (NLP) domain. From what I understand, ChatGPT uses pre-trained models in a unique architecture known as Transformers. These Transformers are different from recurrent neural networks because they can process large amounts of data at once and have long-term memory, which allows for better contextualisation. Gen AI is a sophisticated way to analyse large data sets and has pre-trained models for images, text, voice, and video. In summary, the key architectural elements are:

  • Transformers take centre stage: Unlike traditional AI models, generative AI leverages a unique architecture called Transformers. These powerful models can process large amounts of data simultaneously and maintain long-term memory, enabling them to produce highly contextual and relevant output.
  • Pre-trained models provide instant value: Generative AI offers a treasure trove of pre-trained models for text, images, voice, and video, ready for business consumption. These can be tweaked and tuned over time for better accuracy or wider data sets. This eliminates the need to build models from scratch, accelerating development and time-to-market.
  • Accessibility and integration is simplified. Gen AI stands out as It’s accessible through APIs, allowing organisations to integrate its capabilities seamlessly into their applications.

Business opportunities

In my opinion the fundamental proposition of Gen AI is it humanises the experience very quickly. No need to train yourself on a tool, you just ask it a question like you would with a human. The iteration or training will be on how to articulate questions properly to get the best answers. Which is like human interaction. The key opportunities are:

  • Streamlined model development: Generative AI eliminates the need to start from scratch, enabling businesses to begin with models that are already 60-80% accurate and optimise them for specific use cases. This dramatically streamlines the machine-learning process.
  • Ubiquitous adoption drives innovation: The accessibility of generative AI services like ChatGPT is fuelling a wave of innovation across industries. Businesses can now rapidly build apps that leverage these pre-trained models, delivering contextually relevant experiences to users and accelerating enterprise-wide adoption.
  • Augmenting app experiences: Generative AI is empowering businesses to enhance existing products and services with new capabilities, such as:
    • Virtual assistants that understand natural language and handle complex tasks effortlessly with organisational data
    • Personalised content recommendations that anticipate user preferences and deliver engaging experiences.
    • Intelligent chatbots that engage in natural conversations and provide seamless employee or customer support with organisation data e.g. contract analyses
    • Productivity tool for the workforce to create, discovery, summarise and simplify content.
    • Profiling, understanding challenges with existing data and then implementing an improvement plan

Navigating the Ethical and Societal Implications

While generative AI holds immense promise, I do believe it also raises critical questions:

  • Impact on the workforce: How will existing skillsets adapt as machines become increasingly capable of automating tasks? Will certain roles become obsolete, or will new opportunities emerge?
  • Ownership of data and intellectual property: Who controls the rights to the content generated by AI models? Can organisations use this content for commercial purposes? What are the ethical implications of using personal data to train AI models?
  • Impact on human creativity and communication: Will future generations rely too heavily on AI-generated content, leading to a decline in their ability to articulate thoughts and ideas independently?
  • Data Governance of how Gen AI uses organisational data, does it access all your data? does the users know the boundaries? what if it opens up sensitive data widely for consumption?

The Path Forward: Systems Thinking and Ethical Considerations

Generative AI is poised to transform business and society in profound ways. To ensure its responsible development and deployment, I think we must embrace:

  • Systems thinking: Approaching generative AI from a holistic perspective, understanding its interconnected impacts on individuals, organisations, and society as a whole.
  • Ethical frameworks: Establishing clear guidelines for responsible AI development, ensuring transparency, fairness, and accountability.
  • Collaboration and dialogue: Fostering open discussions among stakeholders to address the challenges and opportunities of generative AI, shaping a future where it benefits everyone’s sustainably.

Generative AI is here to stay and will be adopted by consumers and engrained in our childs life. So, fighting it won’t help and may leave your organisation behind. So, here’s a list of what I think you could do and where you could start. 

  1. Nurture employees – Train them on how to use it safely in its current internet form, regardless of the vendor or tool.
  2. AI or Generative AI Roadmap – Define a strategy or plan on where you think it could benefit your business.
  3. Secure Sandbox or Pilot – Start a small pilot, but with the business. Understand the value and bring the business on the journey with you.
  4. Establish your AI operating model or governance – During or after your pilot establish the key processes, people and frameworks for you realise more and sustainable value from Gen AI.

Generative AI is a game-changer with immense potential to transform businesses and society. By embracing its responsible development and deployment, we can harness its power for a better future.

Generative AI: Unleashing the Power of Pre-Trained Models
Share this
Picture of


Akash is a senior leader experienced in creating emerging products and leading data transformations programs. His expertise is business transformation by leveraging modern data and AI solutions. He has worked in Australia and NZ across multiple industries successfully launching multiple products to market and leading multiple data-driven business transformations programs. Akash has consistently demonstrated strong leadership in building executable visions, creating high performing teams and commercialising products.

It was his passion to drive business transformation through data, that drove him to become the Founder and now CEO of dataengine in 2018. Akash regularly speaks at universities and conferences about the evolution to DataOps as a foundation to analytics in the business.

Want to know more about dataengine?
From Chaos to Collaboration

From Chaos to Collaboration: How Communication Builds Strong Stakeholder Relationships

Serena Roncolato
Serena Roncolato

Enhance the Power of AI Safely: A Guide for Business Leaders

Glenn Johnstone, Chief Operating Officer at dataengine
Glenn Johnstone

Secure Generative AI Adoption: A Guide with GenAI Accelerator, the Knowledge-Based Solution

Tushar Mehta
Tushar Mehta