REAL Talk With Sam Holcman

The term “prompt engineering” has become popular in recent years to describe the process of carefully designing inputs - known as prompts - for large language models (LLMs). The phrase suggests that prompting is a deliberate, technical process similar to real engineering disciplines such as civil engineering, electrical engineering, or software engineering. Yet, when examined more closely, “engineering” may not be the most accurate word.

Recent advancements in Artificial Intelligence (AI) have been touted to result in dramatic increases in productivity and consistency throughout Enterprise Architecture practices. However, our experience with our clients demonstrates that excessive dependence on AI-driven support - particularly for critical decisions - can bring significant risks to business operations and strategic integrity.

Enterprise Architectures and Business Architectures often suffer from a common problem: the outputs produced by architecture teams may be technically accurate, possibly methodologically sound, and rigorously documented - yet they remain unusable for the decision-makers, stakeholders, and business leaders who most depend on them. This issue, known as Human Consume-ability, highlights the gap between creating models and architectures, and delivering insights that can be readily understood, trusted, and acted upon across organizational boundaries.

AI - Artificial Intelligence.  OI - Original Intelligence.  While AI can excel at processing vast amounts of data and identifying patterns, Original Intelligence allows humans to recognize when solutions need to deviate from established norms or when data is incomplete or misleading.  Learn more in this Broadcast

There is always a lot that we can learn from successful technology and software developments. There is also a lot we can learn, unfortunately, from technology and software development “failures”.  In analyzing over twenty well-documented and publicized failures, one fundamental issue came through loud and clear. A major mismatch between the enterprise data representations and processes, and the vendor’s data representations and processes.  There is a pretty straightforward way to address this situation.

Yes, anything you say to AI can be used against you, including things that you thought were deleted.  Yes, hitting the delete button or key really does not do anything.  Let me translate.  Public facing AI and associated models are “dangerous”.  EAI – (Enterprise Augmented Information) – AI within the boundary of your Enterprise is the answer.  Please listen.  And we can help.

What is the oil that makes the artificial intelligence world run?  Data.  Data infrastructure is the key to moving away from guessing to true usefulness.  Data infrastructure is made up of two parts – the actual technology, and the data itself.  Billions are being spent on the technological side.  The data itself is being starved.  Why?  Because it is hard. We will describe the five transformations needed to make data AI ready.

This broadcast is a commentary. One line I am quoting – “AI models should treat content taken from social media and the internet as untrustworthy sources.” Large Language Models (LLM) ARE the issue, and there is no solution to this – what is public is subject to manipulation. What is the answer to this issue? EAI – Enterprise Augmented (Almigated) Information. Also known as Small Language Model. 

Enterprise architecture (EA) and Business Architecture (BA) certifications are often seen as strategic moves - intended to elevate professional credibility, further organizational maturity, and equip leaders with frameworks and methodologies to drive business optimization and transformation. But, unfortunately, most certifications do not live up to these expectations.

The hype around Artificial Intelligence (AI) continues to accelerate.  Billions are being spent in industry, by investors, and by you and I.  Beyond FOMO (Feal Of Missing Out), where is all of this going?

This Broadcast provides some valuable insight and questions you need to ask, as you take this rollercoaster ride.  Fasten your seat belt and be sure you also have a tight harness.