How to buy AI as a non-technical decision maker.
Buying business technology is difficult at the best of times. This article aims to provide you what you actually need to know, in simple terms, as a non-technical decision maker when evaluating AI solutions.
In recent years, artificial intelligence (AI) has transitioned from futuristic hype to a reality that touches every corner of our lives. Whether it’s conversational AI like OpenAI's ChatGPT, integrated AI tools like Microsoft’s CoPilot, or more ambitious projects like Tesla’s Optimus Robots, AI is dominating the tech landscape.
But amidst the excitement, there is an undeniable wave of sameness which makes decision making particularly overwhelming to a non-technical buyer or small-medium business owner. Inspired by a TikTok in which a local business owner showcased an ‘AI hack’ ...which consisted of copying pasting sensitive business information into Chat GPT... I wanted to provide a perspective on how you can make informed decisions without needing a software engineering degree.?This article aims to explain what to consider, the fundamentals of AI and what you need to know, and break down common buzz words so you can feel confident in your evaluation.
AI Everywhere: From Co-Pilots, Agents and Tesla Robots.
Nearly every tech company today has an AI product or feature, often framed as a "co-pilot" or "agent." These tools aim to augment human capability, automating tasks and simplifying decision-making. While innovation in AI has been remarkable, there's an emerging trend of companies offering nearly identical products, with only minor variations in branding and interface.
For instance, Tesla’s recent announcement of The Robotaxi and Optimus is a glimpse into a future where AI agents may transcend screens and enter the physical world. Whether these robots were voiced by humans at the October launch is still being debated and there’s no official release date on the horizon, however, the potential impact of this technology is significant. Imagine robots performing tasks autonomously in industries ranging from manufacturing to healthcare. But where on earth do you start?
Generative AI vs. Agents: What’s the Difference?
There are a lot of buzzwords flying around like 'co-pilot' and 'agent'. It’s important to know what type of AI you’re purchasing and what it’s designed for. Neither type is superior, but they do produce different outcomes.?
To keep things simplistic, Imagine Generative AI is like a writer in an office. You give this writer a brief—"Write me an article about innovation" or "Design an image for a new campaign"—and they go to work. They don’t pull words or designs from thin air; instead, they rely on everything they’ve ever read, seen, or learned (like books, artworks, and data) to create something new based on the patterns they recognise. This writer isn’t doing anything else—no scheduling meetings or making phone calls—they just create based on what you ask.
Generative AI, like this writer, is great at producing new content. Whether it’s an image, a piece of text, or even some code, it takes the patterns from past information it’s been trained on and generates something unique for your specific request.
Now, imagine an AI Agent as a multitasking assistant in the same office. Instead of just writing or creating, this assistant handles various tasks—like booking meetings, responding to emails, or even fetching coffee. You give this assistant a goal, like "Schedule a meeting for tomorrow," and they go ahead and figure out the best time, send out invites, and confirm the details. They can also make decisions on your behalf and even carry out more complex tasks, like navigating an office or working with other assistants (robots or software) to get things done.
While the AI agent might ask the Generative AI for some help (maybe to write an email or create a graphic), their main job is to act, interact, and complete tasks autonomously, based on goals you provide. They’re more concerned with executing actions and making sure things get done efficiently.
In this creative office, Generative AI is your go-to for producing new content—whether it’s writing, design, or data synthesis. Meanwhile, your AI Agent is the one running around, executing tasks, and managing day-to-day actions based on your needs. One creates, the other gets things done.
So...What do I need to do if I'm procuring AI for my business?
Everyone is at a different stage on their AI journey, but with the Australian Chamber of Commerce and Industry predicting 90% of businesses incorporating AI by 2026, chances are you’ll be starting soon if your business isn’t there already.?
Before speaking with a vendor, try and understand which processes you want to optimise. Look at where people waste time in their day, and what tasks people don’t like doing. Also ensure someone in the business is owning the project from both a technical and adoption perspective. Once you know this, you can then reach out to vendors or your current Account Executives. Be clear with what outcomes you want to achieve. In my experience, the more collaborative the customer, the better the outcome, and often, the better the price.?
Once proof of concept is established with a short list of providers, become familiar with how AI use is commercialised. If it’s charged via a consumption model, ensure that you are comfortable that AI usage increase would justify the increase in associated costs. If it’s charged by cost per user, ensure that headcount won’t restrict usage.?
...and what do I need to know if i'm procuring AI for my business?
When it comes to understanding AI, there are three fundamental components to keep in mind: the user interface (UI), the model (large language model/LLM), and the data that fuels the system. The caveat here is that technology is changing rapidly, so if you’re reading this post-2024 the following may not be as relevant.?
领英推荐
User Interface (UI)
This can be in text form (eg. ChatGPT, Microsoft Copilot), a voice based interface (eg. Amazon Alexa, Google Assistant, Siri) or even a Graphical User Interfaces (eg. Self-service kiosks, phone applications). This is how we interact with AI. It could be as simple as a chat box where you type in questions, or as complex as a robot responding to voice commands. The UI is crucial because it determines how intuitive and accessible the AI is to the user. A well-designed interface makes the AI feel seamless, while a poor UI can hinder even the most advanced AI model from being effective. As AI becomes more embedded in our daily lives, UI design is evolving to make interactions more natural, whether through speech, touch, or even gestures.?
The UI is what your technology vendor will be selling, and the following are important questions to ask.
The Model:
The heart of AI is the model, which in many cases is an LLM (Large Language Model). This model is trained on vast amounts of data to understand and generate human-like text, recognize patterns, make predictions, or even create content. Models vary in size and complexity, with larger models typically being more powerful and capable of handling nuanced tasks. However, it’s important to note that models don’t think or understand in the way humans do—they rely on statistical patterns in the data they’ve been trained on. The effectiveness of a model also depends on how it’s fine-tuned for specific tasks, so understanding that not all models are created equal is key to setting realistic expectations. An additional aspect to consider is the feedback loop—how an AI system improves over time.
Typically you will see UI providers partner with LLM providers (ie Open AI) to create the AI experience, which further leans into this feeling of sameness. What to consider:?
Data
Data is the fuel that powers AI. The more high-quality data an AI system has access to, the better it performs. This data can come from a variety of sources, such as text, images, voice recordings, or real-time sensor inputs. It's essential that the data used is relevant and accurate because poor data can lead to inaccurate predictions, biassed outcomes, or subpar performance, regardless of how advanced the model is. In addition, the ongoing cycle of collecting and refining data (called data curation) is critical for keeping AI systems relevant and up-to-date. However, data privacy and security are also major concerns, particularly in sensitive applications like healthcare or finance, where ensuring ethical use of data is crucial.
So, if you’re looking for AI as a business tool, you need to ensure you’re going beyond just the choice of UI to get the outcomes expected. Key considerations are:?
Finding the Balance Between Hype and Substance
While AI promises to revolutionise industries, there’s a growing sense that AI Tech is becoming homogeneous, even for those in the tech space. It can be hard to determine which 'groundbreaking’ innovation is actually worth investigating. The emphasis on launching AI-powered products quickly has led to a deluge of similar solutions—tools that can analyse text, respond to queries, and provide recommendations. Yes, these features are useful, but how much differentiation can there be when every product is marketed as a ‘agent’ or ‘co-pilot’?
As business leaders, developers, and end-users, we need to remain critical about what AI should be. Instead of falling into the trap of AI for AI’s sake, it’s crucial to push for solutions that solve real problems and differentiate themselves meaningfully.
AI will continue to shape our future, but as with any technological revolution, there’s a balance to strike. As a previous Account Executive, I would urge buyers to make the decision right instead of waiting in search of the right solution. For AI innovators, we need innovation that goes beyond the next co-pilot or agent. The real promise of AI lies in its ability to create new possibilities, not just replicate what’s already out there.
Authored by Verity Collins