AI makes work up to 41% faster, according to Bain data.
81% of AI users say they see increased productivity, shows our 2023 AI research.
And according to 2024 Microsoft data, 75% of employees already use AI.
In short, we can safely say that AI is here to stay in the Workplace.
No one – whether an individual contributor who needs to deliver as much quality work as quickly as possible or a CEO looking at the productivity of tens to thousands of employees – will say no to the gains AI has to offer.
So, what is AI? How do we use AI in the Workplace?
And how do we transform traditional teams and businesses into AI-powered ones?
I’ve transitioned myself and my company, done the research, spoken to the experts, and will share how to go from 0 to AI Hero as succinctly as possible.
5. What are the benefits of AI?
6. Which LLM (ChatGPT, Gemini, Copilot) Should I Use?
7. Which AI Tools Should I Use at Work?
8. Which AI Tools Should I Work Specifically to My Role?
9. What Are the Barriers to More Companies Adopting AI?
10. What Trends Will We See for AI in 2024?
11. What Jobs Will AI Replace and How?
13. How AI Will Impact Management
14. AI’s Impact On The Job Market and Salaries
15. As a Business Leader, How Can I Help with the Adoption of AI in My Teams?
What is AI?
You likely have heard of or are using ChatGPT, whose advent raised eyebrows and sparked daily discussions about using artificial intelligence (AI).
However, it's important to note that AI is not a new concept. It has been around for decades, and its evolution has been gradual yet significant.
Artificial Intelligence, or AI, is the ability of machines to simulate human intelligence, including problem-solving and learning.
Artificial Intelligence is technology that enables machines to mimic human intelligence, performing tasks like recognizing patterns, making decisions, and learning from data.
AI was around in the early 1900s, but it really started flourishing around 1950 when Alan Turing (yup, the guy from the ‘Turing Test”) explored the possibilities of “Artificial Intelligence in a groundbreaking paper.
In the decades leading up to this one, AI development intensified, and we quickly found it in many products and services, from Google using it to optimize search to Teslas letting cars drive themselves.
All of these use cases are examples of Predictive AI, which takes vast amounts of data on historical events to predict the future. Another great exampleis Amazon, which as NPR reported in 2018, could predicts which orders will be placed before they occur.
While powerful, Predictive AI has shortcomings, too.
Predictive AI is a branch of AI that focuses on analyzing historical data to forecast future outcomes or trends, enabling machines to anticipate events, behaviors, or needs with a high degree of accuracy.
For example, it takes immense efforts to collect, clean, and train this data. And, it can't predict 'unknown unknowns,' as it can only dig from its training data.
What is Generative AI?
Generative AI is a tool powered by artificial intelligence that can create new content based on the inputs provided by a user.
They use large language models (LLMs) and machine learning (ML) techniques to provide comprehensive answers in human language.
Unlike older language models, ChatGPT and similar apps are based on "Transformers" (the “T in ChatGPT.)
Transformers let the model understand the importance (emphasis) of one piece of data in a sequence as related to others. This allows it to be much 'smarter' than previous models. (This video from Google's Dale Markowitz explains it really well and is easy to understand.)
Generative AI doesn’t stop there. It helps us map out an infinite number of scenarios to a situation, make comparisons, help decision-making with many variables, sift through data, and much more.
New iterations of tools like ChatGPT, Google Gemini, and Microsoft Copilot can even browse the live web to do the research you otherwise would have done.
As a result, Generative AI is taking a lot of work out of the hands of those using it: one in two respondents in our Generative AI at Work study say Gen AI helps them automate Email and Communication, and almost as many (45%) use it for Data Analysis and Reporting. 42% use Generative AI tools for research.
Content creation, often considered the main application of Generative AI, comes in fourth at 39% for writing and editing and 22% for design tasks.
Generative AI’s adoption has been swift because compared to regular AI, the technology is improving daily, creating many interesting use cases, and as experts say, is more "consumer-friendly."
This made it so that choosing AI tools in companies moved from being a tech matter done by the CIO to managers across departments using these tools as an alternative to traditional employee management software.
At the same time, while Microsoft data shows that 75% of employees use AI, this often concerns "Shadow AI," not company-sanctioned AI tools. This makes maintaining security challenging.
ChatGPT
I’ve been speaking a lot about ChatGPT, so what is ChatGPT, actually?
ChatGPT is a chatbot developed by a startup called OpenAI (which most people learned about thanks to the Sam Altman firing & rehiring.)
ChatGPT, which stands for Chat Generative Pre-trained Transformer, is built on OpenAI’s language models and has been made even smarter with additional learning methods.
Give it a question, and it will answer based on more data than any human can ever process.
Doesn’t Google do this, too? Well, no. Google unlocks information by matching your search keywords to web pages containing information about that topic.
ChatGPT gives you a full answer in human(-like) language rather than just links. It’s like talking to someone who has read the entire internet.
Quickly after its launch, The New York Times called ChatGPT "the best artificial intelligence chatbot ever released to the general public."
And the public agreed: ChatGPT garnered 1 million users within the first five days of its release, and to 100 million not long after.
And now, according to our July 2024 Report on the Top 100 AI tools for Work, ChatGPT is the undisputed leader in Generative AI:
ChatGPT is not only accountable for over 67% of all Generative AI traffic and searches, it's even bigger than Netflix, Microsoft, Reddit, TikTok, and The New York Times:
It's making its way to the workplace, too.
In our 2023 study, “Generative AI at Work, 75% of knowledge workers in the US said they have heard of ChatGPT. This climbs to 81% of users and 84% amongst Gen Z and Millennial users.
For a tool that’s only been in the market for a year, 75% awareness amongst all ages and industries is incredible.
It’s also the most used tool.
Among knowledge workers who actively use Generative AI tools, ChatGPT once again reigns supreme: 60% of active Gen AI users use ChatGPT at least weekly.
So ChatGPT is wildly popular in 2024.
But to understand its overwhelming and transformational impact (in case you got used to GPT's power already), take a look at when an earlier version was demoed at TED.
After OpenAI co-founder Greg Brockman explains how ChatGPT works in his TED Talk "The Inside Story of ChatGPT's Astonishing Potential," the Head of TED, Chris Anderson, comes onto the stage, visibly befuddled.
"Oh my goodness, pretty much every single thing about the way I work, I need to rethink," Chris says. "Who thinks that they're having to rethink the way that we do things?," he asks the audience.
Yeah. That’s definitely still the case, and more so than ever.
GPTs
Like the iPhone, which became exponentially more valuable after the launch of its app store, so did ChatGPT get a big boost from launching GPTs.
Announced at at press conference right before Sam Altman's firing, GPTs are small apps that live within ChatGPT.
The simplest way to understand what a GPT is, is that you can train it on something you normally do yourself manually.
For example, do you often review the work of direct reports based on your knowledge or a certain framework?
Then you can now upload your framework to GPT, and rather than checking the work manually, you can let the GPT review it for you.
You can even let your direct report interact with your GPT to get feedback and fix issues before sending it to you.
And creating a GPT is simple:
Building GPTs is as 'no code' as it gets. You create them by talking to ChatGPT, which guides you through the steps:
- Define the Purpose
- Train It
- Launch the GPT
It's as simple as that.
Using GPT is also remarkably easy (and free.)
As of February 2024, you can even tag GPTs directly in your conversations with ChatGPT, and get them to execute your wishes. This makes it even easier to work with GPTs, and should further drive adoption.
For more on GPTs, check out my guide to Create GPTs, or use our Prompt Generator to create 10x more powerful prompts and see more value from your ChatGPT usage.
Hallucination
ChatGPT and many others suffer from hallucinations, making up facts.
McKinsey explains AI hallucinations as follows: “AI confidently generates inaccurate information in response to a user question and has no built-in mechanism to signal this to the user or challenge the result.
Hallucinations in AI happen because of insufficient or biased training data, or incorrect assumptions the model makes.
These hallucinations can have real and damaging impacts, for example, by confirming bias, spreading misinformation, or falsely making someone believe a benign skin lesion is cancerous.
There have been instances where AI-generated content has led to problems.
For example, a lawyer used ChatGPT to prepare a filing in a routine personal injury lawsuit. However, the AI presented fake cases, which the attorney presented to the court. This resulted in the judge considering sanctions against him. Whoops.
Research agency Forrester even predicted hallucination insurance would be a big money maker in 2024. Still, we have to not overthink a few cases of hallucinations versus the huge opportunity for AI to help us work faster – we just have to fact-check its outputs.
There are some strategies to prevent AI hallucinations:
- Choose the right AI models, including purpose-specific tools.
- Implement Retrieval-Augmented Generation (RAG) to anchor responses to verified data sources.
- Master prompt engineering, including techniques like few-shot prompting and chain-of-thought prompting.
- Adopt a multi-model approach, leveraging tools like ChatGPT, Claude, and Perplexity to cross-verify information.
- Establish rigorous fact-checking protocols.
For more tips, check out my guide on Preventing AI Hallucinations and join Lead with AI, our community of business leaders embedding (error-free) AI in their work, teams, and organizations.
Prompt Engineering
Prompting is the input for generative AI to create its outputs.
The best way to think about, as you get the most out of Generative AI if you treat it like a coworker, not software, is that prompting is like briefing your colleague. This means, the better your inputs, the better the outputs.
This is why simple prompts, like "write me a social media plan," don't result in great outputs. While AIs like ChatGPT know everything in the world, this prompt doesn't provide it with enough context to output great work.
This is why more people than ever study prompt engineering.
Prompt engineering is about crafting inputs to AI models to get better outputs, a key skill for getting the most out of AI tools.
CO-DO SuperPrompting
CO-DO Superprompting is one technique, similar to CARE prompting.
CO-DO stands:
- Character: Please imagine you are: [Who or which role should AI take on?]
- Objective: I need to: [Task at hand]
- Do’s and Don’ts: You should: [Do's], and Please avoid: [Don'ts]
- Outputs: The final output should be a (an example or starting point is helpful.)
It provides a structure for creating prompts and ensure you provide enough context for the AI to have what they need to successfully execute your briefing.
To write your first "CO-DO SuperPrompt," check out our free Prompt Generator.
Chain of Thought Prompting
But you definitely don't have to stop your prompting journey there. Chain of Thought Prompting (CoT Prompting) is another useful method.
CoT Prompting breaks down complex questions into steps, guiding the AI through a reasoning process.
For a business problem, you might have the AI outline the issue, analyze solutions, weigh pros and cons, and recommend actions. This improves response quality and shows how the AI reached its conclusions.
Check out our guide to Chain of Thought Prompting here.
Few-Shot Prompting
Another way to increase AI's abilities is by giving ChatGPT, Claude, Gemini or Copilot examples, also called "Shots."
By giving AI examples of what 'good work' looks like, the chances of getting a successful output increase significantly.
This technique can be combined easily with Chain of Thought for the best outcomes.
Context Windows
Context windows are also important in AI models, especially when you're prompting with a lot of information.
The context window is how much information the AI can consider at once.
Larger context windows let AI handle more complex scenarios.
Different AI models have different context window sizes, affecting what tasks they're best for - from quick insights to in-depth analysis.
See our detailed guide to context windows here.
What are the benefits of AI?
So, with that history lesson and frightening look into the future of AI out of the way, what are the actual benefits of AI, as measured in 2024?
According to AI Statistics, there are a ton of benefits to using AI, including:
- AI makes work up to 41% faster (Bain.)
- 81% of AI users say they see increased productivity (FlexOS.)
- Employers are willing to pay an average of 47% more for employees with A.I. skills. (Amazon)
- 87% of AI users say it helped them develop new creative, analytical, and technical skills. (FlexOS)
- 97% of business owners believe that ChatGPT will benefit their businesses (Forbes)
For 100+ Statistics and Trends, see our Guide to AI Statistics.
Which LLM Should I Choose?
OpenAI's ChatGPT 4o's launch disrupted the LLM market in a major way, and its recent ChatGPT o1 addition only increased its lead.
While Google also announced some updates, which could impact companies already running on GSuite, I believe ChatGPT is still the best bet for companies not paying for Google or Microsoft’s ecosystems.
ChatGPT has the best scores on Hugging Face, a user-generated leaderboard for AI performance, and has the most traffic, by far. This means there will be more practical use cases and 'apps' available on this platform versus others.
Here’s my latest overview of LLMs for individuals and businesses looking to find the right fit:
Claude, which finally saw a boost in visitor numbers according to our AI for Work Top 100, may be hurt the most, given that the free GPT’4o beats Claude’s paid model Opus
