This essay is based on a conversation with Geetha Rajan, director of the global strategy team at Freshworks, a SaaS company. She is based in the San Francisco Bay Area. Her identity and employment have been confirmed by Business Insider. The following has been edited for length and clarity.
I am a director on Freshworks’ global strategy team, driving high-priority strategic initiatives that shape the company’s growth, investment decisions, and execution, including AI adoption.
Previously, I spent nearly a decade at PwC advising Fortune 500 companies across healthcare, financial services, and technology on growth strategy and digital transformation. As part of my role, I led the upskilling of over 50,000 employees on automation tools.
Technology transformation is happening all the time, but it took at least a decade for cloud and mobile transformation to be fully adopted. ChatGPT impacted millions of users in its first few months.
Many employers will continue to expect to use AI every day without really understanding the consequences. That’s the pressure, and instead of using it wisely, you actually end up making more mistakes.
Here are some examples of mistakes employees make when implementing AI.
1. Go from 0% to 100% overnight
Many people quickly become Iron Man and try to completely automate their workflow. It’s supposed to be a process. The first step is to treat AI or technology as a super intern. This gives you maximum control over things without giving AI or technology any autonomy.
For example, if you start by giving it structured data and want to validate all output. AI can make beautifully formatted output look illusory.
2. Outsourcing strategy and philosophy
I’m a strategic consultant and advisor. So in terms of ideas and thinking, that’s one of the parts that I don’t typically outsource to AI.
This has been gained through extensive experience in consulting and in the real workplace. First, I would like to mentally write down my model and first principles. I make sure to validate the numbers and even try to extract unstructured data from the AI, but I still stick to first principles and write the first draft very rigorously.
Once the draft is complete, you can ask them to poke holes in the draft and say, “Hey, you’re the most skeptical board member or CFO, please poke holes in my strategy.”
Many of the outputs of AI are highly sophisticated. But if you don’t have that insight, if you don’t see this enough times, you can’t actually tell whether the AI is actually making mistakes. A lot of work is required here. Just take the output of the AI and feed it into your email or analytics.
I’ve made this mistake myself. I had 5 or 10 minutes to spare, so I asked the AI to quickly write down some design principles and throw them on the slides. As I was presenting, I thought, “Wait a minute, I don’t think this makes sense and this isn’t what I actually wanted to say.” In fact, I was ashamed of myself.
You can also easily get caught up in situations where the language used by the AI is not even colloquial or professional.
Sometimes, my biggest worry is what will happen in five years. Then no one actually does the first job and we end up burning the ladder while we’re trying to climb it. AI can do a lot of things, but I think it’s important to work on yourself to learn problem-solving skills and how to use Excel.
3. Approach AI without context
You need to know exactly who you are solving for and what the purpose of solving the exercise is.
For example, if you’re building an AI model to understand your business’ customer segments, you still need to understand the segments at a high level. That’s a part I never want to outsource. If you don’t have that context for yourself, you can just go in a million different directions.
Fundamentals like taste, process, architecture, and how things are built cannot be gained from any tool, whether you’re using ChatGPT or a modern model. If the AI throws out 50 ideas, you need to know which of them will stick. As an employee, it’s your responsibility to choose the right one, so you need the insight and expertise to do so.
