choong pw

While the economy uncertainty has been looming around. It has puts a lot of pressure on the leadership team to increase productivity or cutting costs. The latter is the easy way to win the battle, but lose to the war.

In true leadership fashion, we don’t like the chopping board, and that only lead us to increase productivity. Thankfully, being in tech and most of our work are live in the digital world, AI is here to save the day. But there is a catch, it won’t magically fix any problems. It still requires vision and strategic planning to transform a company to leverage the AI capability.

Things that AI has been well-known for – at least as of today.

1. Automating simple and repeatitive tasks.
2. Automating tedious tasks.
3. Performing analysis.
4. Generating ideas and brainstorming.
5. Expanding on decision making reasoning.

This is what I have for now, maybe there is more to it. It’s about empowering your team to find creative solutions, to challenge the status quo, and to be resourceful. Let’s get into it!

Automating simple and repeatitive tasks

With powerful no-code tool like N8N these day. As a developer, this has been a game changer for me. It come with simple to use AI agent that does almost every automation you can think of. The catch here, learn the fundamental of AI. Example, what is a tool, RAG, LLM model, prompt engineering technique (few shot, single shot, tree of thought), etc.

Here is a screenshot of me attempting to use AI to start a habit of making a weekly blog post. By using the automation to reduce the amount of effort needed to produce a post, hopefully, this can give me the motivation to write more. Fun fact – I am still procastinating a lot.

Word of advice. Be patient, don’t you dare to dream of getting this to work on the first try. Like building any software, is reallly an iterative approach of fine-tuning one small part at a time until you get it right.

Giving N8N a huge shout out, and encourage anyone who want to build something with AI to give this a shot.

Automating tedious tasks

I will categorize code assistance as part of this. Automating tedious work like documentation, code vibing, and scaffolding is a good way to go about it. In the coding world, you have Cline, Co-pilot, Gemini Assistance, and Amazon Q developer – All of which you can get for free. I have share details about vibe conding here, give it a read.

You probably have noticed that almost every SaaS platform you can think of has already embedded AI into its system. Example: for Atlassian you, have Rovo; for Salesforce, you have AgentForce; for Hubspot, you have Breeze; for Google, you have Gemini; and so on. These are mostly accessible from the prompt interface with some degree of customization if you can code.

What I have been doing with my team these days, is thinking about creating structure in your existing system to leverage AI. From coding standards to putting notes and comments in the place where counts. Still in experimental phases, nevertheless with the capability of AI these days, I am confident that they can pick up those clues well. Perhaps even create documentation or summary back for us to read.

Performing analysis

If your company has been advocating the use of AI (Gemini or OpenAI), this is one of the must-use features to improve productivity. Imagine spending hours reading multiple files or documents just to consolidate and summarize information for a meeting later on.

Here are some ideas we have been using it for. AWS month-over-month cost analysis, log file analysis, marketplace analysis, and so on. With tools like NotebookLLM, or OpenAI chat, is quite easy to drop the data as RAG to perform the function above, and start asking questions to get the answer you’re looking for.

Generating ideas and brainstorming

If you have read my previous post, I literally use AI to help me generate winning ideas for a hackathon competition. That pretty much says it all.

If you’re on Miro as well, you guys will love this. Brainstorm by generating sticky note ideas directly from Miro itself. I even use this feature to help me build BBIT (read more about BBIT here) for a problem I am trying to solve at work.

Expanding on decision making reasoning

Lastly, the race between LLM providers has lowered the barrier to getting PHD level thinking capability that used to only be accessible by large companies – now at our fingertips. The reasoning capability is nothing short of impressive.

I have to admit this, their reasoning is better than mine and I do seek advice and validation here. Doing some mental sparring before putting it out for discussion. Things like, process and policy improvement have always been very sticky to work with. With the help of AI, helps me to cover blind spots that can be overlooked.

Takeaway

Before I end on too high of a note. I still advocate for strong fundamental skills before depending AI. Be it, an individual contributor work like coding or managerial work, you’ll need to learn the technical skills to ask and frame the right questions. Don’t just be an empty shell that only knows how to use AI. Otherwise, we might fall out of relevancy in no time.