In early 2026, I attended the Consumer Electronic Show (CES) in Las Vegas—the annual showcase where businesses demo their current and futuristic products and services. I arrived at the show early and attended some sessions on the current state of AI in entertainment, advertising, and other industries.

A common theme emerged: while companies have spent recent years experimenting with many technologies, 2026 marks the shift to production deployment. My observations at the show confirmed this trend, and the past two weeks have proven it truer than ever.

It seems like every other day I reach for an AI tool to speed up my daily work. Just today I used AI twice to accelerate tasks. First, I was building documentation in Word and wanted to create multiple tables from a rough outline. I prompted Copilot:

Create tables from this text. Add a header row with a black background and white text. Add a new row under the header labeled “Description.” Place the contents inside a single cell under the new cell and add borders to the tables.

As Figure 1 shows, Word handled this fabulously well and created a good document from my outline and prompt.

Figure 1: Word generating tables from an outline.
Figure 1: Word generating tables from an outline.

Next I needed to transform the outline into a Lucid Chart. I turned to Google Gemini with this prompt:

Build a chart in Lucid Chart format with 6 bubbles containing these labels: LoadMetaData, LoadRawData, FilterRawData, MergeIntoGlobalDatasets, ProcessOptouts

Here's where things got interesting. While Word created a document directly, Gemini couldn't generate a Lucid Chart file. Instead, it created a Mermaid.js file—a text-based format for generating charts (Figure 2). I then imported those instructions into Lucid Charts to create my diagram.

Figure 2: Google Gemini generated a Mermaid.js file I could import into Lucid Chart.
Figure 2: Google Gemini generated a Mermaid.js file I could import into Lucid Chart.

After some rearranging, I had exactly what I needed (Figure 3).

Figure 3: My labels rearranged in Lucid Chart.
Figure 3: My labels rearranged in Lucid Chart.

These are two examples from today. Last week, I spent time in Gemini analyzing Snowflake performance data with this prompt:

Can you make recommendations from this Snowflake cluster profile? {cluster_by_keys}: …

Figure 4 shows the result of that prompt.

Figure 4: Gemini analyzed a Snowflake cluster profile and made some good recommendations.
Figure 4: Gemini analyzed a Snowflake cluster profile and made some good recommendations.

The analysis was solid. I cross-referenced it with Snowflake's documentation and reviewed the recommendations carefully before implementing them in our production environment. This is a case where I used my judgment as a developer to validate the AI's suggestions. What I've found is that for programming tasks, AI performs well because the body of knowledge is vast and well-represented in training data.

Over the past five weeks, I have repeatedly turned to AI to speed up tasks I already know how to do—not because I can't do them, but because I want to work faster. I also use AI for exploration. I am working extensively in Python now, and I'll ask AI to show me how to accomplish familiar tasks I already know from other environments. For instance, when I got curious about Python's Flask framework, I asked an AI tool to build a simple CRUD application—not to use it, but to learn from it. It's yet another surprising use case I've discovered.

As CODE Magazine continues featuring substantial AI content, we're doing so by design. We're a magazine by developers for developers. Showing you how to deploy these tools effectively in daily work matters more now than ever—exactly what those CES presenters predicted for 2026.