Keep up to date on the latest products, workflows, apps and models so that you can excel at your work. Curated by Duet.

Stay ahead with the most recent breakthroughs—here’s what’s new and making waves in AI-powered productivity:

YouCam Apps – “New Year, New You” AI Creativity Suite

Perfect Corp’s YouCam Apps just rolled out a wave of generative AI and AR features designed to help everyday users refresh their look, create standout social content, and bring static photos to life for 2026. The update includes an AI Beauty Agent that combines facial analysis with astrology‑inspired insights to recommend goal‑oriented makeup styles, plus upgraded text‑to‑image tools that generate personalized New Year scenes and video features that animate still selfies into short, cinematic greetings optimized for social sharing.

Kikuvi – AI Interviewing Agent & New iOS App

Kikuvi is debuting an AI interviewing agent and a new iOS app that streamline the entire “listening” workflow—from designing questions to collecting responses, summarizing conversations, and extracting structured insights. A new quantitative‑and‑qualitative question feature lets teams mix survey‑style metrics with open‑ended answers in a single flow, so they can measure trends while preserving nuance, saving hours on prep and analysis for employee interviews, customer research, audits, and internal discovery sessions.

Timekettle – AI Interpretation & Translation Engine Selector

Timekettle has announced “breakthroughs” in AI interpretation, centered on a state‑of‑the‑art translation engine selector and next‑generation language features for its real‑time communication hardware and apps. The new engine selector intelligently chooses among multiple translation engines based on language pair and scenario, aiming to improve accuracy and latency for travelers, cross‑border teams, and hybrid events who rely on seamless, live multilingual conversations.

DocuBloom – Automated Document Generation for Workflows

DocuBloom is a newly launched productivity tool that automates PDF and DOCX generation from templates using APIs and integrations, focusing on teams that routinely assemble repeatable documents. By connecting to existing systems and letting users define templates once, it can pull in structured data and generate ready‑to‑send contracts, reports, and invoices, reducing manual copy‑paste work and cutting down on formatting inconsistencies across organizations.

SoundHound AI – Agentic Voice Commerce for Cars and TVs

SoundHound AI is showcasing new agentic voice‑commerce capabilities that turn in‑car and TV interfaces into assistants that can order food, book dinner reservations, and handle payments through conversational commands. The system uses AI agents to orchestrate tasks behind the scenes—connecting to restaurants, payment providers, and other services—so users can complete everyday errands simply by speaking, without pulling out a phone or navigating complex menus while driving or watching TV.

Kikuvi’s AI Interviewing Agent – Rethinking How Teams “Listen”

Kikuvi’s AI interviewing agent and new iOS app are built around a simple but powerful idea: most organizations spend huge amounts of time asking questions, collecting responses, and then manually turning conversations into decisions. Kikuvi is demonstrating an end‑to‑end workflow where AI helps with question design, automates response collection, summarizes interviews, and structures insights so that teams can focus on judgment and action rather than repetitive admin work. The mobile app brings these capabilities into the field, making it easier for managers, researchers, and consultants to capture conversations on the go while still plugging into a central system for analysis.

A standout feature in this release is the unified quantitative‑and‑qualitative question engine, which lets users mix rating scales and multiple‑choice items with open‑ended prompts in a single interview or survey. This means a people leader can collect scores on engagement or workload and immediately pair them with rich narrative context, without juggling multiple tools or spreadsheets. The AI agent then aggregates numeric results while also clustering themes and extracting key quotes, turning raw feedback into dashboards and summaries that are much quicker to review than pages of notes.

For productivity, the key benefit of Kikuvi’s new feature set is time compression: tasks that normally require hours of planning, interviewing, transcribing, and summarizing can be handled with far fewer manual steps. Teams running employee one‑on‑ones, voice‑of‑customer programs, internal audits, or field research can follow the same repeatable flow—design questions once, let the AI agent guide interviews and capture responses, then review structured insights instead of raw transcripts. This not only accelerates cycles but also makes it easier to compare interviews over time, because the AI keeps outputs aligned to consistent structures and themes.

Users who want to get the most from Kikuvi’s AI interviewing agent can adopt a few practical habits: start by writing clear, outcome‑oriented questions so the agent knows what to emphasize, and use a mix of quant and open‑ended prompts so that AI summaries retain both signal and story. During early rollouts, teams can treat Kikuvi as a co‑pilot—spot‑checking summaries and suggested themes—to calibrate trust and ensure the system captures the nuances that matter in their culture or market. Over time, organizations can standardize recurring “listening” workflows—such as quarterly engagement check‑ins or customer win‑loss interviews—so that the AI agent becomes a consistent backbone for feedback‑driven decisions.

Low-Stakes Creation: Build Your First Micro-Tool

For decades, the ability to build software was gated behind a massive wall of syntax, logic gates, and environment configurations. If you wanted to build a simple tool to automate a spreadsheet task, you first had to learn what a variable was, how to install a compiler, and why your terminal was yelling at you about "dependencies."

The Return on Investment just wasn’t there. Why spend 40 hours learning Python just to save 10 minutes of manual data entry?

Generative AI has completely inverted this equation.

We have entered the era of Low-Stakes Creation. The barrier to entry for building software has dropped from "months of study" to "minutes of clear articulation." You no longer need to be a software engineer to build software; you just need to be a Problem Owner who can describe a solution.

This article is your roadmap to building your first "Micro-Tool"—a tiny, single-purpose program designed to solve one specific problem in your life—using AI as your hands.

Part 1: The Philosophy of the Micro-Tool

Before we open a code editor, we need to redefine what "building software" means to you.

When most people think of building an app, they imagine the next Facebook, a complex SaaS platform, or a polished mobile game. That is High-Stakes Creation. It requires architecture, scalability, and security.

Micro-Tools are different.

• They don't need to scale to a million users; they only need to work for you.

• They don't need beautiful user interfaces; a command line or a simple button is fine.

• They don't need to be perfect; they just need to function.

A Micro-Tool is a script that renames 1,000 photos by date. It is a calculator that computes the exact material cost for your specific woodworking hobby. It is a browser extension that hides a specific, annoying element on a website you visit daily.

By lowering the stakes, we remove the fear of failure. If your code breaks, no server goes down. You just ask the AI to fix it, and you try again.

Part 2: Spotting Your Opportunity

The hardest part of coding with AI isn't the code; it's recognizing the problem. We are so used to doing things manually that we often don't realize a script could do it for us.

To find your first Micro-Tool idea, look for "The Groan Moments."

• Repetition: "I have to open these 15 tabs every morning."

• Calculation: "I have to copy this number from here, multiply it by 1.05, and paste it there."

• Formatting: "I have to manually delete the spaces from these 500 lines of text."

The Golden Rule: If you have to do it more than three times, and it follows a strict set of rules, an AI can write a program to do it for you.

For the sake of this guide, let's pretend our problem is: “I have a folder of messy PDF invoices, and I want to rename them all to 'Invoice_Date_Vendor.pdf' automatically.”

Part 3: The AI Workflow

Now, we build. We aren't going to write code; we are going to manage the writing of code. Your role is that of the Product Manager; the AI (ChatGPT, Claude, Gemini, etc.) is your Junior Developer.

Step 1: The Contextual Prompt

The biggest mistake beginners make is being vague ("Write a code to organize files"). You must be hyper-specific.

The Prompt Framework:

1. Role: "You are an expert Python scriptwriter."

2. Goal: "I need a script that scans a specific folder for PDF files."

3. Logic: "It should look at the text inside the PDF to find a date and a vendor name."

4. Output: "Rename the file to YYYY-MM-DD_Vendor.pdf."

5. Constraint: "Explain how to run this on a Mac for a complete beginner."

Step 2: The "black box" realization

The AI will generate a block of code. You do not need to understand every line of this code.

This is a controversial take among purists, but for Low-Stakes Creation, it is essential. You need to understand the logic (what is it trying to do?), but you don't need to memorize the syntax. Read the comments the AI provides. Does it look like it's looking for a date? Good.

Step 3: The Setup

The AI will likely tell you to install a language like Python. This is usually the scariest part for non-coders.

Ask the AI for help here, too.

• Prompt: "I don't know how to run Python. Walk me through installing it and running this script on Windows, step-by-step."

Follow the instructions blindly. Once your environment is set up (which takes about 10 minutes), you rarely have to do it again.

Part 4: The Art of AI Debugging

You run the script. It crashes.

In the old days, this was the moment you quit. The error message Traceback (most recent call last): FileNotFoundError looks terrifying.

In the era of AI, this is just part of the conversation.

The Loop:

1. Copy the error message exactly as it appears.

2. Paste it into the chat with the AI.

3. Add context: "I got this error when I ran the script. What does it mean and how do I fix it?"

The AI will apologize, explain the mistake (maybe it looked for the wrong date format), and give you corrected code. You copy, paste, and run it again.

This is the core skill of modern creation: Resilience in the face of error messages. You are iterating. You are not failing; you are refining.

Part 5: From Script to Tool

Once the code works, you have a few options to make it more "tool-like."

If you built a Python script, you currently have to type a command in a terminal to run it. That’s fine, but we can do better.

The Interface Upgrade: Ask the AI: "This script works great. Can you wrap it in a simple Graphical User Interface (GUI) so I can just double-click an icon and select the folder?"

The AI will likely suggest a library like Tkinter (for Python) and rewrite the code to include a window, a "Select Folder" button, and a "Run" button. Suddenly, you aren't looking at matrix code; you are looking at a real app.

Alternatively: The Web App approach If your tool doesn't need to access your local files (e.g., a unit converter or a text generator), ask the AI to write it as a single HTML file.

• Prompt: "Rewrite this logic as a single HTML file with embedded JavaScript. I want to run it in my browser."

• Result: You get a file you can double-click to open in Chrome. No installation required.

Part 6: Example Projects to Start With

If you are struggling to find an idea, start with one of these "Hello World" projects of the AI age. These are tested, simple, and satisfying.

1. The "Clean Up" Script

• Goal: A script that looks at your "Downloads" folder and moves images to an "Images" folder, documents to "Docs," and installers to "Installers."

• Why: Immediate visual gratification.

2. The Text Sanitizer

• Goal: A simple web page with two boxes. You paste messy text (with weird line breaks or formatting) in the left, and clean, plain text appears in the right.

• Why: Useful for anyone who writes emails or reports.

3. The Focus Timer

• Goal: A small desktop window that counts down from 25 minutes and plays a sound when finished.

• Why: Teaches you how to make sound and timers work.

Conclusion: The Shift from Consumer to Creator

The most powerful result of building your first Micro-Tool isn't the tool itself—it’s the psychological shift that occurs in your brain.

For your entire digital life, you have been a Consumer of software. You waited for Apple, Microsoft, or Google to build the features you wanted. If a button didn't exist, you couldn't do the thing.

When you build your first Micro-Tool, you become a Creator. You realize that the digital world is malleable. You realize that if you want a button that sorts your emails by "sadness," you can (theoretically) build it.

The barrier is gone. The cost is zero. The stakes are low.

Open a chat window. Type a prompt. Build your tool

Partner Spotlight: Duet Display

Duet Display turns your iPad, Android tablet, Mac, or PC into an extra high‑performance display and input surface, helping you declutter your desktop while giving AI and productivity tools more room. By extending your screen, you can park dashboards, prompt windows, or long documents on one device while working in your primary app on another, making multitasking with AI assistants, research tabs, and creative tools feel more natural. To explore setups for work, study, or creative projects, visit Duet Display.

Stay productive, stay curious—see you next week with more AI breakthroughs!