My favorite AI tools and tips for each task (January 2026 edition)

AI with ALLIE

The professional’s guide to quick AI bites for your personal life, work life, and beyond.

I’ve been refining my AI toolkit for the last three years. And the biggest unlock is not optimizing my model, it’s optimizing my whole system. It was realizing that different tasks need different tools, and AI labs are moving at breakneck speeds in slightly different directions. Save this breakdown. Reference it often. Instead of asking “what’s the best AI?”, start asking the question “best for what?”

If you're using the same AI tool for every task, you're leaving performance on the table.

I get asked all the time: "What's the best AI model?"

I get it. You want the quick answer. But in my opinion, if you’re trying to really uplevel your AI work and maximize your potential, it’s the wrong question.

The right question is: "What's the best tool for THIS specific task?" And you might even go so far as to say, “Is it worth maximizing on this one task to find the perfect tool, or can we use something ‘good enough’?”

I've spent the last 3 years testing every major model and tool across different workflows. Not fancy benchmarks like the length of human task AI can take on or how much money it can make if it ran its own vending machine, though I appreciate those as indicators. But real work. If you see a model or tool on this list that you haven’t yet tried, this is a great time to download/try it.

Here's my full AI toolkit, broken down by use case. List of prices, availability, and direct links are all at the bottom.

1. Go-To Starter Model: Claude Opus 4.5

My team and I all agree: Claude Opus 4.5 just "gets it."

I continue to be stunned by its capabilities. And it passes the vibe check with flying colors. It matches my writing tone and structure and style at about 85-93%, making it insanely easy to spar with it on posts or client reports. I upgraded to Claude Max ($100/$200 per mo, depending on your usage) just to use it more.

(Budgeting note: I started with the $100/mo and I only hit my limit when I’m really deep into a build or asking Claude to do 100 things direct in my browser. If you’re a non-coder, just start with $100/mo and see how you do.)

It's intelligent, efficient, and great for coding, agents, and everyday tasks like deep research and working with slides and spreadsheets.

The downsides: it runs out of context fast and jumps to action (like making a slide deck) when I don't want it to. But for my daily driver, nothing beats it. Google Gemini 3 Pro and GPT-5.2 Pro/Thinking also get my ‘good enough’ score, in case you have standardized your life or work on Google’s Gemini or OpenAI’s ChatGPT. And pay close attention to all the tricky words in these titles; Opus 4.5 is different from Sonnet 4.5, and GPT-5.2 Thinking is different from GPT-5.2. The naming conventions have fully entered their pharmaceutical drug phase.

Despite loving Claude Opus 4.5, my team still maintains our ChatGPT business accounts as well as personal Google Gemini accounts, and we run API keys across all three providers.

2. Go-To Builder: Claude Code

Give me a permanent soapbox with the words “Allie ❤️ Claude Code” carved into it because I have not shut up about Claude Code for the last month.

I feel like a powerful conductor, wielding multiple terminals, building anything that pops into my mind, morphing my computer into the right interface at the right time. Mostly a terminal user, but will eventually go into Chrome more.

With Claude Code, I have reorganized my entire computer file system, diagnosed a lag issue, built 2 internal apps that saved me 7 hours within the first week, kicked off dozens of agents, built automated workflows between Slack + Notion + Make + GSuite + Whisper + Nano Banana Pro, generated slideshows in minutes based off of massive client folders, even analyzed one year’s worth of data from my personal scale to build out a full health projection model for my skeletal muscle percentage (hey, I wanna make sure that weightlifting is working!).

If you want a 5min Claude Code quickstart guide, I got you. And If you want a deep dive 45min Claude Code tutorial, I also got you. And if you’re a business user that wants a lightweight version to become a little more powerful, I also made a Claude Cowork demo.

3. Long Tasks With Lots of Context: Google Gemini + NotebookLM

When you have a long task (like a conversation you really need to dig into for 2 hours, or 30 different documents), you have to start worrying about ‘context length’, how much an AI can hold in memory about the conversation, which I have covered in a previous newsletter.

I remain obsessed with NotebookLM. Love that it takes in direct links so I don’t have to copy and paste anything. And the ability to click literally just one button and turn your entire collection of documents into the beautifully designed infographic or perfectly acceptable slide deck for a presentation, it’s just chef’s kiss.

I regularly upload YouTube links to those 2h or 4h long podcasts and "talk" to the podcast guest or debate them. Gemini can analyze up to 1,500 pages of text or 30K lines of code simultaneously with its 1 million token context window. It’s heaven.

200K tokens in Claude and ChatGPT just runs out too fast for heavy research sessions, even when Claude compacts (and the compacting is the worst). Gemini 3 Pro is state-of-the-art on reasoning with unprecedented depth and nuance. When I need to go deep on massive amounts of information, this is where I live. And I’ll often bounce assets back and forth between Gemini, Claude, and ChatGPT.

4. Image Anything: Nano Banana Pro

Holy cow.

I made a fun slide deck two weeks ago for my team and used Nano Banana Pro to make every single dynamic slide background. Infographics from Nano Banana Pro are the best, even with a few text issues.

I would not call it enterprise-grade due to text inconsistencies. But for creative work, internal presentations, visual brainstorming — it's unmatched.

5. Video: Tie Between Sora and Veo 3

Google is really just crushing the visual side of things. Veo 3 was announced at Google I/O 2025 and represents Google's most ambitious move into cinematic quality video generation, complete with native audio.

Veo 3 is slightly ahead of Sora in my book. It shows clear improvements in prompt understanding, scene consistency, and physical realism. Camera movements feel intentional, lighting looks cinematic (to my untrained eye, at least), and audio reacts naturally to the environment.

Video generation is not a common use case for me, and I suspect not a common use case for most of you. Really just to make memes. I prioritize convenience very highly here, and the Sora app is easy. I don't scroll the Sora app, however. Used it for 3 days and stopped. Maybe I'm not the right demographic. Or maybe my attention span has been permanently rewired by 7-second TikToks and I can no longer enjoy beautiful things.

🚨 Power use case alert: screen record yourself doing a task, upload it to Google Gemini, and ask for an entire workflow guide based on that task that you can immediately pass to a teammate to help them onboard. Video understanding is dreadfully underused and underappreciated. Gemini is currently the only app out of the large players that has this as a feature.

6. Voice Anything: Whisper API

Whisper API is my first pick, followed by Wispr Flow and Otter AI tied for second. Gemini a distant third. And Anthropic fading into the deep darkness in fourth and forgotten place. Siri buried under the floorboards.

• If I’m building an app that needs dictation: OpenAI Whisper.

• If I want to chat live back and forth with an AI for hands-free fun: ChatGPT.

• If I’m dictating at my desk or on my phone: Wispr Flow.

• If I’m dictating on a walk or recording a meeting: Otter AI.

If you're not using voice-to-text to speed up your workflows, you're working harder than you need to. I draft at least 70% of my first-pass content this way.

7. Talking to Multiple Documents: NotebookLM

NotebookLM wins again here. Upload your docs, your links, your videos — and actually have a conversation with them. Stop copy-pasting between tabs.

A highlight for me is in how it handles source attribution. Ask a question, and it shows you exactly which document (and which section) the answer came from. Very helpful for research use cases. For example, I can upload a company's last three earnings calls, their investor deck, and two competitor analyses, then ask "what's their stated AI strategy and how does it compare to their peers?" and get a cited answer, with key quotes that I can verify, in seconds.

I've tried building similar workflows in Claude and ChatGPT with file uploads. They work, but the citation tracking isn't as clean. For research-heavy synthesis across 5+ sources, NotebookLM is purpose-built for the job.

8. Efficiency Hacks: Claude Skills + Wispr Flow + Context Vault

If you keep re-explaining the same descriptions, wants, needs, workflows, preferences, or analysis processes to your AI, I want you to allocate real time to fix the situation.

For repeated tasks, I create a Claude Skill (or at least a GPT inside of ChatGPT, but Claude Skills are more powerful because they more easily include code).

And no, you don't need to know how to code to create a Claude Skill. A skill can literally just be a zip file with some structured information. Basically a folder with some files, some directions, and some references. Here is my Claude Skills tutorial, complete with prompts and use cases. I've built Claude skills for:

• Visual brand guidelines

• AI-first use case generation

• AI-first index assessment analysis

• Generating images in my style or throwing something in Notion with the right structure

Combine that with Wispr Flow for voice input and context documents that live on your computer, dedicated to different aspects of your life (business, career, family, personal, health, etc.), and you've got a system that actually adapts to how YOU work.

The Bottom Line

Stop trying to find "the one" AI model. Build a stack.

Match the tool to the task. Save this list. Start experimenting.

Let me know which tools you're using differently or which use cases you’re leaning heavily into — I'm always updating my stack.

✅ AI Pricing and Availability

Pricing tiers subject to change.

Stay curious, stay informed,

Allie

Inside AI-First Academy, you'll learn how to rebuild you work from the ground up. We cover the full progression: from AI Basics to custom assistants to full workflow automation. This is the same training other professionals at Amazon, Apple, and Salesforce are using to stop dabbling and start building.

You'll learn how to:

✔ Write prompts that produce usable, business-ready outputs (not generic fluff)

✔ Build custom GPTs and reusable AI systems that handle your recurring tasks

✔ Automate research, reports, meeting synthesis, and multi-step workflows

✔ Prototype real tools and dashboards - no code required

✔ Deploy AI agents that work for you, not just with you

✔ Navigate the best of Claude, ChatGPT, and the latest agentic features with instant access to our 90-minute Agentic AI Workshop

(Want to upskill your whole team? Get a company-wide license)

Feedback is a Gift

I would love to know what you thought of this newsletter and any feedback you have for me. Do you have a favorite part? Wish I would change something? Felt confused? Please reply and share your thoughts or just take the poll below so I can continue to improve and deliver value for you all.