👋 Tomorrow’s Tech, Delivered Today
Hi! Welcome to the 33rd edition of the TomorrowToday newsletter.
We’re here to decode the AI chaos so you don't have to. Think of us as your friendly neighbourhood tech translators - we cut through the chaos, translate the jargon, and spotlight new AI tools that matter for founders, builders, and curious minds.
Buckle up, because the future's moving fast and we're here to make sure you don't get left behind! ⚡
If you enjoyed today’s newsletter, please forward it to a friend & subscribe by following this link.
~10 mins read
🗞️ News Flash
🤖 Clawdbot: Your 24/7 AI Employee That Lives in WhatsApp
/Automation /Agents /OpenSource /Productivity
Clawdbot is absolutely blowing up on X right now, and for good reason - it's the first AI assistant that truly feels like hiring a full-time employee you can message on WhatsApp.
Imagine this: you send a WhatsApp saying "book me a table for two at that Italian place for Friday at 7pm" and it just... handles it. Or "summarise my inbox and draft replies to anything urgent." Or "monitor the rand-dollar exchange rate and ping me if it crosses R18.50." That's Clawdbot.
This open-source framework transforms AI models into persistent personal agents that run locally on your device (Mac, Linux, Windows) and integrate seamlessly with WhatsApp, Telegram, Slack, Discord, Signal, and iMessage. Unlike typical chatbots that forget everything between conversations, Clawdbot maintains context and can execute complex multi-step tasks autonomously.
What makes it special? Three things:
It meets you where you already are. Your AI lives in the messaging apps you use daily. No switching apps, no special interfaces - just message your AI assistant like you'd message a colleague.
It actually does things with high success rates. Clawdbot can browse the web, send messages, manage files, run code, use Claude Clode, book reservations, and execute terminal commands. It's not just answering questions - it's taking action.
It automates recurring tasks. Set up a task once, and it runs automatically based on time triggers, situational triggers, or action triggers. "If I get a message from a customer mentioning X, respond with Y" - that level of automation.
Released in early January 2026, Clawdbot gained viral traction by late January with demos showcasing phone-based control and real-world workflows. One user set it up to monitor their X feed and send daily breakdowns of trending topics, recommended articles, and even memes. Another automated their morning briefing with schedule, weather alerts, and motivational quotes.
But here's the catch: Clawdbot is not plug-and-play. Setup requires terminal commands, environment variables, API keys, and understanding cron syntax. Security implications are real - you're giving an AI access to send messages, read files, and execute code on your behalf. Tech experts recommend running it on a virtual machine first, or if you have the budget, buying a dedicated Mac mini for it.
For technical users and tinkerers, this is the ChatGPT moment for personal autonomous agents. For everyone else? Wait for the consumer-friendly version that's inevitably coming.
Real-life use case: A digital marketer set up Clawdbot to triage their email inbox every morning, categorising messages by priority, drafting responses to common queries, and flagging anything requiring personal attention. The result? What used to take 45 minutes of manual sorting now happens automatically before they've had their first coffee. They also configured calendar management - Clawdbot sends daily briefings with weather integration and can schedule meetings by simply messaging "find a time next week for a 30-minute call with John."
🧑💻 Claude in Excel: The World's Most-Used Software Just Got Smarter
/Claude /Enterprise /Productivity
Excel runs the global financial system. Seriously - if Excel stopped working tomorrow, the collective spreadsheets of Fortune 500 companies would grind to a halt. And now Claude lives inside it.
Claude in Excel is now generally available as a beta research preview for all Claude Pro, Max, Team and Enterprise customers. This isn't a chatbot sitting next to your spreadsheet - it's Claude understanding your entire workbook, from nested formulas to multi-tab dependencies, with cell-level citations for every explanation.
Press Control+Option+C (Mac) or Control+Alt+C (Windows) and Claude opens inside Excel. Ask about any cell, any formula, any calculation that flows across tabs. "What assumptions drive the revenue forecast in Q3?" Claude traces it back with citations. "Why is the NPV calculation returning #VALUE! in cell G145?" Claude explains what broke and how to fix it.
But here's where it gets powerful: Claude can test scenarios without breaking your formulas. Want to see what happens if revenue growth increases by 2%? Claude updates assumptions across your entire model while preserving all dependencies, highlighting every change with explanations for full transparency.
The distribution story here is masterclass-level. Claude now powers Microsoft 365 Copilot's Agent Mode for spreadsheets - Microsoft pays AWS for access despite AWS being a direct cloud competitor. Why? Internal testing showed Claude outperforms OpenAI on Excel financial functions. GitHub Copilot users now "primarily rely on Claude Sonnet 4" when auto-select is enabled.
The numbers tell the story: Anthropic commands 32% of enterprise LLM deployments versus OpenAI's 25%. Two weeks ago, Claude became enabled by default in Microsoft 365 Copilot for commercial tenants worldwide. Anthropic figured out what OpenAI missed - enterprise AI adoption follows distribution, not benchmarks. And 70% of Fortune 500 companies are already Microsoft 365 customers.
The best model isn't the one with the highest scores. It's the one running inside Excel when the analyst builds their quarterly report.
Real-life use case: A financial analyst at a PE firm was debugging a complex DCF model with circular reference errors that had been breaking calculations for weeks. Traditional troubleshooting meant manually tracing formulas across 15 worksheets. With Claude in Excel, they asked, "trace the circular reference in my terminal value calculation", and Claude identified the exact cells causing the loop within seconds, explaining the logical error and suggesting the fix. What would've taken hours of manual debugging was resolved in under two minutes.
💬 Gemini Offers Free SAT Practice: AI Becomes Your Personal Tutor
/Gemini /Education /Accessibility
Google just made SAT prep accessible to anyone with a smartphone - for free.
Gemini now offers full-length SAT practice tests with instant AI-powered feedback and personalised study plans. Developed in collaboration with The Princeton Review, these tests mirror real exam conditions as closely as possible. Complete a test, get immediate performance breakdowns, ask the AI to explain any incorrect answers in detail, and Gemini generates a customised study plan targeting your specific weak spots.
This is crazy because traditional SAT prep is expensive - think private tutors, bootcamps, and giant test prep books that cost hundreds of rands. Making preparation accessible and flexible has obvious appeal, especially for students in under-resourced schools or those studying late at night when traditional help isn't available.
The speed is what makes this different. The shorter the gap between performance and correction, the better students remember what they studied. Gemini walks students through missed answers in the moment, reinforcing learning when it matters most.
For South African students, this signals something bigger: personalised learning pathways in almost any subject are coming. While Gemini's SAT focus targets the US market initially, the underlying technology - AI that adapts to individual learning styles, identifies knowledge gaps, and provides targeted support - will expand to other standardised tests and subjects. Students in Cape Town or Johannesburg will soon have access to the same quality of AI-powered tutoring as students in New York or London.
This also reveals an important pattern in AI adoption: distribution matters more than the best model. Companies with ethical AI practices are positioning themselves through distribution channels. Google embeds Gemini in education tools, Anthropic gets Claude into Microsoft 365, OpenAI partners with schools - the real competitive advantage isn't just building the smartest AI, it's getting it into the hands of users through platforms they already use.
If the SAT continues to exist, equal access to preparation is essential. By putting high-quality practice in anyone's hands, Gemini is democratising test prep at scale.
Real-life use case: A student in a rural South African school with limited access to SAT prep resources uses Gemini to take full-length practice tests on their phone. After each test, Gemini identifies that they consistently struggle with geometry problems involving circles and generates a personalised study plan focusing specifically on circle theorems, arc lengths, and sector areas. The AI explains each missed question in multiple ways until the concept clicks, then provides similar practice problems to reinforce learning. What would've required expensive tutors or specialised books becomes accessible through a free AI tutor available 24/7.
💡 Curiosity Corner
In this section, we aim to spotlight an incredible AI tool or use case and guide you on how you can try it.
Professional animated product videos used to cost thousands of rands and required video editing expertise. Not anymore. This video was made in 10 minutes using AI.
Remotion is a framework that lets you create videos using code - think terminal animations, 3D transforms, spring physics, the works. Combined with Claude Code, you can now build animated promos by just describing what you want. If you can type a sentence, you can do this.
Here's how to set it up:
Step 1: Install the Remotion skill in Claude Code
Open Claude Code inside VS Code or Cursor and run:
npx skills add remotion-dev/skills
This teaches Claude how to write video code - like giving Claude a cheat sheet for Remotion's syntax.
Step 2: Create your video project
Run:
npx create-video@latest
Pick the "Hello World" template when prompted. Remotion Studio opens at localhost:3001 - this is where you preview animations in real-time.
Cleanup step: Delete config files for other AI tools (Cursor, Gemini, OpenCode) and keep only agent.claude.md so Claude isn't confused.
Step 3: Point Claude to your product
Tell Claude about your product:
"Look at /src/pages/index.tsx. What components would work as animated videos?"
Or point it to your live site:
"Go to myproduct.com. What sections would make good promos?"
Step 4: Build your animation iteratively
Start basic:
"Create a macOS terminal window, 1280x1000px, light theme. Blinking cursor."
Then iterate:
"Add typewriter animation typing 'npx skills add remotion-dev/skills'"
"Add 3D rotation, 10 degrees on X and Y axes"
"Rotate Y axis from 10 to -10 degrees over video length"
"Slide from bottom with spring animation, no bounce"
Claude can even run actual commands and add real output to the animation. The key is one change at a time - check it, then next change.
Step 5: Add branding and polish
"Create a new composition with our logo, plus sign, and Claude Code logo"
"When typing finishes, flip the terminal toward the camera"
Step 6: Export your video
When you're happy with the preview:
npx remotion render MyComposition --output promo.mp4
Replace "MyComposition" with your composition name (visible in Remotion Studio's left sidebar). This generates an MP4 you can upload anywhere.
Pro Tips:
Tell Claude your video has three parts (intro, main content, outro) before starting - helps organise the code better
Drop images/logos in the
/publicfolder and reference them with exact pathsBudget time for iterations - each takes 2-3 minutes while Claude writes code
What works well: terminals, screenshots, text animations, numbers counting up
What struggles: character animations, organic shapes, frame-by-frame work
🏢 AI in Enterprise
In this section, we're spotlighting real businesses using AI to solve actual problems.
This week: BBVA - How a Global Bank Automated 9,000 Legal Queries 🏦
BBVA is a Spanish global bank operating across Europe, Mexico, South America, Turkey, and the US, serving tens of millions of customers with 125,000+ employees.
The problem: BBVA's legal department in Mexico was drowning in repetitive queries. The team received over 40,000 legal questions annually from branch managers asking about corporate signatory authority - a compliance process called "bastanteo" that confirms whether company representatives have authority to sign contracts, open accounts, or issue credit. Specialist legal teams had to respond to each query manually, creating delays and consuming scarce legal capacity on routine work.
The solution: BBVA built a generative AI chatbot using ChatGPT Enterprise that provides instant access to standardised, pre-validated legal FAQs and documentation guidance. The content was developed and reviewed by BBVA's legal team to ensure accuracy and compliance, but the chatbot eliminated manual handling of routine inquiries.
The results:
9,000+ queries automated annually - removing the manual bottleneck entirely
3 full-time employees redeployed to higher-value work, producing over 11,000 bastanteos per year
26% of Legal Services division's annual savings target delivered by this single use case
Query handling time reduced from 7.5 minutes to 1 minute in Peru deployment (~80% reduction)
BBVA didn't stop there. They've now deployed over 11,000 ChatGPT Enterprise licenses across their organisation, with employees creating 3,000+ custom GPTs. The bank reports that licensed employees save an average of 2.8 hours per week by automating routine tasks.
The lesson: AI creates the most value when it eliminates specialised bottlenecks in compliance-heavy processes. BBVA didn't try to replace their lawyers - they freed them from repetitive queries so they could focus on complex legal work that actually requires human expertise. The chatbot handles the "have you checked the documentation" questions, while lawyers tackle the strategic challenges.
This is what successful enterprise AI looks like - not flashy demos, but boring efficiency gains that compound across thousands of employees and deliver measurable bottom-line impact.
📜 AI Dictionary
AI is full of jargon, and we’re here to decode it. Each week, we’ll give you a plain-English definition of a buzzy term you’ve probably seen (but never fully understood).
Persistent Memory - noun
We’d like to ask a favour 🤝
If this email lands up in your Promotional or Spam folder, please move it to your Primary inbox. We’re working hard to bring you the best content weekly, and your support is truly appreciated. Thanks!
Thanks for reading TomorrowToday! We’d love to hear from you:
➡️ What would you like us to cover next?
➡️ Have a tool or topic we should feature?
We’re building this with (and for) you. 🚀
See you next Tuesday 👋


