Gemini can now build your actual files - Docs, Sheets, Slides, PDFs - right inside the chat. Microsoft just dropped a Legal Agent into Word, the tool 95% of lawyers already live in. And Mistral, the Paris-based AI giant most South Africans have never tried, just released the orchestration engine quietly running millions of business workflows a day.

Hi! Welcome to the 47th 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.

_{~8 mins read}

/Gemini /Google /Productivity /Workflow

Here's a workflow most of us know too well. You ask an AI to draft a budget, get a beautifully written reply back, then spend ten minutes copying it into Excel, fixing column widths, fighting with formulas, and quietly questioning your life choices.

Last week, Google quietly deleted that whole step. Gemini can now generate actual files - and we're not talking about plain text dumps. We're talking proper Word documents (.docx), Excel spreadsheets (.xlsx), Google Docs, Sheets, Slides, PDFs, CSVs, and even more obscure formats like LaTeX and RTF. You ask, Gemini builds, you download. Done.

Why this matters more than it sounds: until now, AI assistants have basically been talking heads. They give you words; you do the actual work of putting those words somewhere useful. With this update, Gemini becomes a proper agent that closes the loop - you start in chat and end with a finished file. Brainstorm a quarterly report, ask for it as a Word doc, hit download. Plan a budget, get a fully formatted Excel spreadsheet - formulas already in place - dropped straight into your conversation.

Two big things to know. First: it's available globally to everyone right now, no Pro subscription required. Second: this is exactly the kind of feature that doesn't sound revolutionary in a press release - until you use it once, and realise you can't go back.

We tested it on this week's Curiosity Corner challenge below. The result was a downloadable monthly budget in under a minute.

/Microsoft /Legal /Agents /Vertical

There's a quiet land grab happening in legal AI. Startups like Harvey and Legora have raised hundreds of millions building specialist AI tools for lawyers - separate apps, separate logins, separate workflows. They've done genuinely impressive work. But they all share one inconvenient truth: lawyers don't really want another app. They want their existing app to be smarter.

Microsoft just made exactly that bet. On 30 April, they announced the Legal Agent in Word - an AI agent that lives inside the tool 95% of lawyers already use every single day.

The agent is built to mirror real legal work, not generic AI guesses. It reviews contracts clause by clause against your firm's playbook, flagging anything that doesn't conform. It produces redlines with proper tracked changes (preserving formatting, lists, and tables). It separates new edits from prior revisions, so negotiation history stays intact. And it cites the source language for every suggestion, so reviewers can verify what's been flagged - critical for high-stakes work where trust matters more than speed.

It's currently rolling out to US Frontier Program customers on Word desktop. Slow, careful release - because legal work is high-stakes and confidence matters more than first-mover speed.

The bigger story here is the strategic pattern. Microsoft is using Word's massive distribution to ship vertical AI inside the tool, instead of asking professionals to leave it. The lesson for specialist startups: being purpose-built isn't enough anymore. You need to be genuinely better than what's already living inside the tools your customers refuse to leave.

/Mistral /Enterprise /Workflows /Models

Quick reality check: most South Africans know about three AI companies. ChatGPT (OpenAI). Claude (Anthropic). Gemini (Google). And that's about where the conversation ends.

But there's a much bigger world out there - and one of the biggest names you've probably never used is Mistral AI. Mistral is a Paris-based AI company valued at €11.7 billion (roughly R243 billion), best known for releasing some of the smartest open-weight AI models in the world. While the American giants stay locked behind paid APIs, Mistral has carved out a reputation as Europe's most credible AI player - especially for businesses that want to keep their data on European soil and away from US-controlled cloud providers.

Last week, they launched Workflows, and it's worth understanding why. Workflows is an "orchestration engine" - which is just a fancy way of saying it lets companies chain AI together with their actual business processes. So a bank could build a workflow that reviews a loan application, runs compliance checks, flags anything risky, pauses for human approval at exactly the right moment, then auto-generates the contract. Reliably. Auditably. Every step traceable.

The wild stat: Workflows is already running millions of executions per day for customers like CMA-CGM (one of the world's biggest shipping companies), ASML (the chip-equipment giant), and La Banque Postale.

The bigger picture: the AI race is much bigger than the three names you know. Watch the Europeans - they're playing a different game with very different rules.

In this section, we aim to spotlight an incredible AI tool or use case and guide you on how you can try it.

This week’s challenge: Build your full monthly budget in Gemini - and download the spreadsheet in under a minute

Most of us live in an awkward financial middle ground. We kind of track our spending, sort of know where it goes, and mostly hope the bond gets paid before payday runs out. The problem isn't motivation - it's the setup. Building a proper budget spreadsheet from scratch takes about 20 minutes you don't have, and most of us never get past the "I'll do it this weekend" stage.

This week, we're using Gemini's brand new file generation feature (yes, the same one we covered in News Flash above) to skip the setup entirely. Tell Gemini what you want, watch it build the actual spreadsheet, hit download. The whole thing takes less time than a coffee.

Here's how to try it yourself:

Pro tip: Once it's built, ask follow-up prompts like "Add a tab for tracking the next 6 months" or "Add a chart showing my biggest spending categories." Gemini will edit the file in place. No formulas required, no ten years of Excel experience necessary.

You spoke, we listened. “AI in Enterprise” is here to stay. In this section, we're spotlighting real businesses using AI to solve actual problems.

Block Just Asked the Question Most Companies Are Afraid To: What If We Deleted Middle Management?

Most companies treat AI as a productivity tool. Give everyone a copilot. Cut a few meetings. Save a few hours per week. Call it transformation, ship a press release, move on.

Block - the company behind Square, Cash App, and Afterpay, founded by Jack Dorsey - is doing something far more radical. In a co-authored essay with Sequoia partner Roelof Botha , they argue that AI shouldn't just sit on top of your existing org chart. It should replace large parts of it.

Here's their case. Every company on earth, from the Roman Army (eight soldiers, one decanus) to your local insurance brokerage, is built around the same biological limitation: humans can only effectively manage between three and eight other humans. So organisations add layers. Those layers exist for one reason - to route information up and down the chain. But more layers mean slower information flow, and slower information flow means slower decisions. Two thousand years of organisational design has been an attempt to work around this trade-off. Nobody has actually solved it.

Block's bet is that AI finally can. Their plan is to build what they call a "company world model" - an always-on AI representation of the entire business - paired with a "customer world model" built from the millions of real-time transactions flowing through Square and Cash App every day. Together, these models do what middle managers used to do: route context, surface decisions, allocate resources.

Under this structure, Block normalises its org chart down to just three roles. Individual contributors who actually build things. Directly Responsible Individuals (DRIs) who own a specific problem and have full authority to pull resources from anywhere they need. And player-coaches who still build, but also develop the people around them. No permanent middle layer. No status meetings. No "let me check with my manager and circle back."

The lesson for the rest of us: AI isn't a tool you bolt onto your existing business. It's a question your existing business has to answer. The companies who treat it like a productivity hack will save a few quarters of margin and quietly disappear. The ones who treat it like Block is - as a fundamental rethink of how work gets coordinated - will reshape entire industries. South African businesses watching this from afar should pay close attention. The next decade isn't going to reward incrementalists.

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).

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 👋