• Everyday AI
  • Posts
  • IMB Think 2025: AI Updates that could shape Enterprise Work

IMB Think 2025: AI Updates that could shape Enterprise Work

Google’s Gemini 2.5 Pro I/O Edition takes top spot, Amazon working on new AI code generator, Mistral unveils new AI model and more!

Author

Everyday AI
May 07, 2025

Outsmart The Future

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: We’re breaking you the latest AI updates on enterprise work directly from IBM Think 2025. Give it a listen.

🕵️‍♂️ Fresh Finds: Amazon unveils robot with sense of touch, Meta’s plan to automate the ad industry and Google iOS app gets new AI feature. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Google’s Gemini 2.5 Pro I/O Edition takes top spot, Amazon working on new AI code generator and Mistral unveils new AI model. For that and more, read on for Byte Sized News.

🧠 Learn & Leveraging AI: OpenAI’s new o3 and o4 models are powerful. But does that make them the best? Here’s everything you need to know. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about OpenAI acquiring Windsurf, Microsoft unveiling Copilot+ PC AI updates and Musk pushing forward with OpenAI lawsuit. Check it here!

 IBM Think 2025: AI Updates that could shape Enterprise Work 💼

One of the biggest downsides of consumer AI?

It doesn't have up-to-date access to your enterprise data.

Even as frontier labs work tirelessly to connect and integrate AI chatbots with your data, we're a far way off from that happening.

Unless you're using a platform like IBM's watsonx. And if you are using watsonx, your go-to enterprise AI platform just got a TON more powerful.

IBM just unveiled updates across its watson ecosystem at its Think 2025 conference. We've been here covering every step of it, so we're jumping into what you need to know.

Also on the pod today:

• Build Your Own AI Agents Features 🔨
• IBM and Salesforce AI Collaboration 🤝
• New Agent Catalog with 50+ Agents 👥

It’ll be worth your 51 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – Brev is an AI agent that tracks your OKRs, there is an AI Word editor and tracker and Sagehood provides real-time market intelligence.

Amazon – Amazon has unveiled Vulcan, its first robot with a sense of touch.

Meta – At a recent conference, Mark Zuckerberg laid out his plan to automate the ad industry with a black box AI tool

Google – Google’s iOS app gets a new Simplicity feature that uses AI to simplify jargon.

Future of Work – IBM’s CEO says that HR layoffs due to AI led to more investments in other roles.

Social Media - LinkedIn is launching a new AI search tool that lets you find job listings by describing it.

AI in Media – Netflix is launching a generative AI-powered search tool.

AI in Society – A man who was shot dead was recreated using AI video to deliver a message to his killer in court.

AI in Sports – NBC is using Jim Fagan’s AI-generated voice for NBA games.

NBA player Russell Westbrook is launching an AI-enabled funeral planning startup.

1. Google’s Gemini 2.5 Pro I/O Edition Takes the AI Coding Crown 👑

Google DeepMind just dropped Gemini 2.5 Pro “I/O” edition, claiming the top spot in AI coding models with a major leap in performance and usability, according to VentureBeat. This update, rolling out ahead of Google’s I/O conference, boosts reliability in code generation and interface design, pushing it past Anthropic’s Claude 3.7 Sonnet on a key web app development leaderboard.

Developers report it handles complex tasks with senior-level finesse and drastically cuts errors in real-world applications, signaling a shift toward more dependable AI tools for coding and design workflows.

2. Amazon’s New AI Code Generator “Kiro” in the Works 🧑‍💻️️

Amazon Web Services is reportedly developing “Kiro,” a fresh AI-driven code generation tool that promises to churn out code almost instantly using AI agents, according to Business Insider. Beyond just writing code, Kiro aims to create technical design docs, spot bugs, and optimize software, potentially speeding up developer workflows significantly.

While AWS already offers Q Developer, similar to GitHub Copilot, Kiro’s broader multimodal and third-party integrations could shake up AI coding assistance when it possibly launches later this year.

3. Mistral Medium 3 Debuts as a Cost-Effective AI Powerhouse ⚡

French startup Mistral just dropped its new AI model, Mistral Medium 3, promising top-tier performance at a fraction of the cost of competitors like Anthropic’s Claude Sonnet 3.7. Designed for efficiency without sacrificing muscle, it excels in coding, STEM, and multimodal tasks, and can run on modest GPU setups or major cloud platforms like AWS, Azure, and Google Vertex AI.

This move is especially timely as businesses across finance, energy, and healthcare seek affordable, powerful AI to automate workflows and analyze complex data.

4. Hugging Face Launches Open Computer Agent 💻

Hugging Face has unveiled Open Computer Agent, a cloud-based AI tool that navigates a Linux virtual machine to complete user tasks like browsing or searching, though it currently struggles with complex commands and CAPTCHA challenges.

While its speed and accuracy lag behind expectations, this release highlights the growing accessibility of AI agents powered by open models, signaling a shift toward more affordable AI-driven automation.

5. OpenAI Launches Global AI Push to Counter China 🌎

OpenAI just announced "OpenAI for Countries," an initiative to help nations build local AI infrastructure focused on democratic values, aiming to provide tailored ChatGPT versions especially for health care and education. This move, revealed ahead of CEO Sam Altman’s Senate testimony on AI competition, aligns with U.S. efforts to challenge China’s growing AI influence.

Countries partnering in the program will co-fund data centers and gain sovereignty over their data while fostering local AI startups.

6. IBM CEO Urges Boost in Federal AI Research Funding 📢

In the face of proposed deep federal cuts to AI and tech research budgets under the Trump administration, IBM CEO Arvind Krishna is calling for increased government investment in AI R&D, emphasizing its critical role in economic growth and U.S. competitiveness. Recent budget slashes have targeted key agencies like the NSF and the Directorate for Technology, Innovation and Partnerships, threatening billions in AI research grants and programs like the CHIPS Act that support semiconductor production.

Krishna revealed IBM has already lost $100 million from canceled federal contracts but remains optimistic that R&D funding for AI, quantum computing, and semiconductors will rebound within a year.

7. OpenAI Plans to Cut Microsoft Revenue Share by 2030 ✂️

OpenAI is signaling a major shift in its financial ties with Microsoft, aiming to reduce the revenue share it pays from 20% to just 10% by the end of this decade, according to The Information. This move coincides with OpenAI’s pivot toward restructuring its for-profit arm as a public benefit corporation, while still under nonprofit control.

Microsoft, having invested tens of billions and holding exclusive rights to OpenAI’s IP on Azure, has yet to approve the new structure, highlighting ongoing negotiations over protecting its multi-billion-dollar stake.

🦾How You Can Leverage:

Consumer AI is cool and all. But enterprise AI? That's where the REAL money moves happen.

What did IBM just drop at their Think 2025 conference? Only their BIGGEST watsonx updates of the year. 

No big deal.

That's why we HAD to dive deep on IBM's watsonx updates. 

Want the full breakdown? We've got you covered on today’s episode. 

(We’ll have a FULL recap and more in-depth insights next week after we’ve caught up on sleep and have had more time to take everything in.) 

But if you need the quick rundown, here's the no-BS guide to what matters from IBM Think 2025 for (non-technical) business leaders. 

Let's goooooooo.

1 – Build-Your-Own Agents: Zero Code Required 🛠️

Watson x Orchestrate now lets you create custom AI agents in under five minutes without writing a single line of code. The new drag-and-drop interface means anyone can build enterprise-grade AI tools. The technical background is optional.

Watch your agent's reasoning process in real-time while you build, giving you unprecedented visibility into how it accesses your data and makes decisions. We’re big fans of that. 

What it means: 

This is a "Canva moment" for enterprise agentic AI – complex tech finally accessible to everyone.

For watsonx clients, spread AI innovation across departments without hiring more techies. Business units will create their own solutions instead of begging IT for help.

The companies that empower business users while maintaining smart guardrails will leave their competitors in the dust. Time to develop those governance frameworks!

2 – Pre-Built Domain Agents: Ready To Work Day One ✅

IBM's domain-specific agents arrive pre-trained for HR, sales, and procurement tasks. The Better Business Bureau already pocketed $1.5M in annual savings with these agents handling routine processes. A+ buddies. 

Imagine your HR team finally free from the endless cycle of PTO requests and policy questions and 42 ping-ponging emails—these agents handle it all while maintaining access to your company's most current data.

What it means: 

AI is getting specialized, and the ROI is getting real.

Watson clients should identify their highest-volume, lowest-complexity workflows NOW. These are prime candidates for immediate automation.

For the industry at large, it's time to rethink workforce planning. Those administrative tasks eating up valuable human hours? They're about to disappear. Start measuring your baseline metrics today so you can brag about the savings tomorrow.

3 – Agent Catalog: Your AI Arsenal Expands 📚

The new agent catalog delivers 50+ ready-made agents that integrate with your existing enterprise stack. Search, filter, and deploy these purpose-built tools faster than your IT department can say "ticket submitted."

Each agent connects seamlessly to your enterprise data, making the days of disconnected AI tools ancient history.

What it means: 

Enterprise software is going "app store" – compose your toolkit rather than buying one massive platform.

For Watson clients: audit your tech stack, then map agents to specific pain points. Don't just deploy randomly because they look cool (though some definitely do).

The big picture? We're moving toward assembling AI capabilities from specialized providers rather than relying on one vendor. Innovation cycles will speed up dramatically, but integration headaches could follow. Plan accordingly y’all. 

4 – IBM + Salesforce: The Power Couple 🤝

This collaboration brings a supercharged sales prospecting agent that finds leads, pulls contact details, and drafts outreach—all within your Salesforce environment.

It's the perfect marriage of IBM's AI prowess and Salesforce's CRM dominance. Though if you're already using Agent Force, prepare for some interesting conversations about which platform deserves your attention.

What it means: 

The age of AI partnerships is here – even between former competitors.

If you're using both platforms, map out the unique strengths of each offering. Don't double-pay for the same capabilities!

The pattern emerging across the industry: vertical AI expertise (IBM) + horizontal platform reach (Salesforce) = the winning formula. Data interoperability is finally trumping walled gardens. It's about time!

5 – Oracle Cloud gets Watson: Giants Join Forces 🏋

Watson x Orchestrate arrives on Oracle Cloud, enabling AI agents to operate seamlessly across Oracle and non-Oracle applications. Starting with HR workflows and expanding from there, this partnership lets you deploy AI exactly where your data already lives.

What it means: AI is coming to YOUR data, not the other way around.

For the Oracle/Watson crowd, this removes a massive adoption barrier – no more complex data migrations required!

The industry is finally accepting reality: enterprise data environments are fragmented and hybrid. Develop a strategy that identifies your authoritative data sources, then deploy AI agents at those points. Forget those massive data lake projects that never quite work.

6 – Granite 4.0 Tiny: Small Model, Big Impact 🐜

IBM's new 7-billion parameter model delivers performance comparable to much larger models while running on a $350 consumer GPU. This open-source release means accessible, affordable AI without the data center price tag.

Consider it a preview of the full Granite 4.0 family dropping later this summer.

What it means: 

Welcome to AI's commoditization phase – enterprise capabilities on commodity hardware.

Watson clients can now experiment with on-premises deployment for sensitive workflows without breaking the bank.

The bigger trend? AI is following the classic tech adoption curve – what once required specialized infrastructure now runs on everyday hardware. This will supercharge edge AI applications especially. The democratization is real, folks!

7 – Watson x Data Intelligence: Unlocking The 99% 📊

IBM's CEO dropped a bombshell: 99% of business data isn't being used in AI systems. The updated Watson x Data Intelligence platform transforms unstructured documents, emails, and meeting transcripts into AI-ready inputs, boosting response accuracy by 40%.

What it means: 

Most organizations are operating with MASSIVE AI blindspots.

For Watson clients: conduct a data audit ASAP. Categorize by structure type and current AI accessibility.

The competitive advantage belongs to whoever converts their unstructured data into AI-ready assets fastest. Meeting transcripts, support interactions, and internal documentation typically offer the highest immediate value.

That knowledge locked in siloed documents and people's heads? It's finally ready to work for you—instead of collecting digital dust.

Reply

or to participate.