• Everyday AI
  • Posts
  • AI Agents: The Future of Enterprise Work

AI Agents: The Future of Enterprise Work

OpenAI launches o3-mini model, Altman showcases new AI at the White House and NVIDIA CEO meets with Trump and OpenAI’s $40B funding round and more!

Author

Everyday AI
January 31, 2025

Outsmart The Future

Sup y’all! 👋

Big o3 update from OpenAI today. More on that below. 

We’ll be covering o3, Operator, and DeepSeek next week. 

Aside from that, what OpenAI features do you wanna see more coverage of in the near future?

What OpenAI features do you want to see?

Vote to see live results

Login or Subscribe to participate in polls.

Whatever wins, we’ll plan an in-depth episode for February! 

✌️
Jordan

(Let’s connect on LinkedIn. Just tell me you’re from the newsletter when you send a request.) 

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: Why are AI agents booming right now? What are they? We explain how they can boost enterprise work. Give it a listen.

🕵️‍♂️ Fresh Finds: Hewlett Packard’s acquisition targeted by U.S. Gov, Italy blocks DeepSeek AI, Microsoft forms new AI unit and more! Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: OpenAI launches o3-mini model, Altman showcases new AI at the White House and NVIDIA CEO meets with Trump and OpenAI’s $40B funding round. For that and more, read on for Byte Sized News.

🚀 AI In 5: Google AI image model Imagen 3 has flown under the radar. We dive in and see how it stacks up against the competition. TBH, we’re kinda digging this. See it here

🧠 Learn & Leveraging AI: Are AI agents the missing piece you didn’t know your business needed? We break down how they can help. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about Oracle unveiling AI agents, OpenAI partnering with U.S. National Labs and Microsoft adding DeepSeek R1 to Azure AI. Check it here!

AI Agents: The Future of Enterprise Work 💡

Why AI agents? And why now?

You've prolly been seeing all the buzz around AI agents lately.

Same.

Here's the thing, though. There's more to them than meets the eye.

Scott Beechuk joins us to dive in deep and tell ya what you need to know.

Join the conversation and ask Jordan and Scott questions on AI here.

Also on the pod today:

• Current State of AI Agents 🤖
• Humans’ Roles in AI Integration🚶
• Challenges in AI to AI Interactions 🔗

It’ll be worth your 33 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 – Latta finds and solves bugs automatically, Invofox turns files into verified data and CapCut Commerce is end-to-end content creation for commerce.

Trending in AI – The U.S. Justice Department sues to block Hewlett Packard’s $14B acquisition of AI company Juniper Networks

Microsoft – Microsoft is forming a new unit to study AI’s impact.

Apple – Apple’s CEO believes DeepSeek shows “innovation that drives efficiency.”

AI Governance – Italy has blocked DeepSeek AI over data privacy concerns.

Big Tech – Anthropic CEO Dario Amodei is trying to avoid deposition in a copyright lawsuit by the Authors Guild against OpenAI

Meta - A leak from Mark Zuckerberg’s meeting with employees breaks down what was discussed regarding AI and more.

Meta has signed a deal for 595 megawatts of solar power in Texas.

OpenAI - OpenAI is going live with a Reddit AMA soon.

AI in Healthcare – A new study showed that ChatGPT-4 outperformed professionals at diagnostic reasoning.

AI in Science - AI has designed a new wonder material that is as light as foam but as strong as steel

1. OpenAI Unveils New o3-Mini Model 🚀

OpenAI is delivering on its promise with the timely launch of o3-mini, a new reasoning model that enhances speed and accuracy, according to OpenAI's latest announcement. Unveiled as part of the "ship-mas" series in December, o3-mini not only matches the performance of its predecessor in math, coding, and science but also responds 24% faster.

For the first time, free ChatGPT users can experience these advanced capabilities, shaking up the AI landscape just as Microsoft made o1 available for Copilot users.

2. AI's Capitol Hill Debut: Altman Previews Future of Automation 🏛

In a bid to prepare D.C. leaders for AI's rapid advancements, OpenAI CEO Sam Altman and Chief Product Officer Kevin Weil showcased upcoming AI capabilities near Capitol Hill, highlighting their potential impact on science, education, and government services. These advancements are expected to drastically increase efficiency and economic value.

Altman emphasized the transformative potential of new agentic technologies, suggesting they could significantly boost the U.S. economy. With big names like Interior Secretary Doug Burgum and former deputy Treasury secretary Wally Adeyemo in attendance, it's clear that AI's evolution is a hot topic in Washington.

3. NVIDIA CEO Meets with Trump to Discuss AI Policy 🤝

NVIDIA CEO Jensen Huang and President Donald Trump met to discuss U.S. AI policy in Washington, D.C. today. The conversation covered topics like China’s DeepSeek and the impact of U.S. chip export restrictions, which are currently under White House evaluation.

This meeting comes as NVIDIA navigates regulatory challenges, especially regarding their AI chip exports and their role in the U.S. tech landscape. The discussions could influence future investments in U.S. chip production, with potential revisions to the CHIPS Act on the horizon.

4. OpenAI in Talks for Massive $40 Billion Funding Round 🤑

OpenAI is reportedly in early discussions to secure up to $40 billion, valuing the ChatGPT maker at a staggering $300 billion. SoftBank is set to lead this historic funding round with an investment between $15 billion and $25 billion, while other investors will contribute the remaining amount. This move comes amid OpenAI's commitment to its joint venture, Stargate, and follows economic pressures from AI models released by China's DeepSeek.

If successful, this funding would make OpenAI the second-most valuable startup globally, trailing only SpaceX, and mark a pivotal moment in the AI industry's investment landscape.

5. Perplexity Faces Trademark Lawsuit 🧑‍⚖️

AI startup Perplexity is being sued by Perplexity Solved Solutions for allegedly infringing on their trademark, according to TechCrunch. The Texas-based company, which secured its trademark in November 2022, claims the startup's branding confuses consumers, especially since both companies offer similar software services.

Despite receiving a cease and desist letter, Perplexity has not withdrawn its pending trademark application. This lawsuit adds to Perplexity's legal woes, as they are also battling claims from News Corp’s Dow Jones and the NY Post over content copying issues.

6. Google Gemini 2.0 Flash Upgrades Roll Out ✨️

Google has announced that the Gemini app is getting the Gemini 2.0 Flash AI model, marking a significant enhancement to the app's capabilities. This update promises faster responses and improved performance for tasks such as brainstorming and writing.

Users can access the new features on both web and mobile platforms, while the older Gemini 1.5 versions remain available for now. Furthermore, Google's image generation now leverages the latest Imagen 3 AI, delivering more detailed and accurate visuals.

7. Intel's $2.2 Billion Federal AI Chip Boost 💰

Intel Corporation has secured $2.2 billion in federal grants as part of the U.S. CHIPS and Science Act, with plans to enhance semiconductor manufacturing across several states. Despite uncertainties with the CHIPS Act under the Trump administration, Intel remains optimistic about future engagements.

The company aims to utilize these funds for manufacturing and advanced packaging techniques in Arizona, New Mexico, Ohio, and Oregon.

Is Google's Imagen 3 BETTER than MidJourney?

Google image generation model Imagen 3 hasn’t really been talked about.

Is it better than DALL-E 3 or Midjourney?

We show you how to access Imagen 3 and give a head-to-head comparison.

🦾How You Can Leverage:

You ever ponder about how certain repetitive human interactions are so insanely inefficient that we overlook just how… insanely inefficient they are? 

Like, how maybe some human jobs could be better off handled by AI agents and everyone would be happier and better off? 

Case in point — there’s countless call center jobs in the world where all the human agents do is help other humans reset their passwords. 

Repetitive job for the human call center operator. 

Annoyingly long wait times for the human who forgot their password. 

Yuck. 

Scott Beechuk sees a better way — networks of AI agents. 

He’s a Partner at Norwest Venture Partners, and his company invests in multiple companies using AI agents to cut down on mundane interactions and (hopefully) make way for more meaningful connections. 

And now’s a gasoline-induced flaming hot time to be involved in anything AI agents.

In the past few weeks, we’ve seen both announcements and releases in the AI agents space from Anthropic, Microsoft, Salesforce and others. 

So why now? 

And how can companies balance the immense potential of AI agents with the potential peril they bring? 

Buckle up, humans. 

Scott is taking us to school on how AI agents are gonna be impacting enterprise work. 

Here’s what you need to know. 

1 – Convergence is happening now💡

Wait, so why is all of this happening now? 

After all, we heard some ‘AI agent’ rumblings years ago, but nothing ever seemed to materialize. 

Yet in the last few weeks, we’ve seen robust AI agent offerings from some of the largest companies in the world. 

So what’s changed? 

First, companies like OpenAI are making it easier to use their technology. 

Second, the learning curve and requirements for building AI agents is shockingly low and capable agentic systems are able to pop up over night. 

Lastly, Scott said AI agents now have the ability to connect with countless third party enterprise softwares and systems, but can now actually execute tasks with those systems. 

Try this: 

Even we’ve been buzzing a bit more about AI agents here at Everyday AI. 

Why? 

Companies like Salesforce and Microsoft bringing autonomous AI agents to market in 2024 legitimizes this tech that was once more scientific fiction than reality. 

If you want a more well-rounded view on the AI agent space, make sure to check out this episode.

2 – Agents coming for Tier 1 👀

Scott has already seen this happen — AI agents are coming for Tier 1 roles. 

And it’s not always in a ‘AI agent taking a human job’ type of way, but is oftentimes just an AI agent taking on pieces of responsibility that humans would normally take on. 

So what’s a Tier 1 support role?

Think of something like an entry level customer service position or even a Sales Development Rep. 

So many of the first-touch connections that companies have with their customers are routine, repetitive and don’t always require much finesse from a human. 

(Like instructing someone how to reset their password.) 

Try this:

Scott walked us through an example of this from Replicant, an AI agent company that Norwest invests in. 

He gave the example that if you call AAA insurance in North America, there’s a good chance your first conversation might actually be with an AI agent that sounds and reacts more like a human than a tele-operator. 

You can go read more about Replicant’s partnership with AAA here, and then take a deep dive on how AI agents are already disrupting customer service and sales industries. 

3 – Future agent networks need rules ✅

It’s a weird world we’re living in y’all. 

Like… there’s actual AI agents in the wild right now doing roles traditionally reserved for humans. 

AI agents of today can talk like a human, listen intently, respond in low latency (faster than the average human in some cases!), and do it all with access to your company’s information. 

Yikes. 

Right now, we have live, working individual AI agents. But, that will soon change as AI agent networks become mainstream. 

Cool right? 

Yeah! 

Dangerous? 

Maybe. 

But Scott cautioned this is where business leaders have to balance the power of AI agents (or groups of them!) with the current realities of large language model-powered agents. 

Right now, agents can talk to each other and solve problems and execute tasks. But if Agent 1 hallucinates a bit, and Agent 2 hallucinates even further…. You can only imagine the mess you’ll be in when agents 903 and 904 are executing actions on your company’s behalf. 

Double yikes. 

Try this: 

You remember that telephone game? 

One kid whispers a funny secret into another kid’s ear. 

After the message gets passed to 9 or 10 kids, it’s completely different than the original message. 

Scott says that’s what could happen with AI agent networks. 

He said companies who use agentic AI networks first need to develop a tracing system like developers use with APIs. 

The issue is there’s no standard language for agentic AI networks. Normally, network admins and devs can go through logs of their tech stacks, tracing API calls to find bugs. 

With LLM-powered AI agents, not so much. 

If you’re still new to AI agent networks (and aren’t we all?!) ZDNet has a great breakdown of what it all means.

⌚

Numbers to watch

1 Million

Meta’s Ray-Bans smart glasses sold more than 1 million units last year.

Now This …

Let us know your thoughts!

Vote to see live results

Do you attend our livestreams?

Every weekday, we bring you fresh AI insights, exclusive interviews, and breaking news with our Everyday AI livestream.

Login or Subscribe to participate in polls.

Reply

or to participate.