Dear Friend,

Happy Wednesday! In our Applied AI class, we spent time on custom agents, Jasper grids, Google Studio workflows, and a proposal workflow that can turn a single meeting URL into a proposal draft, folder structure, and project tasks.

That theme carried right into this week’s news, too. The most useful AI stories we are seeing right now are about building systems that 10x productivity, improve decision-making, and reduce your cost-to-serve.

I’m also keeping an eye on the policy side. If you are using AI you need to be thinking about governance, documentation, and vendor risk right alongside experimentation. It is all about setting up clear guardrails and keeping human review in the loop.

Let’s dive into the updates.

Microsoft Pushes Multimodal AI Into The Enterprise Stack

Microsoft announced three new foundational AI models that generate text, voice, and images, including a speech-to-text model, a voice model, and an image model. The company is positioning them for practical use and says cost is part of the pitch, which matters for businesses looking to scale AI without letting spending spiral out of control.

My take: We believe in a lean tech stack and getting more value from the tools you already have. Pay attention to this shift because it points to easier consolidation. More companies are going to choose AI that works inside the systems they already use, especially when speed, pricing, and deployment are on the table. For business teams, that means faster content creation, easier transcription, better internal knowledge workflows, and stronger customer experiences, all inside one connected ecosystem.

Read more about Microsoft’s new models here.

Google Vids Gets for Marketing And Training Teams

Google added new AI features to Vids, including prompt-based avatar direction, Veo 3.1 support, YouTube export, and a Chrome extension for screen recording. In simple terms, it is getting easier for teams to produce training clips, internal explainers, product videos, and marketing content with less manual editing.

My take: We are in the Google Workspace and look to leverage it as much as possible. We are using Vids as much as we can and we’ve maxed out our Vids credits. The real win is not making a cinematic video; it is helping a lean team create good-enough content quickly and consistently to drive conversions or help with onboarding and training.

Read more about Google Vids updates here.

Facebook Insider Builds AI Powered Content Moderation

A Facebook insider has raised $12 million for a new platform building a real-time policy layer for content moderation. Its system evaluates content against a company’s rules in 300 milliseconds or less and can either slow distribution, flag for review, or block risky content right away.

My take: The real question will be how well the policy layer handles nuance. Speed is impressive, but accuracy, fairness, and transparency matter just as much. If they can balance all three, this could become a very important part of the future of digital operations. As a moderator of a very busy group, we often see posts that get caught in this trap. Some should be caught and some not and then a bunch of posts that should be caught are not.

Read more about the Facebook insider’s moderation tool here.

MIT Points to a Smarter Future For AI In Energy

MIT highlighted Dean Price’s work on using AI and machine learning to support the design and operation of nuclear systems. The key idea is that AI can identify patterns in complex reactor data, improve early-stage design decisions, and support safer, more economical control decisions without replacing the established safety frameworks that already exist.

My take: I appreciate that this story is grounded in augmentation, not replacement. The companies that get the most value from AI over the next few years may be the ones using it to strengthen expert systems and improve decision quality in high-stakes environments. For businesses, this is a strong reminder that some of the most valuable AI progress is happening behind the scenes in infrastructure-heavy industries.

Read more about MIT’s work in energy here.

Tools This Week

• Gamma – Great for creating presentations, documents, and webpages quickly when you need polished client-facing materials fast.
• Jasper.ai – A strong option for teams that want brand voice consistency, campaign support, and structured content workflows.
• Read.ai – Helpful for turning meetings, emails, and messages into summaries so your team spends less time chasing notes and next steps.

How to Use AI for Email Marketing: Workflows That Save Hours

AI should not replace your email strategy. It should remove the busywork. The article below hows how to use AI to draft faster, test smarter, and create more consistency in your email marketing without adding a bunch of extra tools. The smartest move is to start with one workflow, keep your tech stack lean, and use AI where it saves real time. Let AI help with first drafts, subject lines, segmentation ideas, and testing. Keep humans responsible for tone, accuracy, and the final send. That is how small teams move faster without sounding robotic.

Read the full article

Google Labs 🧪 

How to Use Pomelli’s Product Photoshoot Feature

Pomelli has a new Photoshoot feature that helps you create professional product images fast and for free. You can upload a product photo, choose a few styles, and let the AI generate polished images for your brand.

Step 1: Open the Photoshoot tab

Go to the Photoshoot tab inside Pomelli.

Step 2: Upload your product image

Choose Create a Product Photoshoot and upload a photo of your product.

Step 3: Pick your templates

Click the pencil icon and choose the styles you want, like:

• Studio
• Ingredients
• In-use
• Contextual

Step 4: Choose your aspect ratio

Pick the size you want for your images, such as 9×6, 1×1, or 4×5.

Step 5: Create your photoshoot

Click Create Photo Shoot and Pomelli will generate several product images for you.

Step 6: Edit if needed

If you want changes, click on an image and type the edit you want. The AI can create a new version based on your instructions.

Step 7: Save your images

When you’re happy with the results, download the images or add them to your business DNA.

Important tip

Be careful when clicking the X. It can close the photoshoot and delete your images if you haven’t saved them yet.

Pomelli also lets you generate a product image from a text prompt first, but the easier option is usually to upload your own product photo and start from there.

How to Use Google Workspace Studio to Build AI Automations

1. Open Google Workspace Studio by going to studio.workspace.google.com. You can also find it through the shortcut icon inside Google Workspace.
2. Explore the main tabs:

• Discover to browse templates and pre-built agents
• My Agents to view your saved automations
• Activity to track what your agents have done

3. Create your first agent by starting with a template or describing what you want in plain English. For example: “If an email contains a question for me, label it to respond and ping me in chat.”
4. Set up the three main parts of the automation:

• Starters for triggers like receiving an email, a calendar event, or a form submission
• Steps for actions like drafting a reply, extracting details, or sending a message
• Variables for dynamic details like sender name, action items, or AI responses

5. Connect it to the tools you already use. Workspace Studio works natively with Gmail, Drive, Chat, Calendar, Sheets, and Forms, and also supports some third-party tools like Asana, Jira, Mailchimp, and Salesforce.
6. Test a simple workflow first, such as automatically replying to form emails or extracting action items from messages. This helps you save time on repetitive tasks before building more advanced automations.
7. Check your plan and usage limits. Workspace Studio is available for Google Workspace Business and Enterprise customers, and higher AI usage may require an add-on.

Join the Conversation

Got a burning question, a hot take, or just want to share your latest AI wins? Hit us up at [email protected]. Your insights keep this community growing and thriving!

See you in the Lab,

-Nicole A. Donnelly

Founder, AI Smart Ventures

AI Strategy, Consulting, Training, and Implementation