AI in Business: Walmart AI Seller Tools, Salesforce Agentforce for Manufacturing, Adobe Firefly + Gemini 2.5 Flash, Cleveland Clinic Clinical Trial AI, and MIT Climate Modeling Study
Retail is speeding up listings, manufacturers are tapping agentic AI on day one, creatives get faster visuals, and healthcare is cutting the grind from clinical research. I also pulled in a crisp MIT result that helps leaders invest wisely (hint: bigger models aren’t always better).
This Week’s Insights
Walmart Debuts AI Tools to Help Marketplace Sellers Move Faster
Walmart introduced an AI-powered listing tool to reduce time-to-market and boost conversion. It’s part of a broader push connecting marketplace inventory to physical stores and improving next-day coverage across major cities. For sellers, that means lower listing friction, better visibility, and potentially higher sell-through just in time for the holidays.
Source: Retail Dive
My take: If you sell on Walmart, take this info and create FREE account equity. Standardize your titles, attributes, and imagery now, so the AI can apply consistent rules across your catalog and track whether conversion lifts on AI-assisted listings versus manual.
Salesforce Launches Agentforce for Manufacturing
Salesforce released industry-specific AI agents and templates to optimize demand planning, inventory, service scheduling, and sales. Early customer examples report dramatic cycle-time reductions and fewer admin hours for field teams. Translation: faster quotes, less downtime, and reps focused on revenue activities instead of swivel-chair tasks.
Source: Salesforce
My take: Manufacturers craving “AI that actually does work” should pilot narrow use cases (e.g., spare-parts lookup, rebate management) with clear SLAs and guardrails. Start where your data quality is decent and the outcome is measurable in weeks.
Adobe Firefly Adds Google’s Gemini 2.5 Flash for Image Generation
Firefly Text-to-Image, Firefly Boards (beta), and Adobe Express now include Gemini 2.5 Flash Image, so teams can generate variants, resize assets, and move from concept to campaign without leaving Adobe.
Why this matters
- Multi-model creativity: In Firefly and Express you can pick the model that fits the job Adobe’s commercially safe Firefly models (image, video, vector, audio) or partner models from Google, OpenAI, Black Forest Labs, Runway, Pika, Ideogram, Luma AI, Moonvalley, and more. Each model has its own look; you get the flexibility without bouncing between platforms.
- Pro handoff: Seamlessly open results in Photoshop, Illustrator, or Premiere for precise edits.
- Trust by design: Content you create or upload in Adobe apps is not used to train generative models. Adobe also attaches Content Credentials to wholly AI-generated assets to promote transparency.
Source: Adobe
My take (how to get the win)
Centralize your brand settings (fonts, palettes, product shots) and keep a prompt library for common requests. The payoff isn’t just better visuals it’s cutting asset turnaround from days to hours while staying on-brand.
Cleveland Clinic Rolls Out Dyania Health’s AI Across the System to Speed Clinical Trial Recruitment
Following pilots in cardiology, oncology, and neurology, Cleveland Clinic is deploying Dyania’s Synapsis platform enterprise-wide. In one study, the AI identified eligible melanoma trial participants with ~96% accuracy in 2.5 minutes versus 95% accuracy taking 427 minutes by a specialist nurse, a massive throughput gain that can get patients into studies faster.
Source: Cleveland Clinic
My take: Every patient deserves a fair shot today, not months from now. AI trial matching makes that real: faster IDs, fewer missed candidates, more time for care. Start with one clear bottleneck (trial matching, prior auth, radiology lists). Plug in an AI screener.
Track four metrics: time-to-identify, accuracy vs. baseline, enrollment conversion, clinician hours saved. If the curve bends the right way, scale it. The ripple effects are big, quicker options for families, more inclusive studies, stronger sponsor scorecards—and a care team with precious time back where it belongs: with patients.
MIT: Simpler Models can Outperform Deep Learning for Certain Climate Predictions
MIT researchers showed that physics-based, simpler models can beat deep-learning approaches for regional temperature forecasting (while deep learning fares better on local rainfall). They also propose a sturdier benchmarking method to avoid being fooled by natural variability. For businesses, this suggests you can often get reliable forecasts with cheaper, more interpretable models.
Source: MIT News
My take: You don’t need “fancy” to serve people well. If your signal is stable and physics-driven (think demand tied to temperature), choose the simpler, explainable model. It’s faster to validate, easier to defend to compliance, cheaper to run and often more accurate. Do a clean bake-off: accuracy, robustness, interpretability, run cost. If simple wins, bank the savings and put the time back where it belongs on decisions, not debugging.
Upcoming Learning Opportunities with AI Smart
Take the Next Step with AI. Enroll in Our Applied AI Certification
Applied AI – Level 1 Certification .
What You’ll Learn in the First Few Weeks
We’ll start with the basics to build a solid AI foundation before diving into advanced topics.
Module 1: Introduction to AI
Learn essential tools, terms, and how to fact-check AI content. You’ll also create a simple AI use policy to ensure responsible usage.
Module 2: Prompt Engineering and Brand Voice
Master crafting effective prompts and maintaining a consistent brand voice across AI tools, so everything you create stays on-brand.
Don’t just take our word for it…
Here’s what a recent student had to say:
“I highly recommend this course to anyone in marketing who’s been curious about how to use AI in a practical way. The instructors, Nicole, Andrea, and Mericar, brought a well-structured approach to every session. I had been looking for guidance on which tools to use and how to use them effectively and they delivered that as well as marketing ideas and tips. If you’re navigating AI and want to understand what tools will support your work and how to use them, this course is absolutely worth it!”
— Charis Armer
Ready to get started? Our next 10-week live course begins Tuesday, September 30th, 2025 9am-10am PST. (previously listed as September 16th).
Special Offer: Use code EH7XZTNF for 50% off your registration!
Mastering AI in Your Ops
Develop Your Tech Stack
Bring your automation to life by connecting AI tools to your existing systems for seamless workflows. For example, integrate a CRM with email automation and analytics to create a single source of truth.
If you’re new to this, don’t worry many tools are user-friendly, and with guidance, setup becomes manageable. We recommend Zapier for integrations. Troubleshooting and fine-tuning ensure everything runs smoothly.
Ready to build automated systems?
Our AI Your Ops course will guide you through creating time-saving automations.
Next session starts October 2, 2025
(previously listed as September 18th).
Learn More and Enroll in AI Your Ops
Tools to Try (curated)
- Jasper — AI built for marketers to scale on-brand content and streamline approvals across teams.
- Pictory.ai — Turn scripts, blog posts, or URLs into short, polished videos for social and ads in minutes. \
- Read.ai — Meeting copilot that generates summaries, action items, and topic reports across Zoom, Teams, and email.
Let’s Keep the Conversation Going
I love seeing these positive stories about AI because they show us that this technology is a powerful tool for growth and efficiency, not something to be feared. The key is to be proactive and curious about how you can integrate these new capabilities into your own business.
So, how are you thinking about using AI in your business this week? Let’s keep exploring these opportunities together.
Can’t wait to hear from you!
Cheers,
-Nicole Donnelly
Founder, AI Smart Ventures
