Issue 12: Enterprise AI Basics Explained, Storyboarding Simplified
From Enterprise tech, e-commerce disruption and job losses to your next deck: the big and small effects of AI
Good evening! Today, we have a lot of news to cover, right from online shopping integrations to job shifts, followed by a note on the Enterprise AI Stack. We wind it up with a quick guide on using AI for Storyboarding.
News That Matters
Shopping has a new destination - ChatGPT!
Users in the US can now buy items directly in chat with Instant Checkout, starting with Etsy and expanding to Shopify soon. To make this work, OpenAI launched the Agentic Commerce Protocol, which lets merchants connect their products and handle orders. Shoppers don’t pay extra, while sellers cover a small commission.
This could reshape how online shopping works. The chat window itself becomes the storefront, which means rethinking how you share product feeds, track attribution, and manage customer data. The upside is a smoother buying experience, but the trade-off is giving up part of the customer relationship.
Brace for AI-era job shifts
Accenture has already cut over 11,000 jobs worldwide in the past three months, warning more are likely as it undertakes an $865M restructuring tied to AI and digital work. The company says it is investing heavily in reskilling, but employees who cannot be redeployed into new roles are leaving. In India, TCS is cutting 12,000 jobs, with the employee union reporting even higher numbers. Taken together, these moves show how quickly big firms are reshaping their workforces for the AI era, and how unsettled the transition could feel on the ground.
OpenAI doubles down on AI infrastructure
First, it struck a deal with NVIDIA to lock in up to 10 GW of GPU systems, with NVIDIA also agreeing to invest up to $100B to support deployment. Then, with Oracle and SoftBank, it announced five new U.S. data centers under the “Stargate” project, adding about 7 GW of capacity and hundreds of billions in spend.
Clearly, the AI arms race is accelerating. However, the big open question: is this smart planning for insatiable demand, or the start of an AI bubble?
One Trend of Note
The Enterprise AI Tech Stack: Making AI Work at Work
Every company today is asking the same question: How do we actually bring AI into the business? The answer isn’t picking the flashiest chatbot. It’s building the right Enterprise AI tech stack. Think of it as building blocks: if one block is weak, the whole thing wobbles.
Applications: What employees actually use
Think everyday tools like Copilots in Office, AI dashboards, assistants in CRMs, and voice agents in customer service. It is where RoI becomes visible but poor design or weak integration can derail adoption. Off-the-shelf options like Microsoft Copilot and Salesforce Einstein are common, though many firms are also experimenting with lightweight custom apps.
Models (LLM): Brains behind the apps
LLMs are powerful but generic until paired with company data. That’s where RAG, fine-tuning, and guardrails come in. Proprietary models like GPT or Claude are easy to plug in via APIs. Open-source models like LLaMA or Mistral give firms more control but require hosting, running inference, and fine-tuning.
Data: The fuel
Most business data — invoices, spreadsheets, emails, records — is scattered and messy. AI can’t do much with it until it’s cleaned up and structured. IBM estimates over half of AI projects stall for this reason. Companies are leaning on platforms like Snowflake, Databricks, or BigQuery to make data usable for AI.
Infrastructure: The foundation
Cloud platforms (AWS, Azure, Google Cloud) provide the compute, storage, and security that support everything above. GPUs from NVIDIA or AMD power the workloads. This layer stays invisible to most employees but makes the whole stack possible.
Build vs. Buy: Trade-off across layers
The key decisions show up in two places. At the application layer, firms choose between buying ready-made AI tools or building custom AI apps. The choice at the model layer matters only if you’re building apps. You can use proprietary models through their APIs, or host and run open-source models yourself.
In practice, some enterprises buy AI applications off the shelf, while others are experimenting with their own lightweight custom apps. Even if building their own apps, most rely on proprietary models, with only a small minority hosting open source.
The hidden layer: Process and integration
Even if you have the perfect tech stack, process debt (outdated processes and clunky workflows) can derail adoption. Researchers have found that the majority of AI initiatives fail not because of the tech, but because of outdated workflows and siloed systems. Integration is another common roadblock. AI succeeds when it is embedded directly into the systems employees already use, because employees don’t want another standalone tool.
Why the AI tech stack matters to you
You (most likely) won’t need to know how to code or train models. But understanding the stack helps you spot where AI could genuinely help your team, ask sharper questions when IT or vendors pitch “AI add-ons” and avoid falling for buzzwords that overpromise and underdeliver.
AI in Practice
Storyboarding With Some AI Assistance
It’s midnight, the client meeting is in 9 hours, and you’re staring at a blank slide deck. You scramble through old presentations hoping to find something reusable. Sounds familiar?
That’s where storyboarding helps as a primer to build a good output, or to tell a good “story” so to say. Storyboarding is essential for client presentations, business pitches, training documents, user flows, event planning, speech drafting, film making etc.
A few issues back, we looked at using Gamma.app to create effortless presentations. Today, we are going a step behind: using AI to go from a vague idea to a clear slide-by-slide storyboard.
Here’s a how-to on using your favourite conversational AI tool to sketch a storyboard for a one-hour training on money management for entry-level factory workers:
Anchor the goal and audience.
Prompt: I want to design a one-hour training presentation on money management for entry-level factory workers. Tone should be simple, conversational, mixing English and Hindi. The goal is to explain expenses, savings, and end with a pledge to save / invest.
Define the scope of content.
Prompt: The presentation should cover the importance of money management, key elements of expenses and savings, types of savings, and a final call-to-action / pledge. Suggest a one-hour flow balancing teaching, discussion, and activities. Include stories, analogies, or group exercises.
Refine tone and language.
Prompt: Rewrite the storyboard in simple English with Hinglish phrases to make it relatable.
Expand each section.
Prompt 1: Give me three real examples of savings for slide no. 3.
Prompt 2: Help illustrate expense management with an analogy.
Add interactivity.
Prompt: Suggest a way to do an ice breaker before launching into the training.
Plan visuals.
Prompt: For each section, suggest visuals or diagrams, and how to place them on slides.
Export and build. Download the storyboard in a preferred format (.pptx in this case) and flesh out the slides.
Tips to increase effectiveness:
Start with the target segment / audience profile.
Use the tool iteratively. Don’t expect magic in one prompt.
Brainstorm with the tool on ideas to make the delivery more interactive.
Leverage the tool to build handouts / posters if needed, to go along with the presentation / speech.
Have you used AI for Storyboarding? Do you get involved in Enterprise-AI decision making? What are your experiences? Let us know so we can learn together.
Cheers!