Taking snapshots of the AI future

PLUS: steps and tools for AI-driven demand forecasting.

Author

Nick Lipetzky
July 31, 2024 • Estimated Reading Time: 8 minutes

Hello! This is your non-technical shortcut to understanding AI’s business value, a simple guide to selecting the right AI solutions, and an executive coach for implementing and monitoring AI projects.

“The future depends on what you do today.”

Mahatma Gandhi

Welcome to The Daily Spark! Today, we delve into the complexities of calculating AI’s ROI and highlight JPMorgan’s strategic integration of AI with its LLM Suite. Discover how hyperrealistic voice technology is transforming industries and learn about practical AI applications in demand forecasting. Enhance your inventory management with actionable AI insights, and stay ahead with our concise, non-technical guide to AI’s business value.

In today’s Daily Spark

  • Measuring AI’s Business Value: Understanding AI ROI complexities and current usage stats.

  • Value Creation: How JPMorgan’s LLM Suite enhances productivity and decision-making.

  • Translated Tech: The future impact of hyperrealistic voice technology.

  • AI in Action: Practical steps and tools for AI-driven demand forecasting.

  • Human Side of AI: Optimizing inventory levels and improving supply chain efficiency.

Today’s Headlines Translated

MEASURING AI’S BUSINESS VALUE
The complexity of calculating AI’s ROI

Executive Brief: Thomas Claburn discusses the complexity of calculating AI’s ROI. Al has yet to pay off, with an estimated $1 trillion in capital expenditure commitments generating little economic return so far. Only 5.4% of US businesses reported using Al as of February 2024, up from 3.7% in September 2023.

  • Avoid Traditional ROI Calculations for AI Projects

  • 5.4% of US businesses used AI as of February 2024.

  • Clearview Consulting achieved an 85-90% reduction in data analysis time using AI.

  • Over 50% of AI projects fail due to cost underestimation.

  • AI had almost no economic impact according to The Economist.

VALUE CREATION
The Elements of JPMorgan’s AI “Research Analyst"

Business Case: JPMorgan's deployment of the LLM Suite showcases how AI can be strategically integrated into a business to drive productivity, reduce costs, and create a competitive advantage. By enhancing decision-making, fostering innovation, and improving customer service, AI tools like the LLM Suite can significantly contribute to the overall value chain of a financial institution.

  • Enhanced Productivity: The LLM Suite can assist employees with various tasks such as writing, idea generation, problem-solving using spreadsheets, and summarizing documents. This allows employees to focus on higher-value activities.

  • Improved Problem-Solving and Decision-Making: The chatbot functions as a research analyst, providing information, solutions, and advice. This can significantly improve the problem-solving and decision-making processes within the asset and wealth-management division.

  • Cost Reduction: AI tools like the LLM Suite can lead to efficiency gains from faster turnaround times and reduced operational costs.

  • Scalability: The AI tool can be scaled across the organization without a proportional increase in costs.

  • Enhanced Customer Service: Improved internal processes can lead to better customer interactions, as employees are better equipped with information and can respond more efficiently to customer needs.

  • Knowledge Management and Transfer: Expertise and insights generated by AI can be easily accessed and utilized by different employees, fostering a culture of continuous learning and improvement.

  • Risk Management and Compliance: AI can help in scanning and analyzing vast amounts of data for regulatory compliance, identifying potential risks, and ensuring adherence to regulations.

TRANSLATED TECH
Today’s Headlines with Tomorrow’s Perspective

  • Friend.com’s $1.8M domain signals brands will be crucial in AI era. Prior to the AI era, there were fewer tech companies and products. Companies could rely more on the novelty and functionality of their products.

  • Besides the fact that ChatGPT’s hyperrealistic voice technology will displace human voice actors, it raises several legal and ethical issues. The potential for misuse in creating deepfakes or unauthorized voice replicas poses significant risks. We all need to prepare to question everything via a digital channel.

TECH TERM
Hyperrealistic Voice Mode

Definition: Hyperrealistic voice mode refers to the ability of an AI system to generate speech that is indistinguishable from human speech. This is achieved using advanced speech synthesis technologies that capture the nuances, intonations, and expressions of natural human conversation.

One perspective is that alternative tools like Google’s WaveNet, IBM Watson Text to Speech, Amazon Polly, and Microsoft Azure Cognitive Services will eventually catch up making this tool low-cost and available to anyone.

AI IN ACTION

PROMPT WITH PURPOSE
Demand Forecasting

Copy and paste into ChatGPT, Meta AI, Claude, etc. Edit text to fit your specific case or to experiment for different results.

💥PROMPT
You are an AI assistant helping a retail company improve their demand forecasting. Follow these simple steps to create a forecast:

1. Collect Data:

• Gather the last 5 years of sales data, including dates and sales amounts for each product.

2. Identify Trends:

• Look at the sales data to find patterns. For example, do sales increase during certain times of the year or on specific days?

3. Check for Seasons:

• Note any seasons or holidays when sales usually go up or down.

4. Use Simple Forecasting Tools:

• Use basic tools like Excel or Google Sheets to create charts and graphs of the sales data.

• Try using built-in forecasting functions to predict future sales.

5. Combine Data:

• Include any information about upcoming promotions or events that might affect sales.

6. Make Predictions:

• Based on the patterns and additional information, make a simple forecast for the next few months.

• Write down your predictions for each product.

7. Review and Adjust:

• Compare your predictions to actual sales regularly.

• Adjust your forecasts based on what you learn.

Example Input Data:

• Sales data (CSV file with columns: date, product, sales_quantity).

• Information about upcoming promotions or events.

Example Output:

• Forecasted sales for the next few months in a simple table or chart.

Instructions:

1. Use the sales data to create charts and graphs.

2. Identify patterns and seasonal trends.

3. Use basic forecasting functions to predict future sales.

4. Adjust your predictions based on new information and review them regularly.

Prompt Elements: 

  1. Define the Objective

  2. Outline the Process

  3. Detail Each Step

  4. Provide Examples

  5. Recap Instructions

AI Tools for the Task

💥Amazon Forecast uses machine learning to generate accurate forecasts.
💥Akkio is a no-code predictive AI platform for data preparation, exploration, forecasting, and performance analysis.

HUMAN SIDE OF AI
AI to optimize inventory levels, forecast demand, and improve overall supply chain efficiency.

Example Workflow: Imagine a mid-sized retail company with multiple store locations and an online presence. They manage a wide range of products, including seasonal items, electronics, and clothing. The company has faced challenges with stockouts during peak seasons and overstock issues with slow-moving items.

Implementing an AI-driven inventory management system involves key steps:

  • Identify Challenges & Goals: Address stockouts, overstock, and reorder inefficiencies. Aim to improve demand forecasting, optimize stock levels, reduce holding costs, and enhance customer satisfaction.

  • Select AI Tools: Use Itemery for inventory management and Abacus.AI for demand forecasting and predictive analytics.

  • Data Integration: Collect historical sales data (3-5 years), real-time sales data from POS systems, and external data like market trends and weather forecasts.

  • Configuration: Configure Itemery for product categorization and Abacus.AI for demand forecasting.

  • Forecasting & Replenishment: Use Abacus.AI for daily demand forecasts and Itemery for calculating reorder points and safety stock levels.

  • Automation: Automate replenishment with Itemery, integrating with suppliers for seamless order processing and tracking.

  • Monitoring & Improvement: Monitor inventory levels, forecast accuracy, and supplier performance via dashboards. Continuously improve AI models based on new data.

This approach enhances the human by reducing stockouts, lowering holding costs, improving efficiency, and enhancing customer satisfaction.