Author: Chad Stewart, VP of Salesforce Support Services
Today’s business Artificial Intelligence (AI) applications have two primary uses – predictive, enabling better decision making, and generative, simplifying research and content creation across mediums. Salesforce provides solutions across both categories and after Dreamforce (summary links found here) there is a world of hype about how accessible these solutions are today. But can you flip a switch and start using them? No. AI relies on models, and models are built on data. For your enterprise to do something with AI that provides a competitive advantage, you need to unleash your proprietary data. The following steps outline the preliminary steps that need to be taken, and are frequently overlooked, in today’s marketing buzz world of AI.
Spinnaker Support is a Salesforce Partner that helps customers across industries take advantage of everything Salesforce offers. We immediately began fielding the question “How do I make that happen?”. How indeed. Let’s review the five steps you can take today to prepare for an effective Salesforce AI implementation.
Step 1 – Articulate Core Competencies & Niche Capabilities
While obvious to some, it will be an important part of building your coalition and engaging with Salesforce and any partners. Using AI to strengthen what you are already good at, will increase the value of implementing new solutions. Today there are Large Language Models, built on public information, that can be quite helpful, but do not present a strategic advantage. Literally everyone in the world has access to these.
Let’s imagine we’re a products company that makes the highest quality window fashions in the world. We’ll know more about design, style, fabrics, durability, and of course how our existing products work than anyone else. Undoubtedly there will be data about customer issues, sales volume by products, and partner effectiveness. Maybe you’re a services company that provides software support to enterprises across the globe. You’ll know more about how to solve common issues, leading practice implementation, and measuring ROI. Articulating these competencies now will help you know where to focus your efforts later.
Step 2 – Analyze Enterprise Data Advantages
Based on the above, what data does your organization have as a result of these core competencies, that no one (or at least very few) others in the world have access to? Where does this data live? Start building your coalition and brainstorm, with this data accessible to ask questions of, what efficiencies could be gained? Think across your organization to every customer facing department and determine use cases that could be solved with this data. What does success look like and what are all of the data elements and sources needed to inform the desired outcome?
Let’s continue our examples, for our window fashions company
Step 3 – Prioritize AI Use Cases
Undoubtedly, you’ll have great ideas. Stack rank these based on how much value they would potentially add to your organization. Indicate if they are predictive in nature or generative. Here is the tricky part – for each in your top 5-10, list out all of the data sources across your organization that would be needed to support the use case.
In the theme of AI and our consumer goods example, I’ve asked Chat GPT to generate 10 ideas:
- Demand Forecasting: Use AI to predict consumer demand for products, enabling better inventory management and reducing stockouts or overstock situations.
- Price Optimization: Implement dynamic pricing strategies based on AI-driven market analysis and competitor pricing data to maximize revenue and competitiveness.
- Personalized Marketing: Create personalized product recommendations and targeted advertising based on customer preferences and behavior data.
- Supply Chain Optimization: Optimize the supply chain by predicting disruptions, improving logistics, and ensuring on-time product deliveries.
- Quality Control: Utilize computer vision to inspect and detect product defects, ensuring consistent product quality.
- Inventory Management: Predict when and how much inventory to reorder, reducing carrying costs and minimizing wastage.
- Customer Sentiment Analysis: Analyze social media and customer reviews to gauge consumer sentiment, helping to improve products and customer service.
- Shelf Space Optimization: Use AI to recommend the most effective shelf space allocation for products in retail stores based on historical sales data and customer traffic patterns.
- Fraud Detection: Implement AI-driven fraud detection algorithms to identify and prevent fraudulent transactions, especially in e-commerce.
- Trend Analysis: Analyze market trends, consumer preferences, and emerging technologies to identify new product opportunities and adapt the product portfolio accordingly.
Step 4 – Consolidate Data Islands
Salesforce now provides free licensing to Data Cloud, which enables organizations to connect, federate, and harmonize data from any product and system into a complete view of every customer. The free version has volume limits but can get the ball rolling on your first use case. Associating your disparate data into a consolidated source will enable model building. These models can be trained to predict outcomes, based on past performance, or generate ideal content for sales/service/marketing again based on what has worked for your organization in the past. Take away here is to consolidate your data, so that AI models can have access to it.
Step 5 – Talk to Spinnaker About Implementing Salesforce Solutions
Salesforce has a wide range of integrated AI solution offerings. From Einstein Prediction Builder, to Einstein GPT, these solutions are ready once you are! The worse thing you can do, is stand still. Additionally, if you jump to implement without planning, the result will not match your expectations. In 2006 mathematician Clive Humby said “data is the new oil.” While not useful in it’s raw state, it can be refined, processed, and turned into something of great value and use. Your enterprise has data about your business that no one else does, raw oil sitting in silos, providing limited value. Contact Spinnaker today and start refining, testing, measuring, and increasing that value now.