Top Artificial Intelligence (AI) & Machine Learning (ML) Examples
May 15, 2022
Are you struggling to understand Artificial Intelligence (AI) & Machine Learning (ML)? Start here first.
Once you get down to a specific business context, thousands of valuable questions can be asked and answered with AI/ML and your data. So, to get the creative juices flowing, we are compiling a list of known scenarios. Check back often, as we plan to update this post frequently. Also, let us know if we missed something.
Data Search & Visualization: Using AI/ML found in tools like ThoughtSpot you can search your data using plain English. The AI/ML then creates relevant reports and dashboards on the fly to visualize the data. Gone are the days of waiting for developers to create your reports. Now you can be your own developer.
Automated Data Mapping: Bringing new customers, vendors, and IT systems on board is traditionally expensive and time-consuming. A vital part of any data integration project is the data transformation step, which maps data from one format to another. At the same time, business rules and validations are applied to that data. AI/ML mapping products, like Adeptia, Fivetran, and DeepIQ, consume your historical data mappings and predict future data mappings for you, saving time and money. These tools can save you time and money as interfaces can customer/vendor schemas change over time.
Data Labeling: To use many data sources (e.g., images), the data objects need metadata, called labeling. AI/ML can help augment data objects with labeling that facilitates their use in querying and AI/ML models. For example, Labelbox is an AI/ML ISV that helps in scenarios like this.
Synthetic Data: AI/ML can generate synthetic data to test your systems and avoid exposing your actual data.
Data Compliance: Discovering and protecting sensitive data in our data environments can be challenging. AI/ML companies like BigID can automate this task by searching and flagging data risk so that companies can ensure the assets are securely managed.
Data Accuracy: Using AI/ML, we can scan data sources and rank the data based on accuracy. This will then alert Data Engineers to potential problems and areas of improvement. The more accurate our data is, the better the AI models we can produce and the better our business outcomes.
Intelligent Security: Analyze data feeds to flag fraud and other potential security breaches within your network.
Deception Security: Deploy decoys in your network and then use AI/ML to flag potential security breaches.
Code Generation BOTS: Chatbots are able to understand natural language and write code based on the language. They interpret the meaning behind words and turn them into programming commands. This can help you write code faster, as it can quickly find relevant concepts and suggest appropriate solutions.
Customer Data Platform: You interact with your customers through several channels (website, email campaigns, social media, etc.), but we often don't have a way to associate customers across channels. Using AI/ML, CDP tools help us associate and connect unique customers across the "noisy" interactions they create through many different interactions. For example, a good CDP will enable associating a customer record "ABC," which interacts with your mobile site with a customer with "DEF," which interacts with your website. This allows you to consume the full breadth of customer behavior to improve targeting the customer in future campaigns.
Customer Segmentation: Segment your customers based on data attributes using AI/ML to understand your customer types better. This helps you make better targeting decisions and increases profitability. This may also include determining your loyal customers to ensure they are engaged and don't churn.
Customer Churn: Based on past behavior, how likely will your customer be to cancel your service or stop buying your products?
Customer Engagement: Determine how engaged your customers are with your brand across social media channels. Benchmark engagement against your competitors to determine how well your social media campaigns perform.
Customer Retargeting: Target advertisements to customers who have previously shown interest in a product or a service.
ROMI: Use AI/ML to measure your marketing efforts' effectiveness at converting customers and driving profitability.
Social Media Content: AI recommended social media content based on content created by other bands that drives brand engagement. Also, AI created social media content.
Competitive Intelligence: How does your brand sentiment stack up against your competitors? How do your social media campaigns rank against competitors, etc.?
Marketing Mix/Channels: Using AI, you can predict what mix of marketing spend will maximize return given a specific audience and across available channels.
Personalization: Personalized experiences/offers through websites, social media, or email campaigns based on customer behavior. For example, targeted product offerings based on AI-driven models based on past customer interactions and their propensity to buy. Give them exactly what they want when they want it.
Context-Aware Ads: Using Natural Language Processing (NLP) and Computer Vision, you can determine the context of your ad placement. Ensure your ad fits the context of the website where it is being placed.
BOTS: Answers to 80% of most customer questions about our product and services can be found on our websites, yet customers get easily lost in our complex information taxonomies. AI/ML bots allow customers to ask questions using simple English to retrieve the desired content. This results in greater sales (since most customers do their research) and reduced calls to support.
Demand Forecasting: AI/ML models can cluster products by type of demand (erratic, stable, etc.) and consider factors like product lifecycle, cannibalization, seasonality, promotions, POS, weather forecasts, etc., to predict future demand. This helps "purchasing and supply" to better manage the business. Companies like ORS Group provides ML models that enable these capabilities. Specifically, consensus-based collaborative forecasting supported by AI/ML, real-time forecasting for S&OP optimization and inventory planning, and revolutionary solutions for demand sensing and shaping.
Lead Generation: Determine companies and contacts most likely to need your product and service and make purchasing decisions.
Lead Scoring: Score "rate" leads for their likelihood to purchase based on past performance and current data. This allows your sales team to focus on leads most likely to close.
Sales BOTS: BOTS can suggest sales content and responses based on best practices and experience.
Recommended Next Steps: Based on past performance, suggest the next steps for each opportunity based on market, customer size, opportunity attributes, etc. This can also be determined by best practices developed by top sales performers.
Personalization: Personalized offers with a high likelihood of conversion based on customer profile and behavior.
Digital Assistant: An assistant can help negotiate the next steps with your customer, e.g., setting up a meeting and freeing up precious time for your sales team.
Sales Compensation Planning: Determine the right sales incentives to keep your team motivated.
Sales Analysis: Analyze your marketing and sales data to determine underperforming points in your funnel, underperforming team members, and insights to increase telesales, social media, email campaigns, etc.
Customer Analysis: Analyze all customer touchpoints to understand better your customers' behaviors and what drives them to the subsequent steps/closing.
Operations, Manufacturing & Supply Chain
Supply Chain Planning: AI/ML models can run what-if scenarios using various demand, supply, inventory, fulfillment, facilities, locations, routes, and policy decisions to predict and analyze impacts on decisions. Companies like ORS Group provides ML models that enable these capabilities.
Production Planning: AI/ML models that allow planners to create a constrained material requirement plan to enable Just In Time (JIT) delivery based on constraints: production capacity, product mix, lot sizes, and raw materials. Companies like ORS Group provides ML models that enable these capabilities. Specifically capacity planning schedulers, production capacity optimizations, and the creation of a constrained material requirements plan.
Global Capacity Planning: AI/ML models that create an optimized production plan to enable Just In Time (JIT) production for a worldwide supply network considering multiple constraints. Companies like ORS Group provides ML models that will allow these capabilities—specifically optimizing capacity and enabling JIT production for a global supply chain using web-based collaboration with factories while delivering comprehensive monitoring and control.
Process Improvement: Using AI/ML, you can derive insights into your operations, supply chain processes, and likely improvement areas.
Predictive Maintenance: Predict failure likelihood and dates for your machinery so that you can take action BEFORE a failure costs you downtime.
BOT Purchasing: AI/Ml BOTS that enable procurement automation that reduces capital expenditures while increasing margin. Companies like ORS Group provides ML models that will allow these capabilities. Specifically Human-AI purchase order processing that incorporates sales forecasting and automated safety-stock decisions.
Robotics: AI/ML-based robotic systems can automate repeatable portions of the manufacturing process.
Automated Invoicing: AI/ML processes can automate invoicing and accounts payables reducing human costs and human error (e.g., entering incorrect AP data).
Building Management: AI/ML can help you detect asset failures beforehand and monitor and reduce the overall energy consumption for your buildings.
Frictionless Retail: AI/ML-based retail systems can eliminate humans in the checkout process. This can include AI/Vision technology that tracks consumers and the products they carry/purchase or mobile-based self-checkout systems. Frictionless retail systems reduce operating costs while reducing loss and theft.
Allocation Planning: AI/ML models ensure that inventory levels by product at each store match demand in a Just In Time (JIT) manner. This reduces the required stock while increasing customer satisfaction. Companies like ORS Group provides ML models that enable these capabilities. Specifically, fully automated stock allocation and replenishment, the optimal initial allocation for new products and stores, and reduced stock-outs and average inventory levels.
Allocation Optimization: AI/ML can automatically fulfill store allocation requests while proposing allocation recommendations for wholesale partners. When central stock is scarce, you can prioritize allocations based on service levels, target revenues, or manual customer priorities. This can all be simulated before executing decisions. Companies like ORS Group provides ML models that enable these capabilities. Specifically, central stock load balancing between stores and wholesale partners, maximizing the revenue potential of stock and decreasing holding costs, and fulfill desires service levels for all customers.
Omni-Channel Ordering: AI/ML models that streamline and automate order fulfillment processes from any location across their distribution network, even direct-to-home delivery options. Companies like ORS Group provides ML models that enable these capabilities. Specifically real-time global inventory synchronization, and net profit maximization for every sale, while driving higher customer satisfaction and loyalty.
Retail Analytics: Using AI/ML and analytics to aggregate data from all parts of the retail process and then apply math, stats and econometrics to help drive the process from strategic planning to tactical execution. Companies like ORS Group provides ML models that enable these capabilities. Specifically, providing real-time visibility of all store operations, sophisticated demand and price elasticity planning, allowing what-if scenarios and more profitable campaigns.
Suggested CS Responses: AI/ML can suggest solutions and responses to customer service reps.
Customer Service Chatbots: Answers to 80% of most customer questions about our product and services can be found on our websites, yet customers get easily lost in our complex information taxonomies. AI/ML bots allow customers to ask questions using simple English to retrieve the desired content. This results in fewer customer service interactions.
Intelligent Problem Identification & Routing: Route open CS requests to the most appropriate department and person based on ML/AI and the requested content.
Authentication: Identify and authenticate customers based on voice using AI/ML.
Customer Service Sentiment: Identify the customer's intent and frustration by the content of their interaction. This can be accomplished through voice analysis (tone and words) or text analysis of their written communications.
CS Process Analysis: Using AI/ML to monitor all customer interactions to improve the overall process and the next steps in the process chain.
Survey Processing: AI/ML models that automatically track, score, and process brand, product, or customer surveys. This brings quicker insights and reduces the overall costs of the process.
Screening: AI/ML models can sift through candidates to determine which candidates are most likely to fit your criteria and produce positive business outcomes. This saves recruitment time/costs and reduces recruiting cycle time.
Search: AI/ML model can help you search through candidates quickly to determine the best match for a company and given candidate criteria.
Cycle Time: AI/ML can determine how long it will take to fulfill a given role, given past performance, market conditions, and job criteria.
Search Sizing: AI/ML models can predict how many candidates must be in your pipeline to make a hire, given several factors - past performance, job criteria, market conditions, etc.
Diversity Search: AI/ML can help source candidates meeting diversity requirements based on non-traditional data points. (e.g., pictures).
BOTS: AI/ML BOTS can help candidates locate the proper role for their given skill set and reduce time spent by recruiters moving candidates through the recruiting process. This reduces overall costs and increases the candidate's satisfaction with the process.
Performance: AI/ML models can help track and predict candidate performance based on company objectives.
Retention: Predict which employees are at risk of churn and deploy a strategy to retain them before they churn.
Workforce Sentiment: Monitor the overall sentiment of your workforce. Are they engaged or disengaged from the company's mission and objectives?
Health & Life Sciences
Autonomous Surgery: AI/ML models can enable automated surgery given several inputs – patient dimensions, scans, health data, and best practices. This reduces the chances of human error, increases procedure efficiency, and drives down the overall costs of healthcare.
Patient Recommendations: Analyze patient data (tests, procedures, personal health devices) to predict ailments and recommend treatments and actions (treatments, medications, etc.).
Drug R&D: Using AI/ML models to reduce R&D costs of analyzing drugs.
Drug Effectiveness: AI/ML models can help predict drug effectiveness given a large volume of patient data. Determine how one drug or medical device compares to another in treating disease.
Triage: AI/ML model can help in real-time triage and prioritization, ranking patients based on patient data, capacity, etc.
Early Diagnosis: Predict chronic conditions and recommend treatment plans based on patient data.
Monitoring: AI/ML and IoT devices can monitor patients in real-time and alert medical staff to issues.
Medical Imaging: AI/ML models allow you to detect anomalies from images to predict diseases.
Gene Editing: AI/ML models can model the human genome and predict the impact of gene edits. This could help us eventually remove disease-causing sections of our genome and improve patient care.
BOTS: AI/ML BOTS can answer basic medical questions, help with insurance questions, schedule appointments, etc., saving staff time and costs.
Finance & Legal
Fraud Detection: Detect unusual and potentially fraudulent activity based on user behaviors across various systems.
Credit Application: AI/ML models can help you choose the best customers to target for loan products based on predicted risk profiles created with credit data.
Credit Underwriting: AI/ML model can automate the process of underwriting based on available financial data. This saves money and increases the throughput of claims processed.
Insurance Quotes: AI/ML models can determine the optimal price quote based on historical data and customer risk profiles.
Algorithmic Trading: AI/ML models can assist in making automated trading decisions based on market data and risk profiles.
Expenses: AI/ML models extract data from receipts for tracking and reimbursements.
Compliance: AI/ML models can scan legal and regulatory text to ensure compliance across all content creation channels.
Contract Review: AI/ML models can review contracts and identify risky areas for further analysis, saving expensive and time-consuming human-based reviews.
Audit: AI/ML models can assist public audits/tax returns by detecting anomalies in financial data. This reduces human effort and the cost of human review while increasing accuracy.
Automated Discovery: AI/ML models can quickly review unlimited discovery items and predict areas of interest for further review.
Predict Legal Outcomes: AI/ML models can predict outcomes based on historical ruling data and legal inputs (e.g., discovery items, motions, etc.). This is a valuable tool in determining legal strategy, including settlement.
Legal BOTS: AI/ML bots that can help answer simple legal questions.
Improve Space Designs: AI/ML models can predict which spaces will be most used for your given target profile. This helps construction companies improve their designs before construction begins resulting in a more satisfied customer.
Process Improvement: AI/ML models can be used to analyze data from past projects and identify potential areas of improvement, such as reducing material waste or streamlining the workflow.
Jobsite Safety: AI/ML models can determine safety improvements (for example missing hard hats on site) through imagery.
Risk Mitigation: AI/ML models can identify risks, measure their impact and use predictive analytics to help reduce risks, including suggesting known mitigations. This ultimately helps construction managers streamline their work and reduce expensive delays, issues, etc.
Structural Analysis: AI/ML models can n detect structural anomalies by analyzing data from past projects. For example, it can look for patterns in the design, material usage, and construction of past projects and use that data to identify potential weak points or areas where there may be a risk of failure.
Environmental Impact: AI/ML models analyze the environmental impact of a construction project. It can be used to detect patterns in data related to water and energy usage, emission levels, and toxic materials. This helps contractors make more informed decisions about their construction processes and reduce the negative impact of the project on the environment.
Predict Project Parameters: AI/ML models can be used to improve the planning process for construction projects. For example, it can analyze data from past projects to create accurate estimates of the costs, timeline, and resources required to complete a project. This helps contractors to better plan and allocate resources, reducing the risk of cost overruns or delays.
Automate Building Costs: AI/ML models can analyze data from past projects and identify which materials or combinations of materials are the most cost-efficient, structurally sound, and environmentally sustainable.
Automate Building Inspections: AI/ML models can create reports that provide in-depth insights into the project, enabling contractors to better track progress and identify potential problems before they become an issue.
Material Prediction: AI/ML models can precisely predict the amount of needed materials and improve resource allocation.
Scheduling Repairs: AI/ML models can analyze data from previous projects to build a predictive model for how long certain repairs will take and make recommendations about when repairs should be scheduled. This helps minimize downtime and cost overruns and reduces the risk of potential safety hazards.
Delivery Coordination: AI/ML models can anticipate which materials need to be delivered and when they should arrive on site. This helps streamline the process and ensure that projects stay on track and within budget.
BOTS: AI/ML bots can be used onsite to aid in decision making based on the companies internal data and best practices.
Yield Analysis: AI/ML models can predict yield based on several inputs - historical data, soil makeup, weather, seed types, and chemical treatments.
Yield Optimization: Given the data above, AI/ML helps you facilitate what-if scenarios to choose variables that will give the best chance of maximizing output and profitability.
Market Optimization: AI/ML models can assist in making crop decisions based on market conditions for each crop type. Maximize profitability given limited natural resources.
Automated Harvesting: AI/ML can enable autonomous crop planting and harvesting machinery. This increases efficiency (24-hour operation is possible) and decreases the overall costs to harvest.
Automotive & Transportation
Self-Driving Cars/Trucks/Planes: AI/ML can enable autonomous transportation. E.g., Telsa. This will increase the efficiency of travel while also reducing fatalities.
Transportation Safety: AI/ML models, in conjunction with advanced radar and sensors, warn operators of imminent dangers (e.g., a deer on the road) and take evasive action when operators are not cognizant of the issue (e.g., braking).
Driving Assistance: AI/ML models that assist driving (e.g. lane keeping systems) reduce accidents and insurance costs.
Air & Rail Traffic Control: AI/ML models can predict eventual route conflicts and propose/notify course corrections to avoid conflict. When an immediate collision is imminent, models can notify parties to take evasive action.
Demand Prediction: AI/ML models can predict the energy requirements of the network based on a number of variables.
Energy Production Optimization: AI/ML models can help energy companies' produces the required electricity at the cheapest possible price point given market fluctuations and demand requirement.
Investment Analysis: AI/ML models can help you predict the best acquisition target based on a number of variables - historical performance, company data, risk profiles, etc.
See how others are winning...
Learn why organizations are getting better outcomes using data and how Macula can help.