Are you curious about how Artificial Intelligence (also called Machine Learning) can help your business? Below is a list of proven and practical techniques that can be applied now.
What problems can Artificial Intelligence solve in your business today?
AI Techniques for Online Stores, Membership Businesses and Retailers
- Minimise Customer Loss (aka Churn Reduction) – Does your business sell repeatedly to the same customer again and again? Machine learning techniques can scan your data and give you an early warning about which customers are most likely to leave you.
- Maximise Upsells (aka Product Recommendation) – Do you have an online store that sells a broad product range? Let us help you increase revenue by using all the data points available about your customers as “signals” and training an algorithm that suggests those upsell products most likely to result in additional purchases.
AI Techniques for Professional Service Firms (e.g. Legal, Real Estate etc)
- Price Prediction (e.g. What will this house sell for?) – There are a number of well respected models in Machine Learning to predict the sale price in open market situations such as Real Estate. The same techniques can be applied to predict legal settlements such as the allocation of assets in a divorce.
- Lead Prioritisation (e.g. Who’s most likely to buy?) – Does your business deal with a regular flow of inbound leads and need to prioritise who to spend the most time with? In many cases we can train a model to predict which leads are most likely to convert into customers.
AI techniques for medium to large corporations (hospitality chains, banks, airlines, hotel chains)
- Complex Human Decision Automation (e.g. Should Bob get a loan?) – Does your business rely on people to regularly make the same complex decision like approving a bank loan based on a person’s financials? We can take your historical data and train algorithms to automate the process for you.
- Anomaly & Fraud Detection (e.g. Who’s trying to rip me off?) – Is fraud detection a challenge for your industry? Do you suffer from the “needle in a haystack” problem where there’s a high volume of transactions and you need to find those few that are suspicious quickly and automatically?Let’s work together and train an algorithm that will not only tirelessly scan your data but will get smarter over time as you give it more cases to learn from.
- Price Optimisation (aka “Yielding”) – Do you sell a service (like accommodation) or a product that quickly expires? Are you willing to vary your pricing to maximise overall profit? Machine Learning models can be trained to analyse your historical data and provide a constant real time suggestion of best price to maximise revenue.
- Resource Demand Prediction / Inventory Optimisation (e.g. How many staff will we need this weekend?) – Machine Learning can be used to help predict demand for key resources such as staff or stock helping you to maximise sales and minimise wastage.
General purpose AI techniques for use across a range of industries
- Image Classification (e.g. Sorting images of anything – from faces to fruit!) – In the past few years machine learning algorithms for classifying images have become incredibly powerful. Whether it’s classifying fruit in agriculture or people in security, all you need is a training set of 500+ images and it’s likely we can automate the classification process for you.
- Free Text Analysis (e.g. Automatically analysing customer reviews?) – Does your business have access to large volumes of inbound information in unstructured text format like customer reviews or tweets? Machine learning can automatically classify and identify key themes in your data allowing you to keep your finger ‘on the pulse’ of customer sentiment.
- Chat “Bots” (Halve your customer service costs) – Do your customers ask the same types of questions in lots of different ways? Why not offer your customers the convenience of a real time, interactive chatbot trained to answer all your common inquiries.Advancements in natural language processing mean that today’s chatbots can link deeply with your data and internal systems. And, they’re smart enough to know that “do you have any runners” means the same thing as “got any sneakers for sale?”.
- Voice Recognition – Google, Amazon and a number of the major tech companies now offer high quality voice recognition services (think Siri, Cortana and Google Assistant). Did you know that this voice processing can also be applied to your apps for both receiving instructions and delivering a response.
How is machine learning different from “regular programming”?
In standard programming, we interview industry experts (typically our clients) and they tell us the rules we need to encode in a system. In Machine Learning, we use specially selected algorithms to pour over historical data to extract these ‘rules’ automatically.
These powerful algorithms can often find new patterns and make predictions with greater accuracy than even the best experts in an industry, and once setup these systems run at near zero cost.
What influences the cost of using AI?
There are two key factors that commonly affect the cost of an AI project. Firstly, we need to determine whether your requirements can be met using an off the shelf AI tool from a provider such as Amazon or Google. If that’s not the case, we’ll need to train an in-house ‘model’ to suit your needs. Off the shelf tools are typically cheaper, although they generally incur an ongoing per use cost.
Secondly, the readiness of your data will affect pricing. In most cases, we need initial data to train our algorithms (either in-house or off the shelf) and we’ll often we spend significant time helping clients access and “clean up” their data so that it can be processed by a machine-learning algorithm.