What can I do with artificial intelligence within my organisation?
For over two years now, Pegamento has been implementing AI in various ways and in various orders of magnitude at companies in all kinds of sectors.
It is said that artificial intelligence can 'analyse' data, but in reality, what it is doing is categorising. New (unknown) content is categorised on the basis of data. What kind of data? Data from one or more tables, snippets of text, speech, or images. AI can read and use many different types of data in processes. Examples include pictures, emails, voice messages, and other digital content.
In fact, AI can use all forms of information that we can capture digitally. On the basis of data that contains clearly marked categories, AI makes a prediction.
Now that you know what artificial intelligence does, we will go further into what exactly artificial intelligence is and how it can be implemented. We will also share some case studies.
Here is a quick reading guide:
What is artificial intelligence?
Artificial Intelligence, or AI, is the ability of a system to correctly interpret external data, learn from this data, and use these lessons to achieve specific goals and tasks through flexible adaptation. You can think of this as pattern recognition on a large scale. AI converts human input into input that can be used by computer systems, based on previously detected patterns and a created neural network. It can use relationships between data, and it can also detect patterns in natural language. Interactions with lots of negative words, for example, might be more likely to be a complaint. Interactions with an appointment or reservation are more likely to contain a date. Google Home, Google Assistant, and Siri use speech-to-text conversion to guess what you are trying to say. These applications also work on the basis of previously categorised bits of speech. The prediction indicates to which category a new piece of content should be assigned. For example, that it is a complaint, or an order. It can also identify that you said the word 'Spotify' or the name of your favourite band. You can read more about AI/Machine learning/neural network and RPA here.
What can be predicted?
Predictions come in different forms. Some examples are: the essence of a customer demand, the customer's emotional mood, how important a task is to you or to your customer, the potential value of a new purchase, or how a container can be placed in the most efficient way. These are examples of the types of predictions an AI system can make. The system learns how to make these predictions through training.
Training the AI model
How does the process of training an AI system work? We explain this using an AI mail system as an example. In a workshop with users, we determine which email flows, such as changing address details, placing an order, and swapping a work day, are eligible for being handled by AI Mail Automation and which responses should accompany them. We define categories, such as invoice queries, lost customer cards, or returns, and establish where to find sample messages. Using a link to the email server, we import around 100 sample emails per stream. The AI model analyses the sample emails and trains itself on recognising the categories and recognising entities, such as a phone number or bank account number. We then use test messages to validate and further refine the model.
Mail Assistant, an example of artificial intelligence
What is Mail Assistant?
Mail Assistant is exactly what it says: automation of mail traffic using artificial intelligence (AI). By deploying artificial intelligence, the intent of emails is discovered in a learning process. Once the AI software knows what the emails are about, we deploy smart templates that can handle emails fully automatically. And in instances where this is not possible? The software then recognises this and immediately transfers the email to the right employee - either within the organisation or at an external location.
The advantages and disadvantages of artificial intelligence
Artificial intelligence uses patterns that the AI system has learned through training. While training a system, it is indicated which input is right or wrong based on the best answer. Because an AI system can recognise a problem or customer question, it can give the right answer in most cases once it has been trained. This ensures a higher quality of provided answers and solutions. In addition, the suggestions an AI system gives based on a certain type of situation are consistent, which ensures more consistent problem solving of issues within your organisation.
An AI system can analyse a situation within seconds and, on the basis of training, suggest the best answer. This allows an employee to act faster or eliminates the need for an employee to answer emails, for example. This ensures a more efficient and faster way of working.
Although artificial intelligence offers many advantages, there are also some disadvantages. An AI system is a complex piece of code that takes time to set up and train. It takes substantial time and investment to get a system to act correctly, and it needs human guidance in the early stages. The more complex an intention of an AI model needs to be, the more intensive it is to train it. However, the advantages outweigh the disadvantages. When an AI system has been fully trained, it can act completely autonomously, which brings with it the advantages mentioned above. It is a long-term investment that pays for itself.
Want to read further?
Pegamento implements AI in different ways and in various orders of magnitude at companies in all types of industries. In this white paper, we discuss our findings, recommendations, and implementation strategies
Some client cases in the field of Artificial Intelligence
Using the Pegamento AI Suite, we built a Task Manager. This is a robot that recognises the intent of an incoming call or message in multiple languages and automatically suggests a response to the 1st-line employee. Via a link to the back office reservation system, the employee gets a complete customer overview with all available customer information on the screen. Communication is thus handled faster, better, and more efficiently. Read more about this case here.
For Translink, we developed TRiP, which stands for the Translink Registration Information Platform. This is a complete platform that logs phone calls and emails in the customer services department.
- The customer is helped much faster due to fewer clicks.
- The customer advisor has a better overview of the work.
- Management receives reports containing a dashboard that shows at a glance whether SLAs are being fulfilled and how many customer contacts have been handled in a day.
Read more about this case here.
Rijkswaterstaat identified Pegamento as the party with the best solution for increasing the safety of bridge operation through detection. Rijkswaterstaat had issued a challenge for companies in the market to detect and help prevent dangerous situations in bridge operation through innovative applications. The BCR system can use multiple cameras to track road users using Artificial Intelligence and allows the operator to see what is happening on the bridge deck via a dashboard. You can read more about this case here.
Artificial Intelligence within your organisation?
Artificial intelligence can be applied in a wide range of production processes. Wherever patterns can be found, such as in customer intent, emails, or images, this technology can quickly add value! This technology is innovative, but it is also surprisingly affordable to deploy for process optimisation.
Could it be a solution for your business? What savings or quality improvements are possible in your processes?
Request an introductory meeting with one of our specialists to find out how Artificial Intelligence can be deployed within your organisation.