Please click on the link below to activate the content. Please note that following activation, data will be transmitted to the relevant provider.
Do you know how many inquiries you receive each day?
Experience shows that two thirds of the inquiries a service centre receives are simple routine questions: Where's my parcel? Are you open on Saturdays? You need someone who can systematically answer these types of questions without delay - leaving everyone else free to concentrate on their actual job of giving advice!
Find out more in the video!
digicom ai answers standard questions autonomously.
The focus is on the customer. Sometimes he just needs to quickly get some information, sometimes he needs individually tailored advice. digicom ai structures this communication requirement and answers simple questions immediately. Direct, precise answers increase customer satisfaction. digicom ai forwards complex consulting tasks to your agents as accustomed. This is strategic delegation - it's efficient and occurs in real time.
Which types of questions does digicom ai understand?
- The customer asks questions using natural language and isn't forced to stick to any set phrases or answer multiple choice questions.
- The AI logic used by digicom ai masters the complex rules of German grammar.
- The system currently "understands" 18 additional European languages in Latin script.
- As a rule, digicom ai analyses text in written form. The voice feature, which is currently being developed and tested in the laboratory, transcribes voice messages into text form and answers them in a second step via speech output.
digicom ai reaches reliable decisions based on confidence levels.
Over twenty years of experience with customer dialogue systems has taught us that in service the precision and reliability of your answers is critical for customer satisfaction. To this end, digicom ai first computes the so-called confidence level for each answer, i.e. the likelihood that this answer is correct and is helpful to the customer. If this value falls below a configurable threshold, the question is passed on to an experienced member of staff. This creates a close collaboration between virtual and human agents.
What tasks can digicom ai undertake autonomously?
- digicom ai can answer up to 80% of standard questions, depending on the data quality. The ratio is lower at the beginning and steadily increases by virtue of the continuous learning process. Your service centre staff continue to deal with questions that have legal implications and handle personal consultations.
- digicom ai can be integrated into every part of your communication setup and into every application without media discontinuities.
- All processes run unobtrusively in the background - without additional apps, without registration and without any other hurdles.
- Every contact with your customers is archived and logged by digicom ai and appended to their customer history making it available for evaluation by your systems at any time. The reason for this is, of course, that the AI has to learn.
Where does digicom ai get the answers?
- There's a short initial training phase during which we teach digicom ai the most frequent questions and answers encountered in your daily workflow.
- Following this initial learning phase, digicom ai finds "internal" answers from the data stored in your systems and the conversations documented at the service centre.
- Direct responses provided by the customer ("Was this answer helpful?") are taken into account as well as feedback from the agents.
- External neural networks are only used as a last resort.
digicom ai transforms your data into useful knowledge.
Your data is stored across many systems. digicom ai first uses it to generate information: for example, the information that a package was dispatched to a customer at a particular time. digicom ai's artificial intelligence derives knowledge from this by automatically linking this information with the customer question "Where is my delivery?". What's more: digicom ai profits from the wisdom of the experienced agents and is constantly learning from their responses too.