New obligation for financial advisers: Record investment-related calls with clients. SentiSquare AI analyzes the calls to help better understand customers._

Published on 18|5|2020 by Lucie Kolářová & David Radosta

Czech financial and investment advisers have a new regulation to comply with. From 1st May 2020, they are obliged to record telephone calls with clients in which they deal with investments. That way, CNB aims to ensure financial advisers and intermediaries comply with rules. Most financial advisory companies and advisers were against this. It is a burden both procedurally and financially; in addition, they may need to employ people to evaluate the calls. That is when AI comes to the rescue, transcribing calls, evaluating them and structuring the data to allow for searching for specific information. As a result, financial and investment advisory firms will gain a new world of insights about customers that would otherwise be difficult to access.

SentiSquare Call Analysis produces vital insight for a fraction of the price compared to traditional solutions, and relieves employees from repetitive work.

One of the use cases for SentiSquare Call Analysis are investment-related conversations between advisors and customers. The analysis will be performed as an on-premise solution, ensuring data protection compliance. The SentiSquare Analytics tool structures and visualises the output, also providing semantic search in the transcribed calls. The project thus focuses on monitoring investment calls in terms of compliance, topic detection and easy search the data.

What the customer gets:

1. Off-line speech recognition (convert audio recordings to text).

2. Off-line processing and analysis of call transcripts.

3. Automatic topic detection and keyword extraction (for a quick overview of what the call is about).

4. Automatic categorization of calls into predefined classes (with a focus on the topic of investment).

5. SentiSquare Analytics for easy semantic search and insight into transcribed calls.

SentiSquare Analytics will not only analyse calls, but also help you to better understand your customers in terms of:

1. why are they reluctant to buy a specific product

2. have the benefits of the product been explained detailed enough

3. what alternatives the customer is thinking about

4. what risks does he see

5. how is your agent dealing with customer concerns

All this insight has an enormous value when it comes to train your agents and to improve customer awareness and your product understanding.

How does it work?

Reliably: SentiSquare Analytics user interface is a proven system for working with text data.

Safely: The analysis is performed on the investment company's servers.

Smartly: Provides semantic search in transcribed calls.

Simply: This makes it easy to detect, track and search for all investment calls

Quickly: Deployment within 4 weeks

Algorithms from SentiSquare are based on distributional semantics. This allows the algorithm to understand the meaning of words from raw text only (i.e. without human supervision). Two words are semantically similar when they appear in similar contexts (they are similarly distributed across texts). Meaning is represented as a vector of real numbers. Geometrically speaking, meaning is a point in a high-dimensional semantic space. Words that are semantically similar tend to be close in semantic space. This way, algorithms understand the meaning of any text, even when expressed in different words (synonyms) or in a different style.

Want to know more?

Contact us at tichy@sentisquare.com or call us at 📞 608550338.