Do voice analytics and automation have a place in complex sectors?

We sit down with Jonathan Drechsler, Head of Business Development at Recordsure, to understand the role automation can play in highly nuanced business interactions.

One of the things we get asked about most at Recordsure is the art of the possible. Where does technical feasibility end and science fiction begin?

We’ve all seen the headlines about how much time is lost by organisations each year to red tape and admin tasks. An independent report in 2017 found that office workers lose a third of their week to admin, although of course no two businesses are the same and this figure often sky rockets in more specialist areas. For instance, we’ve spoken to data analysts who are spending 90% of their time on data collection and organisation, with only 10% of their time left for the actual analysis where their expertise is truly needed.

A new way of thinking

Automation and AI work best when applied to repetitive tasks which follow similar patterns. This has meant that in speech analytics, a lot of the application has been focused on automating simple conversations like appointment bookings, chat bots or quick enquiries that follow a rigid, repetitive formula. This has fuelled a misconception that automation becomes less effective the more complex the case is.

Historically, this may perhaps have been true, but things have changed in recent years with new applications of technology that can make departments significantly more efficient. The key? Using technology to make team members more effective, which is an important distinction from using technology to replace them.

There are a number of key areas technology can take the weight of admin off people’s shoulders allowing them to focus their time and energy on the areas they can truly shine:

1. Standardising different channels into one compatible format

Before you can analyse data, you first need to compile it. Odds are that multiple channels have been used in communication, which is brilliant for customer care, but can be a nightmare for reporting. When your data is spread across phone conversations, email, livechat, face-to-face meeting recordings (or more likely, hastily scribbled meeting notes), simply organising it is a time consuming job in itself.

This can be automated though. Audio files from phone and meeting recordings can be transcribed and combined with written records like emails and chat bots. Digitising hand written documents can then create a comprehensive cradle to the grave account of all professional interactions which will save time on the prep work needed before a review can be conducted.

2. Identifying irrelevant data which can be ignored

Again, that balance between a good customer experience and analyst-friendly data. Any free flowing conversation is going to cover a lot of ground, and a lot of it simply won’t be relevant to your audit. If you’re checking whether a customer has been mis-sold a product for regulatory purposes for instance, you won’t need to trawl through the security questions or data collection portion of the conversation.

There is a lot of power at your fingertips now to avoid these traps. Machine learning technology can be used to review transcripts and discard any that are irrelevant for your purposes, then segment topics of conversations into key themes. You can now instantly ignore interactions that aren’t relevant to your goals. On a more granular level, you can review a conversation and automatically break it down into these segments allowing analysts to focus purely on the elements which are actually relevant to them.

3. Extracting key details

Consider the question “how much do you earn?” It’s one that will by necessity be asked in conversations on everything from loan applications to recruitment. Historically, it’s posed one significant issue for data analysts: there are hundreds of different ways you can ask it.

This means that through traditional methods like keyword search, extracting this information would require a huge amount of manual work. Times have changed though, and machine learning technology can now mimic the way a human understands conversation to allow automatic data extraction. Rather than having to break the rhythm of a customer conversation by scribbling notes or going through a full meeting recording afterwards, the information is now available at the click of a button.

4. Zoning in on conversational trends and anomalies

You want to have flowing conversations with clients rather than following a script, but there are still certain patterns natural conversations will tend to follow. Using speech analytics tools to understand these patterns can be a game changer.

An obvious example is during the sales process. If the executives who are listening to the prospect’s needs are getting better results than the ones more focused on describing the product, then trends like this can be instantly highlighted and real-life examples of best practice shared instantly to help train the team. You can also bring in sentiment analysis to see how a prospect is engaged over the course of the conversation to enhance messaging and maximise conversions.

Being able to monitor this efficiently allows you to review 100% of meetings rather than random sampling or having to use mystery shoppers. Similarly, we have used Recordsure’s voice analytics platform to successfully identify potentially vulnerable customers based on their behavioural traits thus providing firms with a robust way to target those in need so they can dedicate additional support where it is needed.

5. Automating internal process reviews

Reviewing activity on a macro level is a job in itself. This review process can be made considerably more efficient by box checking key credentials automatically. There will be multiple pieces of information that you need to retrieve as well as provide during professional interactions: did you collect all the personal details required from the customer? Did you complete all your due diligence? Were the necessary security processes followed?

These will be different from business to business and from role to role, but machine learning technology can be employed to allow an instant indicator of how well these criteria have been met in order to maximise opportunities and minimise risks.

Empower the experts

Many businesses provide solutions to complex problems, solutions which require human judgement that you wouldn’t dream of delegating to a piece of software. In instances like this, the value of AI and automation lies in driving efficiencies that allow human decision makers to use their specialist skills and operate more effectively.

Awareness of this value is surprisingly low, meaning that there is a huge amount of potential going untapped by organisations worldwide. The easiest way to test this is to take a step back and consider how much day-to-day activity requires the full expertise you or your team possess. If you can pull out any elements which feel more of a grind than an art, it’s worth looking into whether there could be a technological solution for them.

Jonathan Dreschler is Head of Business Development at Recordsure. To learn more about the ways Recordsure uses AI technology to empower teams and drive efficiencies, please get in touch.

Sparked your interest?

Find out how our RegTech solution can help you save time and reduce compliance costs.

Book a demo Subscribe