April 2017 technology review

after World Mobile Congress and 4YFN

Innovation Collider SRO, Piotrowski Adam

After attending 2017 4YFN and speaking with more than 400 startups we would like to report some interesting trends in services and solutions available today for banks, insurance companies and telecoms.
This article is meant to be a light read so we’ll try to stay on point as much as possible and give you our perspective on how technology is changing traditional financial services.

Full disclosure: Innovation Collider is a company specializing in sourcing and implementation of many innovative digital strategies. We partner with companies which we believe would bring biggest improvements to our clients. Although we enter into revenue sharing agreements with them we are not bound by any retainers so we are free to choose who we work with.

Innovation Collider visited last year’s editions of 4YFN and Mobile World Congress in Barcelona. Those two amazing exhibitions attracts both visionaries and lunatics and unfortunately it’s hard to tell them apart. While some trends especially in sales automation are getting real traction, others such as AI assisted chatbots, beacons or virtual reality will need at least couple of years to find real use in most common scenarios.

Compare to last year we noticed that

Sales automation is on the rise (again, and again, and again…)

The idea is to provide solutions that not only identify opportunities but also give a robust tools to convert them with minimal effort and resources.

Those tools have to integrate with existing infrastructure and third party solutions. The key is to add value on top of what already exist instead of creating overlapping channels.
A good example of this is focusing on customer touch points such as opening new marketing channels and substituting work-intensive human-to-human conversations with programmable front-ends. 

You might assume that human-to-human interaction would always be preferable but we are observing a different trend among millennials. Call center has to dedicate human resources and the client has to invest his/her own time to receive those calls. A minute of time changes its value throughout a day. It’s the most expensive when we are at work or with family but has little value when we are stuck waiting for a train. 

We still think that the customer should be in the center of all operations and chat communication could allow you to still treat every client individually. Recent developments in deep learning and customer tracking will allow you to stop segmenting your clients and simply create a segment for each and every client.

Those interactions should be set up in such a manner so the customer can skip them, opt-out or at least be able to let know that they are not working ie annoying. This information is key to diagnose inevitable problems and define the evolution vectors.

Intelligent chatbots, contrary to their name, are still kind of dumb, but we are getting there slowly.

In fact artificial intelligence is progressing quicker than you probably imagine. Recent improvements in algorithms but also the emergence of cheap raw computing power are absolute game changers in this field.

Most AI use deep learning and neural networks which aren’t new concepts (1980s) but what has changed dramatically over the last year is the emergence of easy to use tools, which can tap into data streams and provide more accurate results than humans would be ever able to.

This wasn’t possible even 4 years ago and it has caught even the founder of Alphabet Sergey Brin by surprise. It still too early to predict the true impact of Artificial Intelligence on fintech and business in general but we are convinced that its impact will have both positive and negative consequences.
Increased efficiencies may cause supply increase on the labour market(unemployment), but on the upside companies which jump on that technology early will have increased capacity and could gain a substantial competitive edge. A good place to start the search for solutions in this area are chatbots. They can support or substitute humans and are relatively cheap to implement. 


The problem with them isn’t the technology but rather lack of training data which would make them truly interactive, seamless and life-like.

Companies have a ton of voice recorded conversations with their clients from call-centers but they simply can not be used to train chatbots. That’s because voice communication is fundamentally different from text chats. Think about how you would conduct a sales call with a client. Now imagine accomplishing the same task via WhatsApp. Text chats have different structure and are more to the point and less conversational.

They will become a major part of customer interactions as soon as enough human assisted text-based conversations have been recorded and analysed.

To achieve that, we recommend opening those communications channels as soon as possible.
Especially that their true effectiveness can be assessed during a test phase by establishing two user groups - one supported by chatbots and one without that support.

There are more benefits as well. It’s preferable for a growing segment of your market - a new generation which communicates using WhatsApp and other chat apps on the daily basis, and doesn’t respond well to other methods. Also since less resources are required to chat with clients than to actually speak with them on the phone, one sales consultant can concurrently serve 5 client on the chat, compared to only one on the phone. 

Clients will also appreciate no waiting time which plagues most call centers resulting in high balk rates and lost business. Another benefit is the ability to leave the conversation and pick it up again at any time especially when talking on the phone is simply not possible (for instance while commuting)

Inactive conversations can be reactivated using other methods such as interactive videos, discount codes and other retargeting efforts with great results.