CX.AI -Experience the new generation of CX Insights

All Posts By

Lilla Szücs

#13 Interview

CX Enterprise Platforms vs. Specialized Solutions

CX Enterprise Platforms vs. Specialized Solutions

 

Enterprise typically choose so-called enterprise solutions, like we’ve got Qualtrics or Medallia or InMoment. Still there is a universe of specialized solutions just like CX-AI.com. Would you advise to ignore it?

 

Good question. If you want to get the best out of it, if you really want to drive impact, you should think of those specialized solutions for a simple reason: it’s possible to use those systems.

 

Qualtrics are not closed systems. They all have docking stations where you can basically plug, with any kind of other software systems. If you look into this, these systems as any other platform are not the best ones in the market now. So they have good standard modules. So for instance, if you look at those two pieces:

 

they have a text analytics model and they have a key driver analytics model. You can take those modules from the platforms and try to use it.

 

Those the text analytics modules are just supervised learning. It will not be sufficiently precise compared to what is possible. The key driver analysis module is a simple regression (invented 100 years ago). It’s neither capturing non-linearities nor capturing indirect effect. It’s not a causal engine.

 

If you do the comparison you will find that the validity of taking specialized solutions is four times higher.

 

This is just a number, but actually you can see it in real life. If you just use out of the box solutions, you often get those strange results. For instance, positive things like friendliness suddenly have a negative impact. Then becomes strange because you cannot explain this anymore. It becomes obvious that the methodology is, missing something now. Therefore it is a good mix to used what you have, the enterprise system, and plug and play on a specialized solutions for that.

 

The best example for such a solution is www.cx-ai.com.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#12 Interview

How to Setup Text Analytics When Having Multiple Languages?

How to setup text analytics when having multiple languages?

 

There are two things you can think of. I always recommend supervised text analytic systems and there are two ways to train them.

 

One way is that you auto translate everything into a main language. For instance, English, then have an English coder, teaching the AI. That’s one way.

The other way is to have another native coder and teacher. Everyone teaches the AI in native language.

 

Both approaches has pros and cons. There is no way which is just better. The con of the translation is obviously you loose some information, while translating it , but the disadvantage of having native teachers is that you cannot make sure that they really understand every category in every topics, the same way. You cannot make sure that they really code the same way. If you end up at least with more than three languages, we recommend to auto translate into one language.

 

That’s typically what we do and actually what’s not so well known so far is that there is an even much better translation machine than Google translate in the market it’s called Deepl.com.

It has been proven to be much more precise than google translate and for all systems, we use that machine to auto translate every single one language.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#11 Interview

Is Advanced CX Analytics Applicable for B2B or High Value Niches?

Is advanced CX analytics applicable for B2B or high value niches?

 

So does advance CX analytics also work in a B2B company?

 

Absolutely. What’s different in B2B? There are maybe two things which come to my mind. Language is very specific and second sample size is low. As explained a supervised learning approach is tailored exactly for specific language. So if you run this methodology, you can really train like an domain expert, i.e. having an AI categorizing like a domain expert to a B2B. Regarding sample size, of course, if you have very little data, it’s probably not enough, but there are many tricks where you can work with it and if you are a sizeable B2B company, typically you have hundreds or even thousands of feedbacks. Enterprise B2B typically have enough data.

 

One of those tricks and trades to work with lower sample size is so-called split analysis. You take the whole data set and split it in certain subset of your customers, which you want to research on. You model with all data but overweight the split. That handles the instability caused by the low sample size and smooths out by the larger dataset.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#10 Interview

The Role of Context Variables in CX Analytics

The role of context variables in CX analytics

 

So what’s the consequence if you’re not handling the role of context in modeling.

 

The consequence could be for instance, that your insights are biased. So I think I gave you the example recently. Customers with higher income are often more critical and they may give different reasons in their explanation. This lead to negative correlations between those reasons and the rating, although the reason may have positive impacts. If you have the context information in the model, your results become cleaned and true.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#9 Interview

How Does CX Analytics Deal With Sizing the Importance of Topics?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: May 18, 2021 * 9 min read

Contexts information should be a part of analytics. If gender influence what’s important, it should be in the model, it is then a variable saying “one”, it’s a woman, “zero” it’s a man and this variable is influencing the outcome.

So with this context variables in the model, the model can learn, can attribute this outcome to the context, or whether it is really driven by the topic. If you could give the model the information, it can find it out by themselves. That’s the intention to use the data you have. You may argue, Oh I don’t have the data for certain contexts information. You will never have a perfect data set. We are not in the business of proving the ultimate truth. We are in the business of making better decisions.

 

That’s basically what we can do. We can take some data that we already have, and we can work on having better data tomorrow, but everything comes with a price, the price to learn about the ultimate truth is very high. It has the best ROI to get insights that are just closer to the truth, than everything else so far.

"CX Standpoint" Newsletter

b2

Each month I share a well-researched standpoint around CX, Insights and Analytics in my newsletter.

+4000 insights professionals read this bi-weekly for a reason.

I’d love you to join.

“It’s short, sweet, and practical.”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

#8 Interview

Minimum Number of Training Examples Needed to Feed Supervised Text Analytics?

Minimum number of training examples needed to feed supervised text analytics?

 

When it comes to a supervised text analytics, is there a minimum number of training examples that needed to be fed to the AI engine?

 

There has been huge progress made in the last years. If you have heard and have had an opinion formed a few years ago, It’s not true anymore. Actually there are supervised learning systems, which are already categorizing from scratch quite well. You really need to teach them only the specifics of your business and the context of your very question. Typically when it comes to B2B, but also and certainly B to C domains. If it comes to product names or some abbreviation or slang or sometimes it’s also the context and even sarcasm, everything can be trained. Sometimes also the answer depends on the question, that’s why it always needs some training be perfect and human equivalent. When we look at our projects at CX.AI, we typically teach 300, 500 or max 700 examples. Then the machine can categorize infinite examples automatically.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#7 Interview

Customer Audio and Video Feedbacks

Customer audio and video feedbacks

 

What about the other fancy methods that we have around like audio or video feedback or something along the lines of conversational surveying?

 

That’s also on the rise and some companies do it and actually the basic language of everything IS “text”. If you record an audio, transcribe it and feed it in the same process I just described. If it’s video, the main information is also audio, it can be real-time transcribed and fit in the same process. You can in parallel run through other algorithms, which you can also tap on in real time, which is detecting the emotion on the voice. You can see the emotion in the face of a video. All this are available real-time cloud services, where you can, in addition to what they say you can give an algorithm what customers feel using voice or facial information. This can be captured and analysed with your CX analytics. The processes is the same, but in any ways, I think it’s worthwhile to pursue that because, over audio and also video people tend to talk more than in open-ended text fields.

 

It is similarly effective to active listening, which you can also do over audio interview. You can also have a machine asking questions – it’s not rocket science anymore. It’s done as we speak. So if someone wants to build something like that, that’s done in a short set up period.

“Selling Insights Internally” PLAYBOOK

5 Secret Strategies To Better “Sell” Insights To Internal Business Partners.

Not ready yet?

Watch here “How it Works” video.

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#6 Interview

How to Get Customers to Give More Feedback in Open-ends?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: May 18, 2021 * 2 min read

Customers tend to not writing much, let’s be honest in open-ended text fields. Isn’t that a big problem? And if it is, can we actually fix it?

It is a problem, depending on the context, it’s a big problem actually. So for the analytics per se it can be mitigated by just collecting more data.

 

It doesn’t tell you exactly what for person X or person Y what’s their key pain point are, but it can analyze for all customers or certain segments, what’s their key criteria to act on are. For customers or visitors, if they don’t write much, you need to collect more data.

 

One way of collecting more data is to collect more data at one person. There are so-called active listening approaches. What the active listening does in the NPS survey is, it takes the customer text feedback sends it to the deep learning tech service platform in real time, categorizes it in real time. The based on rules the bot reacted. If it’s category is friendliness, it says: “Hey, I’m a bot. I understood, you’re not satisfied with friendliness. Did I get it right? It would be great if you can explain in more detail”.

This probing, which is adaptive, has proven to be very effective, because the response really feels okay. Customers feel “This is not an preprogramed bot here, but they feel understood me somehow, or they feel the bot didn’t get me, which has even the same effect. Customer feel: Oh, he didn’t get me. I need really to explain better. So this, has proven to get double of feedback and also improve the predictive power by another 50%.

"CX Standpoint" Newsletter

b2

Each month I share a well-researched standpoint around CX, Insights and Analytics in my newsletter.

+4000 insights professionals read this bi-weekly for a reason.

I’d love you to join.

“It’s short, sweet, and practical.”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI

#5 Interview

On Key Driver Analysis for CX

On key driver analysis for CX

 

Coming back to that text feedback and the causal machine learning applied. How should companies reliably deal with?

 

You could sit down and just categorize feedback by hand, which is fine as long as you have enough time and the resources to do it and you find people who never get tired to do this. The letter this basically a problem. You, as a human, you get tired. Also, when categorizing you learn and you change your mind over time. If you use another person for categorization, everything gets inconsistent.

 

But it should be consistent and reliable. This is where machine learning AI comes into play. There are basically two types of Texts Analytics AI. The systems you may know are unsupervised learning because 95% of systems in use are that. They basically find clusters themselves without training and human intervention.

 

They find meaningful clusters and these clusters are automatically named and without your contribution, they find it , so it’s very easy to do. BUT: We did a big study around it and found that this approach just has half of explanatory power, compared to when you categorize manually.

 

You don’t want to lose information. You want to drive important decisions from that. Just by the fact that you don’t have enough coders, you cannot sacrifice the quality. That’s at least my recommendation.

 

We searched further and further, and there are so-called supervised learning approaches, which is kind of a deep learning methodology where you train an AI to code like a human. You may know that for instance , Google can spot cats from pictures. Why? Because humans told them what’s a cat and what’s not, nobody knows how the Google deep learning does it actually, because it’s super complex it has millions of parameters, but it’s simply a black box predictor and that’s basically the same approach, you can do it with text. Finally we found the technology, which is often times even better than manual coding, because it never gets tired.

Also produce sentiment scoring, which is better than what humans can do. Typically humans are good at identifying the meaning. But to really say the tone of this sentence is very positive, very negative. It’s hard for humans to switch between the rational and the emotional part. The recommendation I would do, is to really use the best supervised learning approaches to categorize your text feedback, this means also the machine needs to be trained. Someone needs to sit down and teach it.

 

No worries, we do it for you. 😉

Free "How-to" Video on the CX.AI Method

IMPRINT

Our Group: www.Success-Drivers.com

Privacy Policy

Copyright © 2021. All rights reserved.

#4 Interview

How to Find Key Initiatives to Drive CX?

Founder of CX-AI.com and CEO of Success Drivers
// Pioneering Causal AI for Insights since 2001 //
Author, Speaker, Father of two, a huge Metallica fan.

Author: Frank Buckler, Ph.D.
Published on: May 18, 2021 * 9 min read

How is it reliably possible to detect the most important improvement strategies? What you need to do is to first off quantify what people are saying. It means basically you count what is said and put it in buckets.

Once they say, “Oh, this guy said friendly service and buggy software”. That way you take on two things of possible 50 or 100 things, he could have said. That’s the first thing you need to do, to quantify what people are saying.

 

The next step is then across the whole sample, identify which of those top topics are predictive and are uniquely mutually and have the information to explain outcome, which is the customer experience and later churn and upsell.

 

When doing that, you also need to apply the right methodologies. So if you take simply regression, out of your statistical software, this is too simplistic and this approaches assumes independence of drivers and linearity and often things the causal impact is indirect.

 

Let me give you an example. If you have a category, which is great service, and you have a category, which is friendliness that may drive great service and great service then may drive customer experience. At the end, you can use great service to predict, the customer experience, you don’t need friendliness, but still friendliness drives great service.

 

It has an impact and important impact eventually. So you can only reliably measure the importance of topics if you consider it also the relationships between the topics. But still, this is not rocket science. This is established methodology, causal machine learning. With this, you can reliably understand what’s important and what’s not.

"CX Standpoint" Newsletter

b2

Each month I share a well-researched standpoint around CX, Insights and Analytics in my newsletter.

+4000 insights professionals read this bi-weekly for a reason.

I’d love you to join.

“It’s short, sweet, and practical.”

Big Love to All Our Readers Around the World

IMPRINT
Our Group: www.Success-Drivers.com

Privacy Policy
Copyright © 2021. All rights reserved.

FOLLOW CX.AI