How to use A.I. and Intelligent Automation to gain competitive advantage

Pascal Bornet:

When everyone in the company has the opportunity, is empowered to build their own automations, to improve their own work using automation, and their ownership will change. And again, they feel they’re part of the transformation, they are part of the dynamic that is happening. They are not left behind.

Pascal Bornet:

Thank you Alex. Thank you for inviting me.

Pascal Bornet:

Very happy to share my passion, before it’s Sebastian… that pushed me and my co-author to write this book, and I’m happy to share about it today.

Pascal Bornet:

Good and important question. We all hear a lot about artificial intelligence. It became a buzzword that would say over the last two years. And we use it for science fiction as well as companies, and people don’t understand how to pragmatically in reality use it, and get the efficiency gains that we are working for.

That’s the reason why we haven’t called it artificial intelligence, but intelligent automation, intelligent automation really refers to the use of different types of technologies to automate some of the work that is currently being done by people. The official definition is, it’s a combination of methods that includes people, organization, data, processes, and technologies. That includes robotic process automation, machine learning, natural language processing, computer vision, and much more.

Putting or combining those technologies and concepts together, we are able to automate end-to-end processes that are currently being done by people. For example, purchase to pay is an end-to-end process that each and every company around the world has to perform to basically identify vendors, procure goods or services. Then identify if we’ve received those services and goods in the right quantity, quality… And if we can then pay them, those vendors.

Pascal Bornet:

That’s an example of a process that is definitely currently being done by people. And that can be automated. All the processes can be automated. I haven’t seen a process that can’t be automated, not yet.

Pascal Bornet:

Yes, very good question. Let’s get back to our example of purchase to pay.

Pascal Bornet:

I think that’s universal, every company does it. We said that it’s the process to start with, you know that you need to buy the goods or service through a vendor. And the starting point is usually being done leveraging machine learning, to forecast what will be your needs. And especially in big companies you can think of manufacturing companies when they need raw materials, for example, that’s a critical process to get their manufacturing and production activities going.

Forecasting activities, that’s the number one process of this purchase to pay end-to-end process, is leveraging machine learning to predict what will be the needs of which product do we need to buy to which vendor, for which quantity for when? That’s the first starting point.

Pascal Bornet:

Exactly.

Pascal Bornet:

Exactly, exactly. The way I define a process is a sequence of tasks and those tasks can be cognitive tasks, like making a decision, like digesting information to predict or to make a decision, or can be manual tasks or can be tasks that are related to what we see or to what we hear or what we say.

Pascal Bornet:

Let’s go step by step.

Pascal Bornet:

Exactly. And to figure out the decision of buying, which product to who, when, how much. You take this decision, and this means that you need to generate a purchase order. And very often you will use a workflow platform that will help you to connect with the vendors and have them go through this process that goes from step one, you get the purchase order, step two, you get all the answers from the different vendors.

And step three, you will have a first review process with selection of top three vendors, for example and finally get to the selected vendor. Through this workflow platform that belongs to… In the book, we talk about four capabilities that are basically four groups of technologies that help to mimic what the knowledge workers are doing in their work.

The first one was using machine learning, it was thinking and learning. So that’s the first three. Now we are talking about a workflow platform that helps to route the data, collect approvals, store the data, and help and support the work flow of decisions between people. That’s the execution capability.

Then let’s say we select a vendor, and then we receive the goods, and now we need to pay them. So we receive their invoice. And we receive their invoice, unfortunately, not all companies have an exchange of data as you know. You might probably receive a paper invoice, but even you might receive it by mail or a PDF of the invoice. In this case, you will need first of all, computer vision to help you read this invoice automatically.

That’s the third capability which is about visions basically, that’s what we do with our eyes usually. And now machines are doing it automatically thanks to computer vision. And then you need to probably read the email and probably interact with the vendors if there is an issue on the invoice, for example, and here is coming language capability, which is currently being done by virtual assistants that are leveraging natural language processing.

And finally whenever you’re happy with the amount of the invoice, you got your goods in the right quantity, quality, you want now to perform the transaction, to pay those vendors. And this is a very transactional activity that can be simply automated using robotic process automation, for example. And this is the coming back to the execution capability.

Pascal Bornet:

No, no, that’s not-

Pascal Bornet:

Yeah, yeah, very good.

Pascal Bornet:

Those other capabilities you can use to automate a business process. But sometimes one, two or three of them are sufficient.

Pascal Bornet:

Correct.

Pascal Bornet:

Yeah, very importantly. Why those four capabilities? Those are the capabilities that we as knowledge workers are using to perform those processes. Either we read a document, either we talk with someone, either we press a button to send an email, or we think and learn using data and providing some insights and predictions.

Pascal Bornet:

Yes, if you want, we are kind of creating a digital twin of a knowledge worker. That’s another way of explaining it. We are creating a digital worker that is working hands-in-hands with the human worker.

Pascal Bornet:

That’s a critical question. And the answer is, think business first. Technology is just a tool, just something to help. What’s important is what we need. I’m currently performing this process and onboarding clients in the bank. When you want to get a bank account open and you need to produce a lot of documents. How can I get those documents?

How can I make this process of onboarding of clients more efficient, more pleasant for the customers, more pleasant for the employees that are working in the back office. It’s just about this. And to make it easier for the clients, you would think of making it fully digital, for example, so that it’s a workflow online that can be done from your sofa. You can upload your documents.

Everything is simple, if you have a question, you have a virtual assistant to help you. In the back office of the bank, usually you have employees receiving those documents, typing those documents, from image to a digital form. I mean, who wants to do that eight hours per day? I mean, I wouldn’t wish this for my worst enemy.

It’s about helping those people in the back office, having an optical character recognition tool, as you mentioned. Basically computer vision that helps to read the document, extracts the information that we need, and directly processes it into the transactional system that it needs, for example.

Pascal Bornet:

Correct.

Pascal Bornet:

Exactly.

Pascal Bornet:

That’s definitely a very important step, I would say it’s maybe the second one. The first one is to identify what the business needs. And where are they the strongest, because at the end of the day, everything can be automated, but some of the processes have a high value of being automated, a higher value than others.

And I mean, from my experience, if you have a look at the different processes, you can automate in a company and it can go from few hundreds, five hundreds to thousands. If you really go to identify all of them, you see that only 5% of them, roughly 5% of them will generate about 70% of the total value you can generate by automation.

It’s all about focusing on those ones first. And as you say, then those processes we don’t want to just copy what is being done today, because we know it’s not efficient, so let’s rethink them completely. And the best way to rethink them, redesign them, is to focus on the output, the outcome, what do we need at the end.

And the onboarding process that I just described, at the end, I mean, the bank needs to be satisfied with all the criteria that need to be checked for clients to manage regulation, but also to manage the risk of the bank. What do we need at the end? That’s what we need. So now how do we get all these in an automated way? It’s starting at the end, that’s the end, that’s the second step.

Pascal Bornet:

I think when you get there, meaning when you get to a redesign process and you are clear on what are the steps that are necessary to get to your outcome, it doesn’t need to be translated yet into what is the technology to help. But just that you need this, this, this, that. I think 70% of the work is done. That’s the most important.

Then it’s about bringing a solution architect or someone who knows, doesn’t need to be a very technical person, someone that knows that to achieve this, you’ll need these types of technology to achieve this type. It can be one solution architect, it can be also, and I prefer the solution to have around the table, different talents coming from teams that usually are functioning in an isolated manner in companies.

Most large companies have a team for robotic process automation. A team for customer service, which is dealing with chatbots. And another team that is about data science and machine learning. It’s really about bringing all those standards around the table and redesigning together and identifying which technology will we need when, where, and identifying how we connect those technologies together?

Pascal Bornet:

Correct.

Pascal Bornet:

And it’s a critical component of a successful transformation. And when we explain the different capabilities in the book, when we explain the framework to successfully achieve such a transformation, people come first, people are in the centre. I’m used to saying that, those transformations are made by people for people. If you miss the people aspect, you miss the project.

When we talk about people, we talk about two types of people. One is management, and management like CEOs, C-levels, I mean, their involvement, information, education, and sponsorship of the transformation are critical. I’ve never seen a company succeeding in this type of transformation without this. That’s the critical component.

It’s about bringing them to this level and helping them to build their vision, and to build their vision, we need to help them understand what are benefits, quantitative, qualitative, what can be a typical roadmap. And they will be able to understand the benefits, how much investment they’re already… That’s the first but critical component of people.

The second now, all the other employees, and they are extremely critical as well. Again, there is no transformation that happens without people. You need to inform them, educate them, that’s one of the reasons we wrote the book. That’s to create this tool that is necessary for this transformation to succeed.

When you’ve informed and educated everyone in the company, you need to go further; it’s not enough. You need to involve the people in the transformation. Today, and more and more the technologies are getting more user-friendly, you might have heard about low-code, no-code technologies. The philosophy behind that is that everyone in the company is a potential developer, citizen developer.

And I think it’s dimension is extremely important and needs to be leveraged by companies today, for two key reasons. When everyone in the company has the opportunity, is empowered to build their own automations, to improve their own work using automation, and that ownership will change, and again, they feel they are part of the transformation. They are part of the dynamic that is happening, they are not left behind them.

That’s the first thing. And the first thing is, of course, because you have more hands working towards implementing the transformation. Of course, you go faster and you can do it bigger. And the second and most important thing as a result of this is, because the ownership of each and every person in the company is changing and moving towards this transformation to succeed, the mindset, the culture of the company changes into more digitalization, more automation. People understand and are willing, and the mindset is changing, the mindset is so important, the culture is so important.

Pascal Bornet:

Definitely, that’s a great point. Empowering the people to do things is a good starting point, but we need to go further. And the next step is about incentivizing the people to do that. I’ve seen many ways to incentivize people doing this, from contests that give rewards to people, rewards to people in the shape of publisher, in the shape of MBA courses offered, in the shape of, I mean, it can be many things. But most importantly, it should be part of the KPIs of anyone.

Pascal Bornet:

There would be the key performance indicators of each and every employee.

Pascal Bornet:

That’s a difficult question because it’s like machine learning, it’s everywhere now. I think each and every tool now, each and every technology on the market now is making their configurations easier by having user-friendly screens, drag and drop type of functionalities that make it easier for everyone. Each and every technology is following these trends.

Pascal Bornet:

No, that’s a good point. That’s exactly why incentivization is very important, to push people to act in the direction that is needed for them and for the company. And when I talked about the action to the people that are necessary, I talked first about information and education. What does it include?

That includes not only what are those technologies and what you can expect from them, for you as an employee, but also for the company, but also being very transparent and what does it mean for each and every people’s role? And some rules, definitely, those people that we were talking about that are transcribing an image and typing into a transactional system, eight hour per day, those people definitely will have technology to do their work.

And as we said, nobody would love to do that at any time. It’s about giving transparency to those people, and telling them in this specific example, technology will help you now do the job that you were doing that was very transactional. Do you agree with that? The people will say yes. And now we’ve thought about refocusing your time on more value added activities. For example, why don’t we help you now to understand the data that you use to type and work on these data so that you can do some analytics, create some insights from it, and we will teach you how to do that.

Pascal Bornet:

This person will go from the stage of copying data, the whole day without even thinking to a role where this person we have to think, and will have to make some decisions, will have to bring some insights. And from a survey that psychologists have done, satisfaction and happiness at work happens when people have to take even small decisions but decisions, and create some insight, create something from the-

Pascal Bornet:

Psychology point of view, it works. It’s not that science proves it.

Pascal Bornet:

It’s a very good point. The first to start, will be the first ones to learn from it. And of course, will be the first one to collect the benefits from those technologies, or to benefits for companies-

Pascal Bornet:

It’s across the board, productivity improvement, let’s start with that. 20 to 60%, that’s-

Pascal Bornet:

What we’ve seen on the market. Meaning an improvement of net profits. It’s really that’s what’s driving the company’s health, on top of these customer experience improvements, and we’ve seen percentages of NPS scores. Satisfaction of customers increasing by 50% after such a transformation. Employee satisfaction, we’ve seen teams suffering from a high turnover of teams like 20, 25 to 30% of turnover because the work was very transactional, and people were not happy to do that.

To teams that are performing well with still being in a transactional environment, having 10% turnover, so really increasing employee satisfaction. And I think those are four companies, those are the critical ones. I mean, you have also the improvement of quality, you have also the improvement of quality that means better outcome of the work, but also less errors. And somewhere else can cost very much, especially when we talk about the finance department, for example, I mean, yeah, those are, those are the ones that come to my mind now, you cannot release all of them.

Pascal Bornet:

Exactly, and so-

Pascal Bornet:

Correct.

Pascal Bornet:

Exactly, you just forgot one thing, the increase of revenue as well, because a customer being more happy-

Pascal Bornet:

Exactly —

Pascal Bornet:

Yeah, it’s a starting point, but we don’t know when it ends, because I mean, I like a quote, which is, “Intelligent automation is a destination,” not as an ongoing project, not a destination. It’s not a one off project, you need, basically, one of the key recommendations we have and what the… Companies that have succeeded have been the centre of excellence that not only helps to drive the implementation of the roadmap of those concepts and technologies, but also that constantly watch what is under market without the new capabilities coming that we can use to generate more benefits. It’s a never ending game.

Pascal Bornet:

To answer your question, can it be applied to any function in the company and any company? The answer is yes it is. And we’ve demonstrated it in the book at the end of the book we have, I think it’s more than 100 of collecting more than 500 business use cases. And we’ve ordered them by function and by industry. Marketing is one of them, but you have human resources, supply chain, finance, you name it.

In terms of industries, we have banking, telecommunication, governments, and yeah. We’ve demonstrated it through this. And to answer your question for marketing, for example, we have anything that is related to, first of all, understanding the data and creating insights from this data. Let’s think of what you can forecast, what you can plan. Forecast of sales, forecast of understanding of a market, segmentation of a market, segmentation of competitors, of demands, and monitoring of marketing actions, prediction of the impact of marketing actions.

Pascal Bornet:

No, for creativity, for critical thinking, for relationship with people, there are a few capacities that technology is not yet able to achieve.

Pascal Bornet:

And I think it’s going to stay like this for a while like this, and I don’t see technology being able to perform such high level thinking activities yet. It’s about thinking out of the blue, critical thinking strategy, design, pure creativity, and relationship with the people. I mean, for the time being, we have virtual assistants, but they still lack a lot of-

Pascal Bornet:

Of components. Yeah. I mean.

Pascal Bornet:

Empathy for example-

Pascal Bornet:

Is an important one, exactly, exactly.

Pascal Bornet:

Exactly, we did some research around this that we present in the book. Out of all the scope of work tasks that are being done by people in average in companies, more than 30% of those tasks can be eliminated.

Pascal Bornet:

They think of unproductive meetings, unproductive emails, we’re spending a huge amount of time on this. Second, more than 30% of those tasks can be automated, and there is more logistic transactional stuff that we are doing in our days. And the remaining 30% should be augmented by technology. Again, we should keep them, they’re creative tasks, these are strategic components, these are decisions.

For the time being we can’t automate, again, I don’t believe we’ll be able to automate them for a long time in the future. And where we need the clinician to help us to go beyond. Think of when we talk about augmentation, think of a doctor, a medical doctor who usually spends hours reading x-rays to understand if there is a tumour on it. Thanks to technology, these doctors are augmented, is capable of making decisions in just a few seconds, which have huge impacts on people’s lives.

Pascal Bornet:

Can I automate the end-to-end process? But some of those parts in the process would remain manual.

Pascal Bornet:

And the issue with that is, when scaling the process from one decision to 100s of decisions, then you still have this bottleneck of, ah, that’s done by people, that’s not scalable, or you need to hire, you need to hire. And then when your business is slowing down, oh, you need to fire, you need to fire, and then you need to hire the 30 minutes. It’s very important to think holistically, have the big picture first.

Pascal Bornet:

Yes, and don’t keep a task manual in the middle of an end-to-end automated process. That becomes your bottleneck, your weakest point.

Pascal Bornet:

I think the pandemic has been, of course, a terrible event, but if we have a look at the positive aspects, it has pushed the companies to digitize themselves, to automate themselves. The companies that have not been able to collect their cash online, to sell their products online, to motivate their employees remotely, either they are still surviving because of subsidies from the government or they just died.

And so back to your point earlier, it used to be a competitive advantage of using such concepts. Now it’s a survivor, it’s an inductive for the company. I see the future being stronger and stronger and with more and more use of those concepts and technologies, and those concepts and technology sophisticating in time, so that they allow us to scale those transformations.

Today, this is what we are lacking. According to Deloittes, the good news is that more than 50%, 5-0, 50%, half of the companies around the world have started their journey in one way or another in intelligence determination. But the bad news is according to making only 15%, so 1-5 have been able to scale it. Basically have been able to implement it across more than three divisions or functions. The holy grail is about scale now.

Pascal Bornet:

Yes, yes, yes. It’s coming from a that the predictions are that, I mean, basically the life of companies, the duration of company is getting shorter and shorter. More and more companies die every year, but also more and more companies born every year. And the predictions are that’s more than half of the S&P 500 companies will change over the next 10 years.

The question is, are you part of it or not? And to try to answer this question, are you part of it or not? We’ve analysed data from a large sample of companies to understand what makes them different in their adoption of intelligent automation. One of the key criteria, the key indicator we used is the benefits created by employees.

That really shows you how much you can produce as benefits by human resources. The higher, of course, the better, the higher meaning, the better you use your home your human resources, and certainly the more you’re using automation. And we found out that the key driver that explains this is the age of the company. The older the company is, the less it’s going to use automation, and the more likely this company is to die soon.

Pascal Bornet:

And to explain this, our finding is that, all the companies have too much history, too much legacy. It takes too much effort, it’s a huge amount of effort to transform themselves. And we are talking about significant changes that impact everything in an organization, processes, people, data. The analogy to bring all these from where they are today, to where they should be tomorrow is huge compared to starting a new business. Brand new in, I mean, leveraging, straight from scratch those new concepts, leveraging those new technologies.

I mean, we see more and more companies, all those companies changing themselves, not by trying to change the legacy business, but by trying to create a whole new business close to the legacy one, see if it works, make it small, like a startup that is close to the, let’s take a bank, an old bank, old retail banking agencies and so on.

This company creates a new startup and the startups creates the business online, completely online and brings more intelligence and automate and digitalize. And then we see two behaviours, either very often these new businesses will work very well. Either the company uses this new business to say to the legacy business guys, look at this, this is what we need to achieve, so let’s get inspiration from it.

It’s easier for the people because it’s part of the same company, it’s part of their colleagues to infuse this new way of thinking and the new. That’s one. And the other behaviour is to kill the legacy slowly, kill the legacy and slowly grow the new business.

Pascal Bornet:

Alex. This was an honor, thanks a lot for inviting me, and a great discussion, really like your challenging way of questioning.

Pascal Bornet:

Very fancy, very fancy, I like it.

Pascal Bornet:

Thank you Alex, thanks a lot.