It’s not expensive to buy a spy, according to a recent article. You can ‘buy’ a spy for $10,000 a year, or in more significant cases, you may need to pay $40,000 to $70,000, especially if the spy takes a considerable risk. There are other motives for people wanting to sell or give information, not just for military and international politics secrets. Human beings are a significant security risk for businesses. Can we do something to improve this weakest link?

Colonel Vladimir Vetrov was one of the most important spies during the cold war. He worked for the KGB and leaked more than 3,000 pages of documents to French intelligence, including the names of more than 400 Soviet operating agents. He operated from 1981-82, and it is said his information went direct to President Reagan. He played an essential role in exposing weaknesses in the Soviet Union, its dependence on stealing western technology and how an accelerating arms race was driving it to collapse.

However, it seems Colonel Vetrov didn’t do this for the money. He received some small gifts that he gave to his mistress, but nothing significant. He was more embittered with his career development at the KGB and also frustrated by the Soviet system. Several studies and cases demonstrate that embitterment is often a more important motive for spies to leak information than simple greed.

Edward Snowden leaked highly classified information. His motivation was not clear, and now that he is now in Russia, he has indicated he was unhappy that the US authorities spied on its own people. Wikileaks also received leaked information from other people working in governmental agencies. 

Governments and enterprises spend a lot of money developing better solutions for physical and cybersecurity that are becoming increasingly significant. And these investments are definitely needed. But at the same time, it is important to remember; it’s people that leak information and create holes in even the most sophisticated systems.

I have personally seen cases of spying or information leaking during my career. Once, a person at a customer leaked information from our competitors and how some people in the organization worked with the other vendors because he was not happy about his position. In another case, a company warned us that a cleaner in our project office had collected documents and photos from our bid documents. In one extreme case, someone set off a fire alarm in an office, and several laptops of a new project team went missing. All these are old cases.

The question is, who can you trust? It is not an easy question to answer, and it is not black and white. Even the most loyal person can change and start to leak information. We could also say that no one is totally reliable; most people reveal information at some point, either intentionally or unintentionally.

One solution is to keep people loyal. A good salary helps, but even more important is to make people feel they are being treated fairly. Companies try to identify problems to keep their employees loyal and reliable, but it is rarely enough.

That raises the question as to what information is relevant. Many companies hide information that is not very relevant to anyone, competitors or customers. And those parties can usually get that information quite easily, so it is not a good investment to try to hide it at a high cost. It can also increase the risks of leaks if employees feel that irrelevant information is being classified as secret.

There are technology solutions to avoid, identify and reveal spying and information leaks. For example, one old method is to make each copy of the information (e.g. a document) unique in order to identify whose document was leaked. It is also important to track who has copied some confidential information or had access to a system. There are other solutions, e.g. identifying unusual behavior, setting test traps or monitoring communications.

It doesn’t make any sense for companies to take similar measures as critical governmental agencies if it creates ‘bad spirit’ in the organization. One big risk area nowadays is employees using their own devices and personal communication tools. Several simple solutions make sense.

Suppose sensitive discussions between business partners preparing a bid, between a company and its law firm, or amongst board members take place via a messaging app, a Facebook group or another similar service. In that case, it increases the risk of inadvertently sharing information with other parties. Sometimes it can happen accidentally, especially when people are handling multiple groups and discussions simultaneously. It is not realistic in many of these cases to force people to use higher security tools which can be challenging to enforce between organizations. Most security tools have been designed for use within an organization.

Technology is not the only solution to stop people from leaking confidential information. But technology can help to avoid accidental sharing, easy leaking and identify the sources of leaks. These solutions must be easy to use, and they must work with commercial off-the-shelf (COTS) technologies and services. They can help keep information in closed groups, prevent direct sharing, and identify if someone has shared confidential information.
​
Security and trust in people is not black or white, more like shades of grey. There will always be people who want to spy and leak information, whatever it takes. But for the majority, it probably helps to have clear rules, better tools and increase the risk of getting caught. Any company that invests in building security in its physical and cyber environments must also think about building and monitoring trust with its people.

We are all probably skeptical about people who tell us what we should do because they think it’s what is best for us. A good example is adults telling kids and teenagers what to do and not to do to protect them. Apple and Google are doing something similar with privacy. They want to be consumers’ parent to protect their privacy, but they want to keep control. Do consumers really want this, or would they like to control what to do with and where to use their own data? Here lies an opportunity for a new data business.

Apple is introducing new models in the latest iOS versions for users to control the trackers of mobile apps. Basically, a user must allow apps to follow them around the web, collect data and target other apps. Not surprisingly, there are estimates that around 70% of people, if asked, would not allow tracking of this type so that Apple may be onto something. However, this also increases Apple’s control of the ecosystem and makes it even more a closed-garden system by giving Apple control over what app vendors can do and how they do it.

This would have an impact on other companies like Facebook and Tencent, which operate online advertising. Facebook has already warned, this would affect its revenue. Tencent and other Chinese mobile internet companies have developed workarounds for the model.

Google’s Chrome will shortly stop supporting third-party cookies, making it harder to track users on the web. Simultaneously, Google is preparing new solutions to track the browsing history and profile and segment users, enabling advertisers to target ads better. This is coming from Google’s Privacy Sandbox project and gives Google a more critical role in the advertising ecosystem, making it harder for smaller ad companies and advertisers to work independently.

Privacy and user tracking resemble something from the ‘wild west’. It becomes more complex when a few companies can control a significant part of the internet and mobile ecosystems. This may be specifically about web tracking and ad targeting, but Apple’s Health app collects data from wearable apps and enables downloading of health records, and Google Fit aims to do the same.

All this opens the opportunity for a new unholy alliance between consumers and enterprises. Consumers could share their profiles direct with businesses and bypass the internet companies if they could see concrete benefits. This is not a new idea, but it needs easy solutions to become a reality. It is unlikely consumers will do something just for better privacy; they will want to see those benefits quickly.

Let’s take a few examples of what this business and consumer cooperation could mean:

1. Consumers share their wearable data or profiles with insurance companies to get insurance products that are better tailored and priced specifically for them.
2. Consumers share their wearable data with health and fitness operators to get better and more tailored services.
3. Consumers share their personal browsing history to get personalized content and ads.
4. Consumers share their music listening and movie watching history to find more exciting music and films.

These are a few examples of how users can have a direct data relationship without the internet and mobile giants trying to control it. But consumers will need tools to collect their data and share profiles (not raw data). It can’t be something each individual negotiates with enterprises who would dominate, and consumers wouldn’t know the right price to demand. Consumers need weapons (i.e. tools and models) to do this properly. Ideally, this would be an open ecosystem with open source tools, open APIs in an open environment where different parties and developers could provide the means for consumers to keep their data.

All this opens the door to new technology and companies to offer solutions for consumers and enterprises. Could this be the most significant change in the data business since the early days of the internet? Regulators could also accelerate this development by introducing new privacy rules, giving more power to consumers to control their data and restricting the internet giants’ dominating market position.

Current privacy and data discussions and developments can confusing. Even though parties exist that want to protect consumers, they often add restrictions that make their lives more complex, particularly if they continuously need to click approvals. At the same time, data analytics offer more opportunities to consumers and businesses alike to better utilize data for better services, better prices, and make lives easier. The motives of some ‘protectors’ are not very clear and maybe not as ‘innocent’ or as ‘honorable’ as they might appear. There also lies the possibility of ‘data dominance’ simply moving from one actor to another.

Long term solutions for data and privacy cannot be based on the controls and restrictions of a few big companies. Consumers must be able to control and utilize their data. All kinds of companies must also be able to use data if they can offer value to consumers. Otherwise, not only advertising but many other areas, including health and finance services, could also end up in the control of the internet giants.

The article was first published on Disruptive.Asia.

Data is the basis for many operations, but it doesn’t mean data is always reliable. Things can get complicated when you don’t know which data source is reliable and which is not. But we must use data all the time. Sometimes it is possible to increase the accuracy, but the more meaningful solution is to build a software layer to correct data before using it.

I earlier wrote about known and unknown things and data points. The reality is even more complex. We know some data is relevant, and it is available, but we don’t always know how reliable it is. We all know about opinion polls and their error margins. It is just one example, but uncertainty is linked to all data sources and models that utilize data.

In aeroplanes or nuclear power stations, the core systems do not necessarily trust individual sensors or data sources. There can be many reasons why a particular sensor gives incorrect data. For example, a pitot tube that measures an aircraft’s airspeed can transmit incorrect information if frozen, which has caused several plane crashes. Today, a plane typically has several pitot tubes, and the software tries to draw conclusions and give pilots warnings if one or more give inconsistent readings.

Sometimes the situation is more demanding when it is difficult, even impossible, to know if data sources and sensors give accurate data and how large the error margin is. Examples of this are wearable devices. They can measure your exercise patterns, sleep, and body functions like heart rate, temperature or blood pressure. These devices are calibrated using higher accuracy devices during development. But it is still hard to say how accurate they are for different people in different situations. For example, even with top-level research instruments, it is not easy to measure how much light sleep, REM, and deep sleep a person has at night.

We might also have a situation where we have many sensors, but some data might be missing. It is a complex task to combine data from different sources, and it is also tricky to know if available data makes any sense combined. This can occur when having many IoT sensors or an organization’s internal data from multiple sources to measure processes or even financials.

It is often said that intelligence makes up only 20% of AI implementations, and the rest is getting data, combining it and correcting errors. This layer is often underestimated. I have seen projects where 95% of the data is inaccurate, incorrect, or missing data points.

There are several ways to increase the accuracy of data, for example:

1. If we get the same data from several sources, we can have a ‘voting’ model to determine the ‘correct’ data from most sources. The pitot tube system often works like this.
2. We can learn from different sources’ accuracy and take their error margins into account to correct data. Opinion poll models often have correction factors.
3. More complex solutions combine several data sources and make conclusions about what the data can indicate and how well data fits this. For example, if a person is running, this can be concluded from a combination of several data points from wearable devices, e.g. motion sensors, speed, heart rate and oxygen. Another example is that the software in a phone camera system tries to make a photo better based on each camera’s pixel data by correcting individual pixels alone and how the pixels fit together.
4. Sometimes it is possible to have a feedback loop to know how accurate some data is and then have machine learning type models to develop correction factors and models to use data.

These layers combine, correct and smartly use data and become more important as we get more data sources. One could even say it is pretty simple to create AI models if someone has developed this layer to make reliable data available. It is often said that IoT business is not really to sell sensor hardware but to manage data, but what is ignored many times is the critical question of getting reliable data.

It is not easy to make these layers that combine data because each source is different, and it can also require an understanding of the data to be able to analyze and integrate data sources. It is possible to make general models and tools for this, but they often need tailoring for the different data sources and combinations of data sources.

With AI’s hands, these smart data combining models and layers become a vital part of the data and AI business. Data is valuable only if it is reliable. We can trust AI only if it can use correct data. The reality is that no data source is 100% reliable, so we need intelligence, how to correctly and optimally use data sources.

The article was originally published on Disruptive.Asia. 

It has become popular to manage operations with data. New tools to collect and analyze data are continually appearing. But things can still go badly wrong with data. Dashboards and analytics apps rely on multiple assumptions, and if external factors change, models don’t work anymore. COVID-19 activities in many countries have been good examples of this. When you don’t know all the elements, individual numbers can be misleading. You can never manage only by the numbers you have; you must have insight, understand the environment, and be ready to look for changes outside the numbers.

Some years back, US Defense Secretary Donald Rumsfeld made the famous statement about known unknowns and unknown unknowns. Many people laughed at that statement, but it is quite an excellent way to describe reality, not only in war and foreign politics but also in business environments. Many companies focus on things they know, but they are not prepared to handle external factors they are unaware of.

Many countries specified data-based recommendations and rules during the current pandemic, e.g. tier systems to close shops, services, restaurants, or varied travel restrictions. They are based on the number of cases per 100,000 inhabitants, people in hospitals, or the virus’s R-value (reproduction rate). But most governments have been forced to change those rules and thresholds many times. It has given reason for citizens and opposition parties to criticize their actions. The reality is that it is challenging and not very intelligent to manage only by a set of numbers if many factors remain unknown. This scenario makes sense to change rules and metrics as we learn more about the situation.

Many companies focus on optimizing their operations based on the numbers they follow and measure. It is precisely why a disruption in an industry lands an incumbent company in trouble when they focus on numbers in the domain they know and recognize. Still, disruption often changes factors that they don’t monitor or are unknown to them.

The most famous examples involved mainframe computer companies when personal computers came out; for Nokia, when Apple introduced the iPhone and print publishers when online content came. Those companies focused on optimizing their operations, products and metrics in the existing business and environment. For example, Nokia was optimizing the production costs of phones, model ranges for different customer segments and their software features. The iPhone looked too expensive based on their metrics, not suitable for many customer segments and had too few features for users. 

Known unknowns and unknown unknowns are relevant categories for businesses to analyze in more detail. Known unknowns are factors they know about but cannot get details or data on. For example, competitor’s future products, economic growth and the future availability of components.

Unknown unknowns include factors that we don’t know at all but that are likely to impact us. The pandemic, for example, came as a surprise, and we had no idea what impact it would have on our lives and businesses. There are many factors we don’t know about or can even imagine, but they might have a lot of impact on us.  

We can also think of one more category, unknown knowns. It would mean things we know, but we don’t recognize how they impact us. For example, using available data but never thinking it will be relevant, like a company being aware of climate change data but not recognizing it as a factor in their business.

However, if we only focus on the ‘known knowns’, we can still be surprised when something changes and still not understand the real reasons for it. Many businesses and people focus only on the ‘known knowns’ and try to understand and explain everything based on that, but then they miss those three other areas discussed above, and external factors surprise them, or they reach wrong conclusions if they don’t understand that things outside their focus have an impact.

Can we do something to handle unknown areas better? Maybe the most important thing is not to think you know and understand everything, and that you must keep your eyes open for other things too. We can assume at least three categories of actions that you can do better with data and metrics:

1. Don’t focus only on data that explains your current business but try to have a vision and insight to understand how the future could look, even if it changes your business fundamentally.
2. Don’t think your data and models are perfect. They might be useful to measure and optimize your current activities, but there are other factors too, and if external factors change, your existing model might stop working.
3. Prepare for unexpected events with scenarios. You can’t know all things that might happen, but you can make scenarios to act fast when things don’t go as you expect.

The COVID-19 pandemic and its impact on governments, businesses and individuals has taught us how companies can and need to be better prepared for unexpected events. Some of them can be big significant global events, some smaller ones like why some sales go down and why people are no longer interested in a particular product.
Maybe the most important things to remember are: 1) you cannot know it all; 2) your models are not perfect, and 3) when something unexpected happens, don’t think you can explain or handle it using old models only. As the famously coined Darwin quote goes, “the species that survives is the one that is able best to adapt and adjust to the changing environment in which it finds itself.”

UK’s supreme court recently made a ruling classifying Uber’s drivers as workers entitled to rights such as minimum wage and holiday pay. Uber has a long history of legal battles. It has had many fights against taxi regulations, who can offer rides and how. But the struggle over the rights of its drivers is even more fundamental. It hits at the heart of its operating model and cost structure and it’s a good example of disruption versus regulation.

The court ruling says that drivers not only work when they are on a trip but includes the time they are logged into Uber’s app. Amongst other things, Uber now needs to start a pension scheme for these drivers. It significantly increases Uber’s costs, and it fundamentally changes the idea that drivers are ‘entrepreneurs’ who get customers through Uber’s app. Californians voted on a similar question for their workers’ law in November. Uber spent a lot of money to get support for its model, and the results were that the drivers could continue as independent contractors in that state.

Employee rights is only one area where new business models and disruptive startups encounter issues with old regulations. Digital services could and should be global, but financial services and fintech are examples where regulation significantly restricts how and to whom services can be offered. In fintech, we see many restrictions, including what services are not allowed to users. Regulations are supposed to protect citizens, but they also safeguard companies using old models to continue business. Is that fair on people who ask for freedom, not protection?

Regulation is one of the main reasons why fintech services are slow to acquire market share. It doesn’t just limit how services are offered; it also makes it more expensive to provide services. In the global sense, it is hard to understand why you can only use services in your own country. Why is it that you can travel to another country to get financial advisory services and make investments, but you cannot, in many cases, use those services online or talk on Zoom with your advisor in another country? 

There are many other examples of how startups and new business models collide with old regulations. And it is not always the regulations. For example, labor unions or incumbent companies push to introduce new laws and regulations to protect their position. Taxi companies, taxi driver unions and banks are famous for utilizing laws and their lobbying power against newcomers. None of them have a reputation as model citizens or focused on offering their customers the best service.

In most cases, the arguments against newcomers are justified with good intentions such as protecting customers, employees and ensuring fair competition. It is never easy to say what is right and wrong, and the best way to protect someone. Still, it would be more honest to say that in most of these cases, the question is not really about the protection of customers, employees and competition, but about the fight between old and new models.

Many people want to drive for Uber, and similar services, as independent contractors and have their freedom to do other things, too. Then some people like to have more permanent employment and get paid holidays. Many people would like to use new fintech services and global financial services, and then some people just want to walk to their local bank branch and send checks by mail.

As a result, societies become fragmented. It is tough to have one model fit all, but regulation forces one model that everyone is made to follow. That’s OK and easy to understand if the objective is to protect all people. But if it concerns people who do not want that protection and it causes no harm to other people, it is harder to justify. Of course, there are always arguments about indirect impact, e.g. how the competitive environment is shaped.

Let’s be honest; many of these questions are political. They are about conservatism versus the freedom of individuals and businesses. Some of them are also about negative and positive freedom models, i.e. whether a system allows something and offers equal opportunities to different parties. Anyway, a kind of reality in business is that the most efficient model will win eventually, assuming lawmakers don’t restrict people’s freedom by limiting the choice of services they are allowed to use.

The article first appeared on Disruptive.Asia.

People are living and working more and more in digital environments. COVID-19 has accelerated the transition to more virtual and digital interactions. Security is a concern in many services. But part of the problem is that security experts, companies addressing customer concerns and even governments focus on negative messages and want to offer restrictions and hard to use tools instead of focusing on opportunities and making the internet a more trusted environment. The thinking is often too technical and theoretical, not based on human behavior or user experience.

Trust is a fundamental basis for societies and businesses. Countries where people trust each other typically work better than countries with shallow trust. It is hard to make a country or city safer just by adding more police officers or restrictions. If business parties cannot trust each other, they just try to focus on short term quick wins and don’t want to create long term commitments and investments. 

We have the same situation in the digital environment, but many parties still believe that added restrictions, more policing tools, and trendy, trustless transaction solutions would make it better. We can see this on many levels. In many companies, security officers and experts tell us what must not be done, how risky everything is and creating all kinds of rules for the organization. Governments also sometimes adopt very simplified models to use. Some countries even restrict what people can see and do on the internet. But even the US and UK want to move to more populist models like forbidding end-to-end encryption in the fight against terrorism or protecting children. Of course, it is a totally unrealistic request and doesn’t do much to make the internet a safer or better place.

We all know how complex it can be using digital banking apps, identification and signing services. These are usually built from a very technical perspective, making something technically bullet-proof. Still, they are not lazy-user-proof when users don’t use the service or forget the security recommendations while using the service. 

The Financial Times organized its annual European Financial Forum in early February, and one crucial topic was digital finance services. Several speakers emphasized digital trust as a critical component for developing digital services. Nowadays, many things are done online, with email and messaging services, video calls and digital signatures. If parties cannot trust each other, it is quite impossible to conduct digital business.

Facebook deletes billions of fake profiles annually, we all get loads of suspicious emails daily, and companies create bots and fake profiles on LinkedIn just to generate contacts to sell more. Companies use solutions to secure communications and information sharing internally. Still, more and more business is being done across organizations, and most often, email, Zoom and WhatsApp are the typical tools, simply because they are the easiest to use. 

It is quite evident that better trust solutions are needed. But they should be built on natural human behavior and somehow generate trust built up over generations in societies and communities. Cryptography experts cannot create digital trust.

Typically, trust is built up step by step with human interaction. You may be in the same class in school, study together at a university, work together, or live in the same neighbourhood or have the same hobbies. Or you know someone you trust, and they introduce you to someone else, and you immediately trust them by inference. Trust is not black and white. You build it over time, it depends on the context, and you can lose trust quickly. And trust is not based on a set of rules and restrictions; it is based primarily on positive experiences with someone.

We are stepping into a new era of digital trust. Then pandemic has accelerated the need to do this. We need new solutions to build and manage digital trust, and they will need to include both social and technical innovations. And they will also need to work with our daily digital tools, like email, chat, video calls, and data sharing. As trust in society is based on positive experiences and opportunities, we need digital trust tools based on positive experiences, mutual learning and finding more opportunities.

The article first appeared on Disruptive.Asia.

Companies have been collecting data for years. Useful data can offer competitive advantages and be the basis for many services and better customer experience. There have also been many companies that have wanted to become data aggregators, collecting and selling data. But the big data success stories are not in selling data. Sometimes data is almost a toxic asset. What can we learn from the ways that data has been best utilized and monetized? We now have the same question with personal data, and many parties want to repeat the same old mistakes.

Fifteen years ago, in one of my earlier startups, we developed a marketing slogan: Data – the black gold of the 21st century. It was and is still a relevant comparison, but to make money from data is very different from the oil business. There you have separate business lines to drill and refine oil and then sell refined products. We can see something similar in the data business, but making big money in the value chain is very different in the oil and data business.

Google, Facebook and Amazon are the superpowers of the data market. They primarily collect and then build services that utilize data. They might buy some third-party data, but it is not their primary way to get data, and they don’t actually sell data. The reputation of companies that focus on trading data is nowadays quite shaky. As a person who runs data operations for a Silicon Valley giant once said to me, they are more and more skeptical about buying data when they don’t know its sources, how accurate it is, how those companies that are selling it got hold of it and how they conduct their businesses.

Don’t get me wrong, some companies make significant revenue by selling data, and some companies spend hundreds of millions buying data. But it hasn’t been an area to build unicorns and companies that shape the world as was expected maybe 10 or 15 years ago. Then there were a lot of expectations for data exchanges and other creative data trading business models.

Today data is traded more like a commodity than a unique source of value add. Companies buy outside data to enrich their data and help their solutions to utilize data better. The real value is achieved when companies build solutions to use data in marketing, sales and operations. One could even claim, the winner doesn’t have the most data, but the best tools to utilize the data. Of course, the Internet giants have heaps of data. Still, banks, telecom carriers and retailers have lots too (and the opportunity to collect more), but they have generally been slow to utilize it. Those successful companies also offer the data’s value to their users, like Google search, maps and other services, and Amazon’s better customer experience.

We are now seeing early days of personal data, i.e. how people can utilize their own data. Some initiatives and companies want to build solutions based on ideological views; people have moral rights to own and control their data. Those haven’t done too well; only a small set of people are interested in these ideological projects. 

Then there are those companies that want to help people collect their data and sell it. This has many practical challenges, including how to get a data market to work with enough demand and supply. Pricing is also a complex challenge, as are the associated terms and conditions, whether you sell your data for one purpose and how to track its use. It is not easy to get this personal data market working correctly. The user value promise is often disappointing, like being paid a few dollars monthly to watch ads. 

The most obvious option that has worked with the big data businesses for over ten years is forgotten. Why not offer people better tools to collect and utilize their data. When some companies want to help people to control and use their data by selling it, it is similar to recommending Google, Amazon and Facebook to sell all data they collect. Those companies have achieved their current position and power by having top tools to utilize the data they get. It is the same with individuals. If you want to empower them with their data, you need to offer the best tools to utilize that data personally.

Utilizing personal data will include many concepts, and we don’t know them all yet. We need an open market to innovate and develop those tools. But it can have, for example, tools to plan better personal finance, find the best prices, manage better health and wellbeing, and get help in all kinds of daily needs and activities. The longer-term vision is to build personal AI that offers a dashboard to guide all daily activities.

As with data businesses, personal data could also be enriched with external data sources. For example, public data like price comparison, traffic, public health and map data combined with personal data making it more powerful. Data model training for Machine Learning and AI improves when it can use data from many users. 

In many ways, the best way to utilize personal data is similar to what the leading data companies have done for years. But it seems that with a new business opportunity, many parties first go to very complex models, like justifying data with ideological thoughts or wanting to build a blockchain-based data exchange with digital rights management systems. Often the simplest and best solution is to copy one that has worked earlier elsewhere.

The article first appeared on Disruptive.Asia.

Artificial Intelligence (AI) is popping up everywhere, at least in discussions. Intelligent systems are being used in many places, and they are becoming smarter. But the real bottleneck is not the intelligence or ‘brains’ of the systems; it’s that AI also needs ‘hands’ to do things.

AI has become a very popular keyword over the last five years. Most company management groups and boards want to see some AI development in their organizations. Unfortunately, the reality, and actual use cases and expectations are not always in line. The biggest problem is not having smart enough machine learning (ML) or AI models to analyze data, handle tasks and make decisions.

Let’s take a simplified AI task. A system collects data, analyzes the data, makes needed conclusions and decisions and sends the results for operative use. If a whole system is built to work around AI, like a self-driving car, the capability to analyze the data and make decisions can be the bottleneck. But most systems are different.

We can take another example utilizing AI – automating insurance claim processing. We have the same phases, but data and interactions with other systems are much more complex:

1. A policyholder fills out a claim, probably a web form, but in some cases, it can still be a paper form. They also have some other documents, e.g. receipts, a report of an offence or a medical report. To get it all to a digital format, e.g. OCR (Optical Character Recognition) and NLP (Natural Language Processing), might be needed.
2. The insurance company collects data from other sources. For example, they can use a person’s insurance history from a national database, credit rating data, criminal records, data from other similar incidents. All kinds of data that can be used to see that information in the claim makes sense, is in line with other data sources, within a statistical margin of expected behavior and is not fraudulent.
3. Then the system analyzes the data and makes a decision. The decision can be to pay a certain sum, not pay, or send the case onward for further investigation.
4. When the decision has been reached the system must then send a letter or email to the policyholder, store the decision and all documents, start the payment process, and inform third parties (for example, national insurance database, health care provider, other parties in the incident, police). 
5. After this, the policyholder might not be happy with the decision and can trigger a new process.

In this example, we can see that the data analytics and decision-making is a small part of the overall process flow. There are many other parts, especially getting data from several sources, formatting the data, entering decision data to other systems and triggering actions in different systems. And what makes this even more complex is that typically the data is in many different formats and a part of the information is missing or is inaccurate (just think the claim form the policyholder fills and add attachments). Even the case of a data value being “null” needs to be handled, “null” is not “zero” and depending on the data set, it can have meaning or not. There are many handlers needed.

One of my companies implemented this kind of system several years ago. Although it was quite a digitally advanced insurance company and environment (Scandinavia), there was still a lot of work to be done. A typical rule of thumb in the data business is that 60% to 80% of the work is to pre-process the data. This is reality when you try to implement AI in any enterprise with many existing systems, and some of them can be quite old-fashioned. Just think SAP, Netsuite and links to banking systems.

We can even think of a more modern solution to get data from several wearable devices (Apple Watch, Fitbit, Withings, Garmin, Oura, etc.) to one place and bring it into a format that you could build ML/AI solutions on top. Even to collect all that data is not as simple as you would think, even when people talk about open APIs. APIs are still not so common, and while an API will be structured, the quality of data included can vary from one source to another.

A term I have started to like is ‘AI hands’. It means solutions, how to get data collected from many old and new systems, format them in one place and then get the processing results to operative use in other systems. Companies often forget or ignore the development of ‘hands’ when it is fancier to talk about the latest innovations for the ‘brains’. As always, great thinking is rarely enough; you must collect and organize information first and then get things done based on your thoughts.

In reality, these ‘hands’ are like software robots (RPA) that can work with different systems and devices. These include additional software components (e.g. OCR, NLP, data cleaning, APIs) to get the data and trigger actions (e.g. sending emails, start payment, start delivery).  Other useful tools are webhooks that can trigger background tasks, for example, in the serverless environment and such as verifying data and running NLP. This means the capability to work with a vast number of different systems and formats. 

Open source is often the best way to support many kinds of needs from small and rare systems to major systems. There are many data formats and even unformatted data that no company can implement in their proprietary system. Here, open source is the only option. These ‘hands’ and ‘brains’ should be based on commonly used and widely available programming languages (e.g. Python) that help get ‘brains’ and ‘hands’ work together utilizing open source components.

To get more AI and ML use, we need more and better ‘hands’ for AI. Management groups must also invest in these capabilities if they want to implement and utilize AI. And it is the same with consumer services, someone must offer the solutions where the data is available in a usable format, and there are tools to get results in real use. In last year’s Gartner Hype Cycle, many AI solutions were on the hype peak. AI ‘hands’ are needed to improve productivity.

The article first appeared on Disruptive Asia.

Automation and digitization should increase the productivity of work. But productivity growth has been flat or declining in most developed countries during the past 20 years. This has been visible in countries where most jobs, and especially new jobs, are not in manufacturing, but in services and information work. So, it would be fair to assume that technology and digitization don’t help improve productivity. Henry Ford, Jeff Bezos and Larry Page didn’t win big because they optimized old operations; it’s because they created totally new operating models. Opportunity lies in developing new ways to do things, not optimizing old ones.

World-famous economists, like Daron Acemoglu, Greg Mankiw and advisors of many governments, try to understand reasons for slower productivity growth. I won’t attempt to understand all the macro-economic factors, but to focus on small practical questions like what could be the bottlenecks with digitization and automation of information work.

I wrote earlier about how we need real digitalization, not consulting projects. The problem of many automation and digitization projects is that they just try to optimize the existing processes and implement them in legacy IT systems. Both those processes and systems were developed before the current opportunities of digital services were readily available. The optimal model would be to build new processes with the latest technology focusing on the company’s real value to its customers. If you automate old processes that are unnecessary to offer customers value, it doesn’t improve productivity. That’s why genuinely digital companies like Amazon, Facebook, Google, Netflix, Alibaba and many startups win business from old companies.

It takes quite a lot of courage from management and investors to disrupt old models instead of just trying to ‘optimize’ them. The reality is that to fine-tune old models with old IT could give you a small percentage improvement in productivity, but if you want to achieve much more, maybe 100 or 1,000 per cent gain, you must create new models to operate with the latest technology.

I also wrote earlier about the trending low-code and citizen-development, and how it can rarely help implement robust well-planned solutions. This is another example, why automation of processes doesn’t always bring significant value when citizen-development is trending in automation. Suppose a company must create new models to operate so that customers can communicate digitally with it, and they digitize all internal and supplier interactions. In that case, it doesn’t work if each employee (i.e. citizen-developer) starts to automate their routines from the pre-digital era.

It’s a sad fact that real automation also makes some work unnecessary. If you just let employees automate something they don’t like, it doesn’t make a company significantly more effective. Of course, by getting rid of boring routines, each individual and department can become more effective. But in reality, significant changes need much more fundamental changes. A record shop doesn’t become a new Spotify simply because employees automate some of their routine work. And a bricks-and-mortar retailer doesn’t become a new Amazon when employees automates their routines. Those companies need a new way to operate with new processes and new roles for their employees. Uncovering existing processes and automating them might bring some savings, but if you create new ways to operate based on new tools, you can create a whole new business.

AI, digitization and automation (including RPA, robotic process automation) are at the heart of these changes. They are hype terms nowadays, and it is easy to make fun of them. Their reputations suffer if those technologies are not appropriately utilized; they become window-dressing, like lipstick on a pig. Suppose you put a little bit of AI and a little bit of automation on top of your old processes and systems. In that case, it is not making them more digital or intelligent, and it’s just adding one more layer of complexity and arguably, technical problems. Some companies would like to use machines to observe people and use AI to create automation to perform the same tasks. It sounds like an exciting tech vision, but it’s a strange idea that the optimal model for machines would be to copy how people have done something traditionally.

Henry Ford didn’t build a car for everyone by asking old-style workshop car builders to automate some of their routines. Jeff Bezos didn’t digitalize retail by asking guys who receive telephone orders and fill paper order forms to use VoIP calls and scan order papers. Google founders didn’t revolutionize the online ad business by making an online copy of the yellow pages. They created new models from scratch, how they could offer the best value to their customers with the latest technology. But many companies still try to develop their operations by adding new tricks to old models.

Automation, AI, and digitization will change most businesses, and they will significantly change the way information works. Improving existing processes is a multibillion-dollar opportunity, but creating new, more effective models to operate in hundreds of billions or trillions. Improvements bring short term wins; new operating and business models create companies that prevail in the future.

All these require courage from management and investors. They must be brave enough to discard old models to operate and old systems. It is nice to promise each employee that nothing will change or promise two per cent stable growth to investors. Still, as we have seen in retail, this model leads to huge collapses, significantly when competitors change the business and market rules. Those leaders who want to create big successes should start to build their operations based on software robots, AI and digital processes, not just hope the old models can be done a little bit better. And they should start today.

The article was first published on Disruptive Asia.

A personal trainer gives you instructions on what to do at the gym. In most cases, she or he asks only basic things from you, like your target, to either lose weight or grow muscles, and maybe how often you have visited the gym before. A growing group of wellbeing consultants tell you, how to sleep, eat and work better. They might ask you to keep a sleep and food diary. These days, people have more and more wearable devices to measure daily activities, heartbeat, sleep, blood glucose and many other things. But there is still a very weak link between data, wellbeing and training services. However, this will change.

I have read about sleep consultants whose primary task is to teach people to repeat some words when they try to sleep. They say it helps you to relax and sleep better. However, people nowadays have several devices that measure their sleep, heart rate when they go to sleep, sleep intervals, even body temperature and how tough their day has been. Wouldn’t it be better if those sleep consultants could utilize your data, and not only teach mantras?

During the COVID lockdown, many fitness centers were closed. They started to offer online services, including virtual personal trainer sessions, online exercise classes and videos on how to train at home. But this is mainly one-way communication. The fitness center doesn’t take your data to create a more personalized plan for you. Why not? Technically it would be quite feasible, but they would have to develop new services for this model. Many customers would be ready to pay more for personal services than standard classes.

The world is full of services to lose weight. People pay for online services to get instructions for daily eating and exercising. Some services help track your calories when you record your daily food entries. Most services are still elementary and don’t use data available from wearable devices. Nowadays, you can even track blood glucose in real-time. It would be quite useful with exercise, heart rate and sleep data for personal weight control services.

The wearable market is increasing. The smartwatch market, in particular, is growing steadily, approximately 20% annually based on market research and is expected to reach almost $100 billion by 2027 from $150 billion this year. Smartwatches take market share from some other early devices that only measured steps and heart rate data, basic things. At the same time, new categories are growing, like smart rings (e.g. Oura) and blood glucose, metabolic health apps (e.g. Levels and Veri). Withingswas part of Nokia for some years, but Nokia sold it back to its founders and wrote it off, just when the market started to grow. It is a company that has a more extensive range of products from watches to digital blood pressure and under-mattress sleep tracking equipment.

So, people are starting to gather a lot of personal data. But many people are still confused, how to utilize this data. Apple Health is a service that helps combine data from several devices if you have an iPhone. But it is probably the most confusing and worst UX product Apple has. As with business data, people need tools to utilize the data, and raw data is hard to understand.

There are also other health data sources. DNA tests offer information on personal genetic profiles. Digital health care records are starting to become available in some countries. This data could also be combined with wearable data.
This sounds like a perfect match. Wellbeing services should start to become more personal and based on real data, not only some standard instructions, because people are, in fact, individuals and different. Wearable devices provide more and more data points that are hard to interpret. Both those parties could improve their businesses if they learned to utilize the other party’s services better.

How can this happen in practice? There are, at least, three ways to do this:

1. Wearable device manufacturers could start to offer more apps and services to utilize data in daily life. They will probably do something in this area, but it is not their core business, and people should be able to combine data from many sources, not only use data from device-specific silos.
2. Wellbeing services could start offering services to collect data from different sources and develop ways to utilize them. But most of these service providers (gyms, personal trainers, or wellbeing consultants) are not data technology experts.
3. There will be players who help collect data from many devices and sources and offer it in an easy format. Third parties can make applications for people and wellbeing service providers to utilize the data. This is the most viable path to work with many data sources, have data technology competence, and work with many wellbeing service providers. This is also the best solution to guarantee the privacy of the data.

Any professional business consultant usually analyzes a company’s numbers and processes before starting to give instructions. It would be bizarre to have a consultant that would try to get a company to better health, without looking at its existing data. But in wellbeing consulting it is still very typical. This will change in the next few years, and we’ll see wellbeing services based on actual personal data. And this market will grow fast; people are ready to pay for better overall health and wellness.

The article first appeared on Disruptive Asia.