Earlier this year, several Belgian IT journalists started paying considerable attention to the new generation of artificial intelligence, or NextGen AI. The tone of the articles suggested that NextGen AI is a disruptive threat to many professions, from teachers and lawyers to communications specialists. Here, we delve into this pressing issue.

Most of us will indeed have already read about ChatGTP, a product made by OpenAI, into which Microsoft is investing a lot of money. It is an example of what I term NextGen AI. Original AI tools were based on what is known as Artificial Narrow Intelligence and were mainly used to automate tasks or create schedules. But with NextGen AI, things go a step further with what is called Artificial General Intelligence, wherein the power of intelligence is already far broader and more profound. If you input the right questions, you can use it to write texts, draft CVs, create contracts or answer letters of complaint. Thus, it does not work like a traditional search engine (think Google or Bing), where you are presented with a series of links in response to your queries. Instead, you immediately get a full text that appears ready to use. I say it appears because this is where it hits a slight snag.

In its current iteration, ChatGTP can undoubtedly boost productivity for communications consultants. Those who don’t use it will be at a disadvantage to those who do, if only from a time-saving perspective. Starting with a text that has already been prescribed based on a set of instructions means the final, finished text can be produced more quickly – though human intervention is still required in proofreading and correcting or supplementing texts where necessary. Sander Duivenstein and Thijs Pepping, researchers at Sogeti Labs, wrote in an opinion piece in the Dutch newspaper NRC that there is a big difference between a calculator and Chat GTP. Whereas the outcomes of a calculator are deterministic, those of NextGen AI are probabilistic. The result will always be the same if you repeatedly enter the same calculation into your calculator. With ChatGPT (and similar technologies), this is not the case. Everything depends on the context you put in, meaning that results will vary according to how you formulate your question. NextGen AI will use artificial intelligence to give you the most likely outcome (i.e., probabilistic), which may vary from one time to the next. ‘ChatGTP doesn’t understand anything,’ the two researchers argue, ‘It is more like a parrot repeating a set of coincidences.’

It is precisely why human intervention is still needed, as nothing can replace real expertise in particular subject areas. ChatGTP is a valuable tool for storytelling or formulating press releases, for example. But this technology is evolving at lightning speed. Undoubtedly, NextGen AI will continue to gain momentum and become increasingly self-teaching on its march towards perfection. Just think of developments such as deep fake and the metaverse, both relying heavily on AI and NextGen AI. Therefore, as knowledge workers, we had better keep a close eye on it, following each new development.

As with many fellow consultants, I already have ChatGTP open by default on my second screen. Who knows that ChatGTP did not write this blog post on NextGen AI? You will have to guess that for yourself…

This article was originally published on ITdaily in French. To access the French version of the article, click here.

Numerous studies have emphasized that climate change is expanding, accelerating, and escalating to an extent that some changes are already irreversible. The United Nations has declared it a “code red for humanity,” and averting a calamity will only be possible if the world takes prompt action. Predictably, investments in digital transformation are set to increase by 50% from 2022-2025, leading to a further surge in energy consumption. The increasing digitalization has made it apparent that software applications are also leaving a significant impact on the environment.

At the same time though digital solutions are expected to help having the right insights to decrease footprint: Fifteen digital technologies are set to reduce C02 by 5 times than the total digital emissions by 2030 | Agoria and Digital technologies can cut global emissions by 20%. Here’s how | World Economic Forum (weforum.org)

The fast-paced development of international standards and disclosure regulations related to sustainability, coupled with mounting expectations from multiple stakeholders such as customers, investors, regulators, and employees, is putting immense pressure on businesses to invest in sustainability. Consequently, sustainability has now become one of the top ten business priorities for CEOs, according to the 2021 Gartner CEO and Senior Business Executive Survey. As an example, France has initiated the implementation of a distinct green software legislation that will mandate companies to disclose the environmental cost of their products to both users and the government. This legislation is expected to come into effect shortly.

Sustainability – Some Facts and Figures

• In a recent survey from IDC 83% of respondents agree sustainability is the most important criteria in IT buying decisions. Cloud computing hereby is a major differentiator when it comes to sustainability compared to traditional on-premises infrastructure.

• A new Gartner, Inc. survey  (https://www.gartner.com/en/newsroom/press-releases/2022-11-14-gartner-survey-finds-87-percent-of-business-leaders-expect-to-increase-sustainability-investment-over-the-next-two-years) revealed that :
  • 87% of business leaders expect to increase their organization’s investment in sustainability over the next two years. Customers are the primary stakeholder group creating pressure for organizations to invest or act on sustainability issues, selected by 80% of executives, followed by investors (60%) and regulators (55%).
  • 86% of business leaders see sustainability as an investment which protects their organization from disruption.
  • 83% said sustainability program activities directly created both short- and long-term value for their organization.
  • 80% indicated that sustainability helped their organization optimize and reduce costs (nbr 1 being energy consumption).

Sustainability is one of the hottest topics in IT currently

Even though there have been ongoing discussions regarding the environmental impact of IT, including data centers and public clouds, the conversation about the contribution of software, rather than hardware, to the sustainability of the IT industry is relatively recent.

However, it is crucial to note that software systems also emit carbon emissions indirectly through the energy consumed by the physical hardware they run on.

Greening of software

According to the Green Software Foundation, green software refers to software that produces minimal greenhouse gas emissions. To decrease the carbon footprint of software, the primary measures to be taken are

• using fewer physical resources
• reducing energy consumption,
• and utilizing lower-carbon energy sources.

It’s likely that several developers don’t consider the carbon footprint or energy usage of the code they create. However, the reality is that software significantly influences energy consumption and consequently, environmental sustainability. Nonetheless, sustainable software development is expected to become an increasingly prominent topic in the future as companies search for approaches to become more eco-friendly at every layer of the technology stack, including software development and not just hardware.

The majority of hyperscalers have managed to optimize their hardware and reduce energy consumption, despite significant growth in computational demands within their datacenters. However, due to ongoing concerns about energy consumption, it is important for us to examine the software applications being run on this hardware. Every developer has a role to play in addressing climate change by implementing best practices in Sustainable Software Engineering.

“The way we design and build our applications has a direct relationship to how much carbon they emit. With a better understanding of the impact our application designs have on the environment, we can make choices which have a more positive impact on the planet.”

— Paul McEwen, Global Head of Technology Services for UBS

^{(*) Quoted from https://greensoftware.foundation/articles/carbon-aware-computing-whitepaper-how-ubs-succeeded-in-measuring-and-reducing-car (article licensed under Creative Commons (CC BY 4.0))}  

Sustainable Software Engineering

Microsoft has recognized ‘Sustainable Software Engineering’, also referred to as Green Software Engineering, as an emerging discipline, and rightly so. This practice is just one aspect of sustainability in IT. The Principles of Sustainable Software Engineering encompass a fundamental set of skills required to create, develop, and maintain sustainable software applications. These applications aim to minimize greenhouse gas emissions while delivering the same level of business value and generating a positive impact on the environment.

There are many advantages to building sustainable applications

• They are almost always cheaper to run
• They are often more performant
• They are often more optimized

The Sustainability Workload Documentation by Microsoft would be a nice guideline on how businesses worldwide can plan their sustainability priorities in the future.

In follow-on blogs in this series, I’ll talk more about how developers can contribute to a greener environment by applying above principles.

While there is a huge adoption of Generative AI across organizations and industries – our research reveals that over 95% of executives are engaged in Generative AI discussions in their boardrooms – we can observe clearly a shift in the way people perceive AI now.

I have been working in the field since many years, and the unprecedent enthusiasm around Gen AI is impressive – 74% of executives believe the benefits of generative AI outweigh the associated risks. Beyond the positive feedback around it, there is a massive need for information, education and guidance. Especially for organizations to successfully and responsibly implement Generative AI across their data value chain, considering ethics, privacy and security from the start.

However, when you leverage Generative AI in a secured and trusted environment the opportunities are immense. From tasks and workflow optimization, to content production, product innovation and R&D, it is revolutionizing the way we create, interact and collaborate, completely shifting the way organizations operate. What if you could as a CXOs leverage Gen AI, across your organization, in a safe, secured and controlled manner, to fit your business reality?

For pretty much every company, it is already an undeniable fact: the cloud will form an essential part of business operations in the coming years. But at the same time, many companies are unsure about how to make this a reality. How do you start your journey to the cloud? And how do you make your cloud project a success story? We asked Paul Vanderborght, Managing Director Cloud Infrastructure Services at Capgemini Belgium, and his colleague Danny De Paepe, Cloud Solutions Sales Lead, for the best transportation and rehousing tips.

Tip 1: Start from what your business wants

A journey to the cloud – and you can definitely call it a journey – is usually a long and complex one, requiring careful thought and planning. Just like with any other IT project nowadays, you start off by looking at your business needs. The cloud overcomes many challenges, and you can choose from a wide range of cloud solutions. In these choices, you must always be guided by what your company wants to achieve with its cloud project. Do you want to stop using an old, electricity-guzzling data centre, and so do your business in a more sustainable way? Do you need extra flexibility in your infrastructure, so you are better equipped to deal with peak loads? Do you hope to have a lower TCO through the cloud? How important is 24/7 availability and support for your business? Would you like to respond more quickly to the products and services of your suppliers, who are increasingly developing their software for the cloud first? It is crucial that you have complete and clear answers to these questions and that you take account of all the relevant business needs before you set off on your journey. It’s only once you have these answers that you can make decisions about technology, such as how much of your infrastructure you should put in the cloud, what type of cloud appeals to you, whether you should opt for one or several suppliers, and so on.

Tip 2: Know what you have

After the business analysis, you have to conduct a somewhat technical exercise: getting a good overview of your application landscape. What applications do you have? What applications communicate with each other or with external parties? All this needs to be mapped out. Most businesses do this mapping exercise well, for the majority of their applications. But often, the devil is in the detail. Older applications are frequently overlooked ­– even though they are crucial for a successful cloud migration. Luckily there is help at hand: expert guidance for companies such as Capgemini, and automatic technology that facilitates an end-to-end view of the application landscape.

Additionally, these tools are not just limited to applications. They also map out infrastructure and security, and make an initial proposal about what components are best migrated to what type of cloud. Based on this, we can work with the client to plan out an optimal journey to the cloud. Some applications require a lift and shift approach, meaning they are migrated to the cloud ‘as is’, others an upscaling, which entails rewriting the application and placing it in a container, and others still require a full rewrite from scratch for the cloud, or simply need to be replaced by an external cloud service.

Tip 3: Think about the foundations

You can compare your journey to the cloud with building a house. First and foremost, you need good foundations. These foundations must be in the right location, and they must be deep enough to support the building you are constructing on top. In cloud environments, we talk about the landing zone: where are you going to place each application, so that they have optimal support? Making the right choices in this phase often determines whether your cloud journey will succeed, or fail. You can make ‘quick and dirty’ decisions about everything, but the chances are that these decisions will be ill-considered and require costly rectification down the line.

It is also important to make your landing zone future-proof. Bear in mind that, in the long term, you will want to build an extension on your house. If your foundations are not strong enough, you will simply not be able to. At the same time, your foundations must be flexible, so that you do not get bogged down from the start. New technologies may emerge that were not anticipated, and then your foundations must allow their use.

Tip 4: Ensure you have decent security

The journey to the cloud is also the ideal opportunity to review your security choices. For many of our clients this is a wake-up call: a moment to consciously reflect on what they need in terms of security and what exactly they already have in place. Here, they must take into account the nature of their activities – security needs to vary strongly, depending on the sector – as well as their customer base and geographical reach. Companies who serve customers across the world need different security provisions than those who serve a narrower, local clientele.

Tip 5: Build in good governance

People often ask us if the cloud ends up being more expensive than keeping everything on their premises. Our answer to this is a resolute one: you can always make a positive business case, by bearing the following in mind. If you take your kids to the supermarket, you can be sure that half of the sweets and chocolates on the shelf will land in your trolley. The same applies to the cloud: if you allow everyone to choose what they want, you will end up with far more than you actually need. The challenge is to make everything easy to consume – which should remain a demonstrable advantage of the cloud – and at the same time ensure that consumption remains limited to what you need as a business.

In other words: with the move to the cloud, you must provide good governance so that its use remains under control. Happily, there are also tools available for this, which give you a full and accurate overview of who is consuming what in the cloud. This means you can keep a continual oversight of use by different users, and how much this is costing. This not only ensures that everyone pays attention to what they are using, but you can also quickly identify when a cloud service is no longer being used, and so a way to save money. With such governance in place, the cost of the cloud is usually much lower than you think. In fact, the businesses we have helped migrate to the cloud have saved up to 30% (or more) when doing so.

This does require that there is good awareness within the different departments of a company and that they take responsibility for what they are doing. If they know what the cost of a particular extra cloud service is, they will think more carefully before using it and perhaps look at the business case. This means less impulsive consumption behaviour, and so a lower bill at the end of the month.

Tip 6: Dynamic environments require dynamic management and maintenance

You need to provide support in another way. Your cloud environment is a very dynamic setting. At the same time, the expectations that everything will be quicker and more easily available are higher than ever. As an IT service, this is very difficult for you to handle on your own. To continue to deliver a professional service, you need automated cloud management and support. The basis of service automation is carrying out operational tasks with artificial intelligence (AiOps), which saves enormous amounts of time and entails a much smaller chance of mistakes being made than with manual, human interventions. As Capgemini has developed this AiOps environment itself, it is much easier for us to adjust this to the needs of each client, whether they have a few hundred servers or tens of thousands.

Tip 7: Look for a good travel companion/construction supervisor

Capgemini supports its clients at every stage of the journey, from analysis and implementation to support and management. Furthermore, we help with both the technical and business sides: we look together at what the business needs are, but also provide advice about what is technically possible, and what is not. So the client knows they are in safe hands, even when their most important application is being migrated.

Interested in transforming your business with Cloud? Get in touch!

This article was originally published on Computable in Dutch. To access the Dutch version of the article, click here.

Working with the folks at the Capgemini Research Institute is always fun, when we set up the surveys the intent is always to find things out rather than prove them. With the latest report on Consumer adoption of Generative AI that mindset has led to me losing a friendly wager. Because I was absolutely sure we’d see a generational divide with GenZ in particular being ahead of the game. This is because that is what we nearly always see with new technologies. Instead what we saw is that actually adoption was pretty flat across all groups.

That is over 50% of the 10,000 people interviewed who have tried Generative AI tools, realistically in the 7 months since ChatGPT launched.

50% penetration in 7 months. That is frankly amazing. But what I find more amazing, and heartening, is that this adoption isn’t a “TikTok” curve with younger generations far outstripping older ones, its pretty even across the board, with Boomers actually being a few points ahead of the game. When we look at where they are using it, well that gets even more interesting for me.

Although its a statistically small difference, its great to think of GenX and Boomers getting the edge in Fortnite by using Generative AI! Seriously though, this again shows a very even adoption across the age ranges, really aligning to the ease of use that Generative AI tools enable. It also plays to one of the old adages of technology: There are two industries that drive new technology adoption, Computer Gaming is the other one.

On satisfaction there are some interesting differences, with Millennials and Boomers noticeably preferring the gaming experience but also noticeably liking the search experience much less. Clearly a lot more research is going to be done over that search category, as that represents a very clear an obvious market battle ground, but with two thirds of Boomers already liking the experience it isn’t a case of bad needing to become good, its good needing to become great. As we know however when we saw Google take over from Altavista, the market impact of good to great is dramatic.

OK but there is too much trust here

One finding in there that worried the Trusted AI person in me was the level of trust that all generations have in the created content.

Again we aren’t seeing differences based on age, with nearly 75% of people across all age ranges saying they trust the content. For a technology to have this sort of approval after only 7 months is remarkable, and a little bit concerning. I might be more on the Trusted AI side, but I’ve had enough experience with several models now to say that my trust level in the content isn’t terribly high and I’m validating what I see. There is a risk here that misplaced trust becomes a problem. That this is constant across the generations is extremely interesting though, as often we’d expect older people to distrust new technologies more.

Caveats and more research

So this is a very detailed report that I advise you to read all of the report, and I’ll finish with a few caveats on the data, firstly out of the 10,000 people surveyed 600 were Boomers, that is still a significantly significant sample, but it is less than the number of GenZ people interviewed, which means that the GenZ numbers are statistically more accurate. The other piece is that the education levels are statistically high, with over 75% having at least an undergraduate degree. This means that while within those groups the age discrepancy is low, you can’t say that this is constant across education ranges.

Much more research remains to be done, but one thing is for sure, the ease of use of Generative AI is clearly something that isn’t age dependent.

This article was first published on Medium on June 21, 2023.

Data is everywhere. Over the last twenty years, the sheer volume of data we capture, process and store from different sources has continued to multiply at a staggering rate. Data is not just about the operational transactions that keep our business running or something we need to store, mask or delete to comply with regulations. Data informs our decision making. Data can help our business to grow, pivot or surge ahead of the competition.

To ensure that we make the right business decisions, understanding and managing our data is imperative, and that’s where data governance comes in.

What is data governance?

Data governance is, at its core, knowing your data and how you use it. While this sounds pretty fundamental, a surprising number of businesses don’t feel confident that they have this knowledge of how data is used by their people, processes and systems.

In 2020, Capgemini’s Data Powered Enterprise Digital Report found that only 25% of business executives surveyed said that their organisation had a complete picture of all the data inventory.

It’s a bit like managing your money. You need to know where it’s coming from (to ensure that you’re not getting it from the “wrong” places), how you’re looking after it and whether you’re using it to invest in the right things to be confident that your money is working for you.

Similarly, data governance is about knowing where your data has come from, how you’re maintaining it and how you’re harnessing its power to help your business thrive.

Do businesses understand their data?

There are a number of reasons why businesses might lose control of their data over time:

• Mergers and acquisitions might not take full account of the data that exists in the organisations being merged and can lead to data duplication, inconsistencies or gaps.
• Lack of control or documentation of IT systems can lead to systems being developed in isolation in an inconsistent way, without knowledge of the wider picture of data inputs and outputs.
• Business reorganisations and restructuring can create gaps in the business or the data. Removing roles might mean that you inadvertently remove processes and data, leaving gaps that need to be filled.
• A silo mentality within organisations can lead to a lack of overall visibility of the data in the organisation as a whole.

Why do we need data governance?

Through the eighties and nineties, data governance was conceptually separate from business operations.

Since the start of the millennium and with the advent of social media and lower connection costs, data is now all around us and data governance has therefore taken on a whole new significance.

As well as capturing data and storing it as part of our operational processes, data is also bought, sold and generated automatically by human activity and machines in huge volumes, including of course data produced and used by artificial intelligence and machine learning. It has become preferable to take a horizontal view of our data rather than the previous vertical model.

As well as changes within organisations and the external evolution of the data landscape, there are regulatory and compliance reasons for understanding and managing our data effectively. GDPR (the EU General Data Protection Regulation) is probably the most well-known example of a regulation that has defined rules for sourcing, storing and processing data, but it’s by no means the only one. Many regulations are sector-specific, such as within the pharmaceutical industry where data is exchanged between organisations and must be structured and transferred in a pre-defined way.

While the data within operational systems such as ERPs is more or less stable, the evolution of the data landscape outside of an organisation increases the complexity and therefore the need for control.

What are the benefits of data governance?

As well as complying with regulations and ensuring that your data is producing accurate insights, the benefits of having a deep understanding of your data landscape include:

• Increased capability to change the business model at enterprise or local level
• Decreased time to market for new initiatives
• Reduced cost of fixing “bad” data, some of which are hidden costs
• Competitive advantage over similar organisations that do not have accurate insights and/or the same agility for change
• Avoidance of fines from regulators
• Use of data as a predictive rather than a reactive force

According to Gartner’s 2019 report “Augmented Data Catalogs: Now an Enterprise Must-Have for Data and Analytics Leaders”:

“Organizations that offer a curated catalog of internal and external data to diverse users will realise twice the business value from their data and analytics investments than those that do not.”

And in their 2021 report entitled “Market Guide for Active Metadata Management”, they state:

“Through 2024, organizations that adopt aggressive metadata analysis across their complete data management environment will decrease time to delivery of new data assets to users by as much as 70%.”

What are the different elements of data governance?

Data governance is a broad concept, but it can be broken down into the following topics:

• Data discovery and lineage
• Data catalogue
• Business glossary
• Data quality
• Master data management
• Data lifecycle management
• Data protection.

A more recent concept that has entered the sphere of data governance is data mesh. This is an architecture that decentralises data to enable more specialised, domain-based ownership and insights, moving away from a centralised approach.

Which tools can help us govern our data?

It’s no longer possible for humans to manage all aspects of the data within an organisation, so data governance requires tools to help with this significant task.

At the foundation of effective data governance is the data catalogue, which brings together a comprehensive view of the data within an organisation through multiple lenses.

There’s a wide range of data governance tools available, some of which are highly specific (e.g. data catalogue creation) while others are modules within larger applications (such as master data management within an ERP).

Some of the major vendors in the data management tool space are:

• Informatica
• Alation
• Collibra
• Purview
• Semarchy
• Profisee

Some of these solutions started out with very specific, standalone functionality, such as data quality, but over time evolved to encompass a wider remit as the concept of data governance gained ground. Most of these are provided as SaaS (Software as a Service) or PaaS (Platform as a Service) solutions.

How do we know which tools to use?

Capgemini has been helping businesses with their data governance for many years and has partnerships with over ten industry-leading vendors. With its centre of excellence providing consultation, implementation and operational support, Capgemini offers the competency and experience to help your business select the right combination of tools to meet your data governance needs.

How do we implement data governance?

Data governance is an organisational problem rather than a technical problem, making it a journey rather than a one-time implementation.

A data-savvy, forward-looking entrepreneur starting a brand-new company could put all the right pieces of the puzzle in place at outset, and this will stand them in good stead for a while. However, the continuously evolving data landscape means that a business must have the right methodologies and assumptions available to maintain governance over time and continue to use data to their advantage.

A phased approach to data governance is often recommended, starting with one tool and moving to others as needs evolve. There’s no one-size-fits-all solution, and Capgemini always tailors its recommendations to suit the client based on country, business sector, applicable regulations and data complexity.

Conclusion

By gaining a comprehensive understanding of their data, business leaders can benefit from accurate insights, improved business agility, reduced time to market and significant cost savings through not having to fix bad data.

Given that only 25% of businesses are estimated to have a comprehensive understanding of their data, this means that there’s a huge opportunity for companies who manage to master data governance to gain a competitive advantage.

In an ever-evolving world, effective data governance helps companies to harness the power within their data and change their business at the flick of a switch.

In our constantly changing times, the relationship between a customer and a brand remains the most important, as new technologies are being introduced to connect these two. Amidst many technological advances, sometimes, it’s hard to build a trustful relationship that is manageable to maintain. Marketers get swamped with new technologies and CRM managers get overwhelmed with the amount of new data sources. That’s the moment when a tool that can unity data, customers, and marketing come in handy and that’s a Customer Data Platform.

What is a CDP?

The CDP Institute defines a CDP as “a packaged software that creates a persistent, unified customer database that is accessible to other systems.”

In other words, it is a tool that assembles and consolidates first-party customer data from multiple sources into a single view of each customer. The data sources may include:

• Behavioural data, for instance clients’ interactions with your website, app, live chat, or agents
• Transactional data, such as purchase information coming from your point-of-sales systems or e-Commerce
• Demographic data, such as the name, birth date and month, and address

How can your company benefit from a CDP?

In a single sentence, a CDP allows you to deliver personalized experiences at scale using data-driven insights, and without the real necessity of a third party. Taking a closer look, we can say that a CDP will empower you to:

1. Combine Data from different sources

   A CDP will ingest data from a variety of channels such as email, mobile applications, website, social media, ERP (Enterprise Resource Planning), CRM or DMP systems, and many more, whether they are structured, unstructured, or semi-structured. Using business rules and machine learning, a CDP will create a single view of each customer, so you can understand the clients’ behaviour, perform segmentation, assign a scoring, or orchestrate a meaningfully customer journey.

2. Integrate with other MarTech tools

   CDPs are intended to easily integrate with other marketing tools, thanks to pre-built connectors or APIs. That way, a CDP can become the single centre of customer data, for your organization.

3. Democratize Data

   As a single source of truth, a CDP will allow your entire organization to use this data and personalize your communication with your clients in every touchpoint, whether it’s sales, marketing, customer service, support, and so on.

4. Marketing Activation and Personalization

   Once your team have all the customer data gather in a single place, structured in profiles and easily accessible, they can act on it by creating audience segments and personalizing the customer experience through the entire journey, in every channel.

Why would you need a CDP?

Many organizations have access to plenty of customer data, but they might struggle to gather, understand, and act upon it. They may wrestle with tracking the customer across their different platforms, make recommendations based on their past behaviour or missed an interaction after a website visit.

CDPs can help organizations to tackle many of their daily problems. How can you identify if your organization needs a CDP? These are the main signals your organisation may need one:

• Disorganized data

  As the number of data sources continues to increase, it becomes increasingly complex to understand your customer and creates significant connections. If you find yourself in this situation and your data is disconnected, a CDP will be able to structure your data to make sense of each communication.

• Customer identification

  If your team is struggling to link customer identities across different channels, a CDP can easily solve these inconveniences by bringing together data from multiple sources and creating a unified profile for each customer.

• Complex segmentation

  Your organization might have an organized data structure and is able to identify the client in every interaction, but the team might struggle to make use of that data by segmenting their audiences and personalizing the communications. A CDP will offer an intuitive playground to segment your audiences, suggest clusters and produce personalized engagement.

Briefly, a CDP, together with the right data strategy, will allow your team to create a unified view of individual customers across online and offline touchpoints, allowing for real-time insights and interactions at just the right moment.

Why adding another tool to the current stack

You might be wondering, why add yet another technology if you’re already using a CRM or a DMP or even both. Here’s a quick illustration of the main differences between these three platforms that will help you evaluate a CDP for your organization:

As you can see, CDP combines what’s missing in CRM and DMP and making the most use of the data your organization has.

How to choose a CDP that works for your organization

With the quick rise of popularity of CDPs many companies have introduced their platforms, so it might be hard to choose the right one for you in the current market. In this article, we’d like to give you an overview of the main players in the market.

Adobe CDP and Salesforce Data Cloud are two leading customer data platforms that help businesses collect, unify, and activate customer data for better marketing and customer experiences.

Adobe CDP is a cloud-based platform that enables businesses to collect data from various sources, create a unified customer profile, and activate those data across channels for personalized marketing. It’s best suited for mid-sized to large enterprises that are looking for a scalable and flexible solution to manage their customer data.

Salesforce Data Cloud, claims to be a first-ever real-time CRM. It allows businesses to access and leverage third-party data for targeted marketing and sales efforts. It offers a range of data solutions, including B2B and B2C data, audience segmentation, and data enrichment. It’s best suited for businesses of all sizes that are looking to augment their first-party data with third-party data to improve their marketing and sales efforts.

Overall, both Adobe CDP and Salesforce Data Cloud are powerful platforms that can help businesses improve their marketing and customer experiences through better data management and activation. The choice between the two will ultimately depend on a company’s specific needs and goals.

As you can see CDPs is the new word in the digital market and they’re here to stay. At CapGemini, we’re happy to provide you with the expertise on any kind of CDP and help your business to confidently step into the new world of the customer 360.  Let us know if you’d like to learn more about the Customer Data Platforms, and we’d be happy to answer any of your questions.

Over the last few years, there seems to be a trend toward the increasing democratization of data modeling and analytics. In this spirit, SAP has introduced SAP Datasphere (formerly named SAP Data Warehouse Cloud) as its new cloud data warehouse solution. Brought together, SAP Datasphere and SAP Analytics Cloud (SAC) provide a complete data warehouse in the cloud including integration, modeling, and consumption. It thereby sets the pace for next generation analytics.

From IT to Business: A distribution of responsibilities

Last years have been highlighted by a fast popularization of data analysis, which lead to the question of making it easier for non-technical users to access and process data. Until recently, analytics were driven by IT, and users had limited autonomy. The focus was on large-scale reporting and the tools offered were aimed at data engineers. Gradually, the trend has shifted to self-service, allowing increased autonomy for business users through more user-friendly interfaces. Little by little, responsibilities are moving away from the IT department, leading to a true autonomy of the end users.

This trend is also visible in the so-called data mesh. A decentralized data architecture that structures data by specific business domain. Spaces in SAP Datasphere perfectly support this decentralized architecture. For business users, spaces can be considered as dedicated working environments to model and analyse data according to logical areas or business lines. For IT, those spaces consist of compute and storage resources that can be allocated to a specific project or team. Data mesh architectures facilitate self-service applications from multiple data sources, extending access to data beyond the more technical resources, namely data scientists, data engineers and developers. This creates teams that work independently and take full end-to-end ownership of their domain data. They are responsible for both the operational source data and the analytical endpoints. By making data more accessible through this domain-centric design, the data mesh reduces data silos and operational bottlenecks.

The end goal for these trends is similar, getting more people involved in the analysis process and thus unlocking the full value of the data. As a result, decisions can be made faster and more informed. There is no doubt that SAP Datasphere is designed for both business users and developers. It has a semantic layer that facilitates data analysis, and it allows everyone to speak the same language in terms of metadata, dimensions, and metrics, thus avoiding confusion over technical terms.

SAP Datasphere – A 3-layer architecture

SAP Datasphere seems to be breaking new ground in the world of data warehousing due to its ease of configuration and operation. It is an out-of-the-box solution that comes with standard, pre-installed business content, offering a fast time to market. Powered by SAP HANA, this tool provides end-to-end processes that span three different layers: data integration, data modeling and data consumption. Hence, SAP Datasphere comes as a link between the sourcing and the consumption of data.

Data Integration layer

SAP Datasphere facilitates the integration of data from different landscapes. The data can be integrated virtually or via replication. Both Smart Data Integration (SDI) and Smart Data Access (SDA) are supported. Integration possibilities are extensive, giving this tool a large degree of flexibility. It allows data from multi-cloud, hybrid and on-premise environments to be combined. And depending on changing requirements, the spaces can easily be expanded, reduced, or even put on hold. This offers some elasticity when it comes to managing computing resources and storage.

Besides, SAP Datasphere offers rich metadata management tools to ensure data quality and comes with data lineage capabilities. Using the lineage features, users get a clear overview of objects and spaces and can quickly retrieve data dependencies and navigate to the relevant objects, while understanding where data is coming from and where it will land.

Brought together with SAP Data Intelligence you can cover way more scenarios. Users can harness the power of heterogeneous data across multiple systems, integrating structured and unstructured data.

Data modeling layer

SAP Datasphere comes with two modeling layers, the data layer and the business layer.

The data layer, referring to the data builder, targets data engineers that model with a more technical approach. Models are created and maintained here. Python scripts can be used to apply custom transformations that are not supported out of the box. What’s more, the data builder will allow users to acquire and combine data, create tables/views on various data sources and build data flows for reporting purposes. One of the modeling options offered by SAP Datasphere is modeling using Graphical views. This option is very similar to the Calculations views feature available in previous toolsets. The main idea is to combine different tables or views into a single output in a graphical manner. These graphical views seem more intuitive compared to SQL views, which require some more technical knowledge. Once done, these views can then be consumed directly by SAP Analytics Cloud.

The business layer, i.e., business builder is meant for business users that want to semantically enrich those models. Business users can, amongst others, create business entities such as dimensions and analytics datasets in the business builder. What’s more, SAP Datasphere includes a new data catalog that acts as a centralized repository.

Next, the Datasphere incorporates Artificial Intelligence and predictive intelligence into the data warehouse. Since it has SAP HANA Cloud embedded into it, once data has been curated in SAP Datasphere, it can be enriched by various ML algorithms. For instance, as it is tightly integrated with SAP Data Intelligence cloud, users can benefit from certain extra features. Amongst others, the predictive capabilities that come with SAP HANA APL and PAL libraries. Users can write python to train a predictive model and integrate such models in the data flows. Both R and Python operators are available in the Data Intelligence data flows.

In most cases, the data integration will be handled by the IT team. They will manage data replication and federation and create a baseline model for each business unit in the Data Builder. Then, business users get access to add their own virtual models to the existing foundation and enrich the semantic layer.

Data consumption layer

SAP Datasphere is tightly integrated with SAP’s analytical platform: SAP Analytics Cloud.  Consuming SAP Datasphere content is done through the optimized story experience. Unified stories enable interactive exploration of data and help users to find insights and visualize information with charts and tables. Findings and presentations can be commented, shared, and presented to people from within and outside of the organization.

With the integration of SAC and it’s planning capabilities, businesses can easily adopt the hassle-free planning and data simulation across KPIs, and publish the planned data into the system with less involvement of IT.

Finally, advanced analytics is also possible via machine learning and R integration. Augmented analytics help organizations to gain insights from their data faster than ever before. Instead of spending hours or even days screening data and creating complex reports, augmented analytics can quickly identify patterns and relationships in the data, enabling users to derive key insights in real time. The smart features from SAC help users to automate some data wrangling steps, identify best influences, or even perform analyses using natural language capabilities.

Customers are not limited to SAC as a reporting solution as SAP Datasphere can easily connect with 3rd party tools such as Microsoft Power BI, Tableau or Microsoft Excel.

Target audience of SAP Datasphere

SAP Datasphere is aimed at both small and larger companies. Small companies find benefit in lowering the initial costs that can be contracted by implementing a data warehouse. Larger companies are given the benefit of modernizing their data warehouse without having to completely re-build their flows. Besides, it allows anyone to connect with applications and data sources, without the need for technical expertise.

SAP Datasphere is based on the decentralization of responsibilities and thus contributes to a real convergence between IT and business, reassigning responsibilities. In that sense, it provides business users with a highly abstract infrastructure, removing the complexity of managing the lifecycle of the product and enabling domain autonomy. SAP Datasphere offers seamless integration and radically simplifies the data warehousing landscape, providing graphical low-code modeling tools, a rich semantic layer, and self-service visualization capabilities. Together, they form the next-generation analytics platform.

Capgemini is already helping customers around the globe in setting up their next-generation analytics solutions with SAP Datasphere and SAP Analytics Cloud. For instance at one of Belgium’s leading Water distribution companies we set up a next generation analytics platform bringing together master data, transactional data, and streaming data to enable real time tracking of water flow and water consumption throughout the network. It unlocks intelligence, simplifies innovation, and lowers the total cost of ownership compared to peers.

Contact us now

Contact us

Select a value Alumni Business Career Customer Support Media Relations Others Partner Alliance Privacy Sell to Capgemini Purpose of contact *

Purpose of contact is not valid.

First Name *

First Name is not valid.

Last Name *

Last Name is not valid.

Email *

Email is not valid.

Job Title *

Job Title is not valid.

Phone (optional)

Company / Organization *

Company / Organization is not valid.

Select a value Aerospace & Defense Automotive Banking Chemicals Consumer Products Distribution & logistics Education & research Government General services Healthcare Insurance Industrial Manufacturing Life Sciences Media & Entertainment Natural resources Other Financial Services Public Security & Police Retail & Wholesale trade Tax & Welfare Travel & Transport Technology Telecoms Utilities Industry *

Industry is not valid.

Select a value Assistant/secretary CDO/CMO CEO CFO Chairman/Board of Directors CIO/CTO Consultant/Senior Consultant Director/Principle Consultant EVP Manager Operational Other CXO Purchaser Senior Manager/Managing Consultant SVP Vice President Position/Level *

Position/Level is not valid.

Country

Country Afghanistan Aland Islands Albania Algeria American Samoa Andorra Angola Anguilla Antarctica Antigua And Barbuda Argentina Armenia Aruba Australia Austria Azerbaijan Bahamas Bahrain Bangladesh Barbados Belarus Belgium Belize Benin Bermuda Bhutan Bolivia Bosnia And Herzegovina Botswana Bouvet Island Brazil British Indian Ocean Territory Brunei Darussalam Bulgaria Burkina Faso Burundi Cambodia Cameroon Canada Cape Verde Cayman Islands Central African Republic Chad Chile China Christmas Island Cocos (Keeling) Islands Colombia Comoros Congo Congo, Democratic Republic Cook Islands Costa Rica Cote D'Ivoire Croatia Cuba Cyprus Czech Republic Denmark Djibouti Dominica Dominican Republic Ecuador Egypt El Salvador Equatorial Guinea Eritrea Estonia Ethiopia Falkland Islands (Malvinas) Faroe Islands Fiji Finland France French Guiana French Polynesia French Southern Territories Gabon Gambia Georgia Germany Ghana Gibraltar Greece Greenland Grenada Guadeloupe Guam Guatemala Guernsey Guinea Guinea-Bissau Guyana Haiti Heard Island & Mcdonald Islands Holy See (Vatican City State) Honduras Hong Kong Hungary Iceland India Indonesia Iran, Islamic Republic Of Iraq Ireland Isle Of Man Israel Italy Jamaica Japan Jersey Jordan Kazakhstan Kenya Kiribati Korea Kuwait Kyrgyzstan Lao People's Democratic Republic Latvia Lebanon Lesotho Liberia Libyan Arab Jamahiriya Liechtenstein Lithuania Luxembourg Macao Macedonia Madagascar Malawi Malaysia Maldives Mali Malta Marshall Islands Martinique Mauritania Mauritius Mayotte Mexico Micronesia, Federated States Of Moldova Monaco Mongolia Montenegro Montserrat Morocco Mozambique Myanmar Namibia Nauru Nepal Netherlands Netherlands Antilles New Caledonia New Zealand Nicaragua Niger Nigeria Niue Norfolk Island Northern Mariana Islands Norway Oman Pakistan Palau Palestinian Territory, Occupied Panama Papua New Guinea Paraguay Peru Philippines Pitcairn Poland Portugal Puerto Rico Qatar Reunion Romania Russian Federation Rwanda Saint Barthelemy Saint Helena Saint Kitts And Nevis Saint Lucia Saint Martin Saint Pierre And Miquelon Saint Vincent And Grenadines Samoa San Marino Sao Tome And Principe Saudi Arabia Senegal Serbia Seychelles Sierra Leone Singapore Slovakia Slovenia Solomon Islands Somalia South Africa South Georgia And Sandwich Isl. Spain Sri Lanka Sudan Suriname Svalbard And Jan Mayen Swaziland Sweden Switzerland Syrian Arab Republic Taiwan Tajikistan Tanzania Thailand Timor-Leste Togo Tokelau Tonga Trinidad And Tobago Tunisia Turkey Turkmenistan Turks And Caicos Islands Tuvalu Uganda Ukraine United Arab Emirates United Kingdom United States United States Outlying Islands Uruguay Uzbekistan Vanuatu Venezuela Viet Nam Virgin Islands, British Virgin Islands, U.S. Wallis And Futuna Western Sahara Yemen Zambia Zimbabwe

Country is not valid.

Your Message *

Your Message is not valid.

Page title

Page URL

I agree to Capgemini collecting and processing my personal data to allow me to receive information on Capgemini services. For further information, please see our Privacy Notice.  

Submit

Thank you for contacting us.

We are sorry, the form submission failed. Please try again.

Disclaimer: This blog was NOT written by ChatGPT, but by a group of human data scientists: Shahryar Masoumi, Wouter Zirkzee, Almira Pillay, Sven Hendrikx and myself.

Whether we are ready for it or not, we are currently in the era of generative AI, with the explosion of generative models such as DALL-e, GPT-3, and, notably, ChatGPT, which racked up one million users in one day. Recently, on March 14th, 2023, OpenAI released GPT-4, which caused quite a stir and thousands of people lining up to try it.

Generative AI can be used as a powerful resource to aid us in the most complex tasks. But like with any powerful innovation, there are some important questions to be asked… Can we really trust these AI models? How do we know if the data used in model training is representative, unbiased, and copyright safe? Are the safety constraints implemented robust enough? And most importantly, will AI replace the human workforce?

These are tough questions that we need to keep in mind and address. In this blog, we will focus on generative AI models, their trustworthiness, and how we can mitigate the risks that come with using them in a business setting.

Before we lay out our trust issues, let’s take a step back and explain what this new generative AI era means. Generative models are deep learning models that create new data. Their predecessors are Chatbots, VAE, GANs, and transformer-based NLP models, they hold an architecture that can fantasize about and create new data points based on the original data that was used to train them — and today, we can do this all based on just a text prompt!

We can consider chatbots as the first generative models, but looking back we’ve come very far since then, with ChatGPT and DALL-e being easily accessible interfaces that everyone can use in their day-to-day. It is important to remember these are interfaces with generative pre-trained transformer (GPT) models under the hood.

The widespread accessibility of these two models has brought about a boom in the open-source community where we see more and more models being published, in the hopes of making the technology more user-friendly and enabling more robust implementations.

But let’s not get ahead of ourselves just yet — we will come back to this in our next blog. What’s that infamous Spiderman quote again?

With great power…

The generative AI era has so much potential in moving us closer to artificial general intelligence (AGI) because these models are trained on understanding language but can also perform on a wide variety of other tasks, that in some cases even exceed human capability. This makes them very powerful in many business applications.

Starting with the most common — text application, which is fueled by GPT and GAN models. Including everything from text generation to summarization and personalized content creation, these can be used in education, healthcare, marketing, and day-to-day life. The conversational application component of text application is used in chatbots and voice assistants.

Next, code-based applications are fueled by the same models, with GitHub’s Co-pilot as the most notable example. Here we can use generative AI to complete our code, review it, fix bugs, refactor, and write code comments and documentation.

On the topic of visual applications, we can use DALL-e, Stable Diffusion, and Midjourney. These models can be used to create new or improved visual material for marketing, education, and design. In the health sector, we can use these models for semantic translation, where semantic images are taken as input and a realistic visual output is generated. 3D shape generation with GANs is another interesting application in the video game industry. Finally, text-to-video editing with natural language is a novel and interesting application for the entertainment industry.

GANs and sequence-to-sequence automatic speech recognition (ASR) models (such as Whisper) are used in audio applications. Their text-to-speech application can be used in education and marketing. Speech-to-speech conversion and music generation have advantages for the entertainment and video game industry, such as game character voice generation.

Although powerful, such models also come with societal limitations and risks, which are crucial to address. For example, generative models are susceptible to unexplainable or faulty behavior, often because the data can have a variety of flaws, such as poor quality, bias, or just straight-up wrong information.

So, with great power indeed comes great responsibility… and a few trust issues

If we take a closer look at the risks regarding ethics and fairness in generative models, we can distinguish multiple categories of risk.

The first major risk is bias, which can occur in different settings. An example of bias is the use of stereotypes such as race, gender, or sexuality. This can lead to discrimination and unjust or oppressive answers generated from the model. Another form of bias is the model’s word choice. Its answers should be formulated without toxic or vulgar content, and slurs.

One example of a language model that learned a wrong bias is Tay, a Twitter bot developed by Microsoft in 2016. Tay was created to learn, by actively engaging with other Twitter users by answering, retweeting, or liking their posts. Through these interactions, the model swiftly learned wrong, racist, and unethical information, which it included in its own Twitter posts. This led to the shutdown of Tay, less than 24 hours after its initial release.

Large language models (LLMs) like ChatGPT generate the most relevant answer based on the constraints, but it is not always 100% correct and can contain false information. Currently, such models provide their answers written as confident statements, which can be misleading as they may not be correct. Such events where a model confidently makes inaccurate statements are also called hallucinations.

In 2023, Microsoft released a GPT-backed model to empower their Bing search engine with chat capabilities. However, there have already been multiple reports of undesirable behavior by this new service. It has threatened users with legal consequences or exposed their personal information. In another situation, it tried to convince a tech reporter he was not happily married and that he was in love with the chatbot (it also proclaimed their love for the reporter) and consequently should leave his wife (you see why we have trust issues now?!).

Generative models are trained on large corpora of data, which in many cases, is scraped from the internet. This data can contain private information, causing a privacy risk as it can unintentionally be learned and memorized by the model. This private data not only contain people, but also project documents, code bases, and works of art. When using medical models to diagnose a patient, it could also include private patient data. This also ties into copyright when this private memorized data is used in a generated output. For example, there have even been cases where image diffusion models have included slightly altered signatures or watermarks it has learned from their training set.

The public can also maliciously use generative models to harm/cheat others. This risk is linked with the other mentioned risks, except that it is intentional. Generative models can easily be used to create entirely new content with (purposefully) incorrect, private, or stolen information. Scarily, it doesn’t take much effort to flood the internet with maliciously generated content.

Building trust takes time…and tests

To mitigate these risks, we need to ensure the models are reliable and transparent through testing. Testing of AI models comes with some nuances when compared to testing of software, and they need to be addressed in an MLOps setting with data, model, and system tests.

These tests are captured in a test strategy at the very start of the project (problem formulation). In this early stage, it is important to capture key performance indicators (KPIs) to ensure a robust implementation. Next to that, assessing the impact of the model on the user and society is a crucial step in this phase. Based on the assessment, user subpopulation KPIs are collected and measured against, in addition to the performance KPIs.

An example of a subpopulation KPI is model accuracy on a specific user segment, which needs to be measured on data, model, and system levels. There are open-source packages that we can use to do this, like the AI Fairness 360 package.

Data testing can be used to address bias, privacy, and false information (consistency) trust issues. We make sure these are mitigated through exploratory data analysis (EDA), with assessments on bias, consistency, and toxicity of the data sources.

The data bias mitigation methods vary depending on the data used for training (images, text, audio, tabular), but they boil down to re-weighting the features of the minority group, oversampling the minority group, or under-sampling the majority group.

These changes need to be documented and reproducible, which is done with the help of data version control (DVC). DVC allows us to commit versions of data, parameters, and models in the same way “traditional” version control tools such as git do.

Model testing focuses on model performance metrics, which are assessed through training iterations with validated training data from previous tests. These need to be reproducible and saved with model versions. We can support this through open MLOPs packages like MLFlow.

Next, model robustness tests like metamorphic and adversarial tests should be implemented. These tests help assess if the model performs well on independent test scenarios. The usability of the model is assessed through user acceptance tests (UAT). Lags in the pipeline, false information, and interpretability of the prediction are measured on this level.

In terms of ChatGPT, a UAT could be constructed around assessing if the answer to the prompt is according to the user’s expectation. In addition, the explainability aspect is added — does the model provide sources used to generate the expected response?

System testing is extremely important to mitigate malicious use and false information risks. Malicious use needs to be assessed in the first phase and system tests are constructed based on that. Constraints in the model are then programmed.

OpenAI is aware of possible malicious uses of ChatGPT and have incorporated safety as part of their strategy. They have described how they try to mitigate some of these risks and limitations. In a system test, these constraints are validated on real-life scenarios, as opposed to controlled environments used in previous tests.

Let’s not forget about model and data drift. These are monitored, and retraining mechanisms can be set up to ensure the model stays relevant over time. Finally, the human-in-the-loop (HIL) method is also used to provide feedback to an online model.

ChatGPT and Bard (Google’s chatbot) have the possibility of human feedback through a thumbs up/down. Though simple, this feedback is used to effectively retrain and align the underlying models to users’ expectations, providing more relevant feedback in future iterations.

To trust or not to trust?

Just like the internet, truth and facts are not always given — and we’ve seen (and will continue to see) instances where ChatGPT and other generative AI models get it wrong. While it is a powerful tool, and we completely understand the hype, there will always be some risk. It should be standard practice to implement risk and quality control techniques to minimize the risks as much as possible. And we do see this happening in practice — OpenAI has been transparent about the limitations of their models, how they have tested them, and the governance that has been set up. Google also has responsible AI principles that they have abided by when developing Bard. As both organizations release new and improved models — they also advance their testing controls to continuously improve quality, safety, and user-friendliness.

Perhaps we can argue that using generative AI models like ChatGPT doesn’t necessarily leave us vulnerable to misinformation, but more familiar with how AI works and its limitations. Overall, the future of generative AI is bright and will continue to revolutionize the industry if we can trust it. And as we know, trust is an ongoing process…

In the next part of our Trustworthy Generative AI series, we will explore testing LLMs (bring your techie hat) and how quality LLM solutions lead to trust, which in turn, will increase adoption among businesses and the public.

This article first appeared on SogetiLabs blog.