The future of public administration lies in partnerships鈥攏ot silos鈥攚ith citizens, businesses, and civil society. In an era of rapid digital transformation, while the guiding principle of providing accessible, inclusive and high-quality public services remains fundamentally unchanged, the way public administrations are creating value for their citizens is undergoing a profound evolution.

As technology evolves and societal challenges grow more complex and interconnected, traditional siloed structures are increasingly being replaced by dynamic ecosystems where value co-creation is critical to the success or failure of public interventions.

In 2021, . Today, this approach has become a widespread foundational principle. Key technological enablers are driving this shift, empowering public administrations to move towards a collaborative approach of public service delivery that brings together governments, businesses and citizens to address challenges more effectively. From leveraging interoperability to dissolve boundaries and advance data-sharing ecosystems to the rise of GovTech, proactive service delivery and the transformative potential of government AI, these key trends are laying the groundwork for a smarter, more inclusive and efficient public governance designed to meet the demands of modern, interconnected societies.

Five trends in public administration

Time for action in an increasingly interconnected world

To fully harness the potential of these digital trends, public administration leaders must adopt an action-oriented approach. A combination of political commitment to digital transformation, inter-agency collaboration and leveraging robust PPPs to bridge resource gaps and accelerate innovation will be key. Together they will help to overcome budget constraints, siloed institutional frameworks, cultural resistance to change and complexities in measuring and reporting progress that still afflict public administrations worldwide. While strategically investing in cutting-edge technologies like AI, leaders must also champion a culture of continuous learning and upskilling among civil servants, ensuring they are equipped to leverage effectively these emerging tools. Ultimately, aligning digital strategies with citizens鈥� needs through human-centered service delivery will enable administrations to build trust, improve efficiency, and deliver meaningful public value in an increasingly interconnected world.

While Mulder trusts his gut; Scully trusts the facts – in fraud detection, we need both. Hybrid AI blends the intuition of LLM with the structured knowledge of a knowledge graph, letting agents uncover hidden patterns in real time. The truth is out there鈥攏ow we have the tools to find it.

Fraud detection can be revolutionized with hybrid AI. Combining the 鈥渋ntuitive hunches鈥� from LLMs with a fraud-focused knowledge graph, a multi-agent system can identify weak signals and evolving fraud patterns, moving from detection to prevention in real-time. The challenge? Rule sets need to be cast in iron, whereas the system itself must be like water: resilient and adaptive. Historically, this conflict has been unsolvable. But that is about to change.

A multi-agent setup

Large language models (LLMs) are often criticized for hallucinating: coming up with results that seem feasible but are plain wrong. In this case though, we embrace the LLM鈥檚 gut-feeling-based approach and exploit its capabilities to identify potential signs of fraud. These 鈥渉unches鈥� are mapped onto a general ontology and thus made available to symbolic AI components that build on logic and rules. So, rather than constricting the LLM, we are relying on its language capabilities to spot subtle clues in text. Should we act directly on these hunches, we would run into a whole world of problems derived from the inherent unreliability of LLMs. However, this is the task of a highly specialized team of agents, and there are other agents standing by, ready to make sense of the data and establish reliable patterns.

When we talk about agents, we refer to any entity that acts on behalf of another to accomplish high-level objectives using specialized capabilities. They may differ in degree of autonomy and authority to take actions that can impact their environment. Agents do not necessarily use AI: many non-AI systems are agents, too. (A traditional thermostat is a simple non-AI agent.) Similarly, not all AI systems are agents. In this context, the agents we focus on primarily handle data, following predefined instructions and using specific tools to achieve their tasks.

We define a multi-agent system as being made up of multiple independent agents. Every agent runs on its own, processing its own data and making decisions, yet staying in sync with the others through constant communication. In a homogeneous system, all agents are the same and their complex behavior solves the problem (as in a swarm). Heterogeneous systems, though, deploy different agents with different capabilities. Systems that use agents (either single or multiple) are sometimes called 鈥渁gentic鈥� architectures or frameworks.

For example, specialized agents can dive into a knowledge graph, dig up specific information, spot patterns, and update nodes or relationships based on new findings. The result? A more dynamic, contextually rich knowledge graph that evolves as the agents learn and adapt.

The power is in the teaming. Think of the agents Mulder and Scully from The X-Files television show: Mulder represents intuitive, open-minded thinking, while Scully embodies rational analysis. In software, there always have been many Scullys but, with LLMs, we now have Mulders too. The challenge, as in The X-Files, is in making them work together effectively.

The role of a universal ontology

We employ a universal ontology to act as a shared language or, perhaps a better analogy, a translation exchange, ensuring that both intuitive and analytical agents communicate in terms that can be universally understood. This ontology primarily consists of 鈥渇lags鈥� 鈥揼eneric indicators associated with potential fraud risks. These flags are intentionally defined broadly, capturing a wide range of behaviors or activities that could hint at fraudulent actions without constraining the agents to specific cases.

The key to this system lies not in isolating a single flag but in identifying meaningful combinations. A single instance of a flag may not signify fraud; however, when several flags emerge together, they provide a more compelling picture of potential risk.

鈥淭his innovation shifts the approach from simple fraud detection to proactive prevention, allowing authorities to stay ahead of fraudsters with scalable systems that learn and evolve.鈥�

Hybrid AI adaptability

The adaptability of the system lies in the bridging between neural and symbolic AI as the LLM distills nuances in texts into hunches. They need to be structured and amplified for our analytical AI to be able to access them. As Igor Stravinsky wrote in his 1970 book Poetics of Music in the Form of Six Lessons, 鈥淭hus what concerns us here is not imagination itself, but rather creative imagination: the faculty that helps us pass from the level of conception to the level of realization.鈥� For us, that faculty is the combination of a general ontology and vector-based similarity search. They allow us to connect hunches to flags based on semantic matching and thus address the data using general rules. Because we work in a graph context, we can also explore direct, indirect, and even implicit relations between the data.

Now let鈥檚 explore how our team of agents picks up and amplifies weak signals, and how these signals, once interwoven in the graph, can lead the system to identify patterns spanning time and space, patterns it was not designed to identify.

A scenario: Welfare agencies have observed a rise in fraudulent behavior, often uncovered only after individuals are exposed for other reasons like media reports. Identifying these fraud attempts earlier, ideally at the application stage, would be extremely important.

Outcome: By combining intuitive and analytical insights, authorities uncover a well-coordinated fraud ring that would be hard to detect through traditional methods. The agents map amplified weak signals as well as explicit and implicit connections. Note also that the system was not trained on detecting this pattern; it emerged thanks to the weak signal amplification.

One of the powers of hybrid AI lies in its ability to amplify weak signals and adapt in real time, uncovering hidden fraud patterns that traditional methods often miss. By blending the intuitive insights of LLMs with the analytical strength of knowledge graphs and multi-agent systems, we鈥檙e entering a new era of fraud detection and prevention 鈥� one that鈥檚 smarter, faster, and more effective. As Mulder might say, the truth is out there, and with the right team, we鈥檙e finally close to finding it.

Start innovating now 鈥�

Implement a universal ontology

Create a shared ontology to bridge neural (intuitive) and symbolic (analytical) AI agents, transforming weak signals for deeper analysis by expert systems and graph-based connections.

Form specialized multi-agent teams

Build teams of neural (real-time detection) and symbolic (rule-based analysis) AI agents, each specialized with tools for their role.

Leverage graph technology for cross-referencing

Use graph databases to link signals over time and across data sources, uncovering patterns like fraud faster, earlier, and at a lower cost than current methods.

Interesting read?

乌鸦传媒鈥檚 Innovation publication, Data-powered Innovation Review – Wave 9听features 15 captivating innovation articles with contributions from leading experts from 乌鸦传媒, with a special mention of our external contributors from, and .听Explore the transformative potential of generative AI, data platforms, and sustainability-driven tech. Find all previous Waves here.

Europe is building a data-driven economy in a changing geopolitical context. As it strives for both innovation and sovereignty, decentralized ecosystems offer a way to create value with data, while safeguarding freedom of choice.

Data has the potential to transform processes, businesses, economies, and society by unlocking new kinds of value creation. It鈥檚 also how we are going to make AI work as a crucial component of the future European data economy鈥攂ut only if that data is built on strong foundations that ensure its quality and relevance.

Of course, value creation depends on the data that鈥檚 available to you, and you might not have all the data you need. That鈥檚 why data needs to be shared and combined. In this article, we consider how data spaces meet this need, offering what the as the 鈥渁bility to provide the essential foundations for secure and efficient data sharing鈥�. While our focus in this article is on European data spaces, we recognize that this is becoming a relevant topic around the world.

Why a decentralized data economy makes sense for Europe

Data spaces are, in effect, decentralized ecosystems that have a powerful resonance in the world today. Indeed, recognizing their huge potential, the European Commission established a series of domain-specific/sectoral common European data spaces designed to help 鈥渦nleash the enormous potential of data-driven innovation鈥�.

We see three main drivers for these common data spaces in Europe: geopolitics, commercials, and choice. In the first instance, in light of the unstable geopolitical landscape, data spaces give you assurance that all your (data) eggs aren鈥檛 in one basket.  You select the datasets you want to reside in what data space. Interoperability and portability can help avoid the dreaded lock-in effect where changing from one service provider to another might be prohibitively complicated. Commercially, data spaces address any exposure to potential monopolistic lock-in effects by individual companies cornering the market in data platforms. Then there鈥檚 the matter of choice. You choose who you interact with in a common data space, which puts you in control of who to share data with.

Why we need data spaces

Sharing data is key to data-driven growth. Indeed, it鈥檚 a vital aspect of the European strategy for data. But over-reliance on data platforms predominantly controlled by a limited number of international technology firms introduces potential vulnerabilities regarding data security, access, and strategic autonomy. We may also lose the ability to share data on our own terms, in accordance with our own values鈥攆reedom, privacy, control.

An alternative future for Europe is to share data on a sovereign basis. And that implies across industries. That鈥檚 why we鈥檙e so excited to be working on the DSSC and on Simpl, the open source, smart and secure middleware platform that supports data access and interoperability among European data spaces.

Beyond technology to value creation

Let鈥檚 not forget that a data space is only an instrument. It鈥檚 what you do with it that matters. In a data space you will be able to aggregate, combine and correlate data that you can鈥檛 today because it is stored in different places. And that鈥檚 where we begin to create significant value from data, specifically in a number of areas, as follows.

1. Global challenges: Data spaces will prove inordinately useful in tackling grand challenges that cut across sectors and geographies. Here we鈥檙e talking about achieving mission-oriented policy goals, such as reducing healthcare inequality and achieving net zero/carbon neutrality targets. For example, the (EHDS) will be an enabler of patient empowerment, with better access to and control over health data. Further, increased reuse of EHDS data for research and policy making will improve public health interventions. A 2025 from the World Economic Forum in collaboration with 乌鸦传媒 suggests the EHDS could generate 鈧�5.5 billion savings over ten years. We鈥檝e already seen the huge value of data sharing in a global crisis when, in the Covid 19 pandemic, our governments needed data from many areas at once to form policy鈥攈ealthcare systems, pharma, mobility, employment and economic data. There will be future pandemics.
2. Innovation: Data spaces will undoubtedly contribute to data-driven innovation across the EU as it continues on its mission to build the Single Market for Data. The states, 鈥淐ommon European Data Spaces will enhance the development of new data-driven products and services in the EU, forming the core tissue of an interconnected and competitive European data economy鈥�.  In this respect, the combination of data from different sources across sectors can produce fascinating new applications. Think, for example, of the traffic flow in a city, where the observation of vehicle movement and a subsequent adjustment of traffic lights can help avoid congestion, and the monitoring of parking lots eases the burden of finding a parking spot, possibly connected to a recommendation of a charging port for the car鈥檚 battery. The energy grid could then be supplied with better anticipation of demand peaks and control energy distribution accordingly.  The seamless integration of real-time public transport data can then be used to recommend the best option for getting from A to B.
3. Efficiency: Data spaces will help in the more efficient use of resources and improve public services. A great example here is that of road surface observation. By correlating data from cars鈥� electrical sensors, it becomes possible to monitor, in real time, the deterioration of the road, and carry out preventative maintenance to optimize spend / return. And returning to the healthcare sector, access to comprehensive patient histories in a shared data ecosystem has the potential to lead to better and faster diagnosis and treatment.
4. Science and research: Shared data can create new evidence bases for scientific and medical research. Let鈥檚 consider the following scenario鈥擨 drive to work in a convertible most days; the farmer of the field sprays an experimental fertilizer; later I develop  neurological issues but doctors are unsure how to treat them. In the future we might be able to correlate this illness with the exposure to the fertilizer by aggregating mobility data, air quality data, times that the farmer used the fertilizer, and the contents of that fertilizer.

Questions at the edges of our data economy

The potential for value is clear, but there are numerous challenges still to overcome鈥攁nd they are not principally digital ones. One unknown factor is what it will cost to set up and run a common data space. At this point we don鈥檛 have an adequate way to price data, so this question remains unanswered. Other questions include: How can we quantify the value of new data-driven business models vs traditional business models? And how can we pinpoint the strengths and weaknesses of data ecosystems and technologies?

The answer to all of these questions at present is that we are all on a journey with common data spaces. We improve every day and the answers will come. But it is hard to imagine that the massive contribution of sharing data to the common good will not outweigh the costs and barriers that need to be overcome.

Above all, the decentralized model depends on participants鈥� willingness to share data. That means they must trust the other participants and the infrastructure. There is no other way to build trust except enabling people to say no. Letting people choose in itself invites trust.

Europe can do data differently

Data spaces are a way for Europe to reap the benefits of data for economic growth and positive societal outcomes, while affirming European values in the digital domain. They remain an integral part of the European strategy aiming to make the EU a leader in a data-driven society.

Find out more

Peter Kraemer will speak about the future of data sharing in Europe at the Data Spaces Symposium in Warsaw on 11-12 March. Register at

The key to success in digital, service, and product monetization lies in a sound monetization strategy that connects consumer and industry insights.

In today鈥檚 interconnected world, the challenge of successfully monetizing digital products and services has taken center stage. Digital technology continues to drive transformation across industries, but businesses must align their monetization strategies with these innovations to fully unlock their potential. It is surprising that many companies struggle to do so effectively. Research shows that just 53% of companies successfully implement digital monetization, revealing a critical gap. This article examines the obstacles and strategies necessary to navigate the monetization landscape successfully.

Why is monetization so challenging?

Monetization has become increasingly complex in a digital-first world, where companies are under constant pressure to innovate and meet evolving customer expectations. More than just generating revenue, successful monetization ensures business sustainability and enhances the overall customer experience. Without the right approach, businesses risk missing out on the full potential of digital monetization.

Whilst many companies who are implementing the right monetization strategies have successfully established new revenue streams; broadening their customer base and strengthening their overall market position, many are still getting lost in this monetization jungle, and therefore not realizing the full benefit. This can be attributed to several key challenges that companies face when implementing their strategies. Let鈥檚 explore these challenges in greater detail and examine the common barriers to successful monetization.

With these challenges in mind, let鈥檚 now take a closer look at the different approaches companies can use for monetization.

Direct and indirect monetization strategies

In tackling these challenges, it is essential to distinguish between direct and indirect monetization strategies. Direct monetization involves generating revenue directly from a product or service, while indirect monetization leverages secondary or complementary revenue streams. 

• Direct Monetization: This includes strategies such as one-time purchases, subscription models, pay-per-use approaches, and freemium services. Each of these models has its own advantages. For instance, a one-time payment offers transparent pricing, while subscription models ensure consistent revenue streams. The freemium model, where users are offered a basic free version and then enticed to pay for additional features, is particularly effective for building a large user base and crafting a strong and relevant service monetization offer. 
• Indirect Monetization: Companies can generate revenue indirectly through advertising, transaction-based models, or data monetization. Data monetization, in particular, is becoming increasingly valuable as companies collect, enrich, and sell customer data to third parties or use it to enhance their services internally. This creates new opportunities to drive revenue while improving the customer experience. 

Having discussed the key strategies, we now turn to an equally critical component – how these strategies align with pricing models. 

Aligning monetization with pricing strategy 

Any successful monetization strategy must also be aligned with a robust pricing strategy. Businesses need to carefully consider several factors to set the right prices, including: 

• Cost Structures: A comprehensive understanding of all associated costs 鈥攕uch as research and development, production, and marketing 鈥� is vital for setting profitable prices. 
• Competition: Pricing should reflect a company鈥檚 market position relative to its competitors. By analyzing the pricing strategies of rivals, businesses can adjust their own prices to remain competitive. 
• Customer Value Perception: Pricing strategies must consider how customers perceive the value of a product or service. A strong understanding of customer preferences can help segment the market and tailor pricing to different groups. 
• Product & Service Lifecycle: As products and services evolve, pricing strategies need to be adjusted to reflect their lifecycle stage, ensuring the right balance between market penetration and profitability. 

Let鈥檚 now explore how these strategies can be applied in different industries, each facing its own unique challenges. 

Industry 乌鸦传媒: Applying monetization strategies across sectors 

Monetization strategies must be tailored to the specific needs of each industry. 乌鸦传媒 from the automotive, manufacturing, and life sciences sectors provide valuable lessons on how companies can successfully implement monetization strategies. 

• Automotive Industry: In the automotive sector, monetizing digital services through connected car data is a key focus. Subscription-based services, such as real-time navigation or diagnostics, provide recurring revenue streams while improving customer satisfaction. 
• Manufacturing Industry: Manufacturers are transitioning from hardware-centric to software-driven business models. By leveraging IoT (Internet of Things) and AI technologies, they can create new revenue streams from data-driven services and platform-based solutions.听
• Life Sciences Industry: Monetization in life sciences often involves the development of digital health platforms and the monetization of patient data. Outcome-based pricing models, which tie the cost of services to their success in improving patient health, offer an innovative approach to generating revenue. 

How can businesses ensure they are on the right path to mastering monetization? 

Moving forward: Strategies for navigating the monetization jungle 

To successfully navigate the complex terrain of monetization, businesses need a strategy that aligns with both their overall business goals and customer expectations. Achieving this requires action from two perspectives: 

1. The Business Perspective: Companies must start by conducting a thorough analysis of their cost structures and the market to determine competitive pricing. This involves understanding all cost components associated with delivering digital products and services and assessing how these costs can be distributed. Once the costs are clear, businesses can use competitor pricing analysis to find the best market position and adjust their prices accordingly, which in turn is how product monetization can be scoped before deploying any monetization strategy 
2. The Customer Perspective: Building strong customer relationships is essential to successful monetization. Understanding customer needs, preferences, and their willingness to pay is key to tailoring pricing strategies and maximizing profits. Companies should also leverage tools like customer surveys and willingness-to-pay analysis to refine their strategies.

Conclusion: Mastering the art of monetization

As companies continue to expand into digital products and services, monetization strategies must evolve to meet new demands. Whether through direct or indirect monetization, success depends on a company鈥檚 ability to align these strategies with its broader business goals, pricing, and customer relationships.

Mastering the art of monetization is essential for any business looking to thrive in today鈥檚 digital world. By overcoming the challenges that many companies face, businesses can unlock new revenue streams, enhance customer satisfaction, and ensure long-term growth and success.

The healthcare sector is undergoing transformative changes, driven by technological advancements, data ecosystems, evolving patient needs, and a growing emphasis on sustainability. The focus is on shifting care from hospitals to patients’ homes and enhancing quality of life with minimal disruptions.

Healthcare providers are embracing digital solutions and collaborative models to create more integrated, efficient, and patient-centered systems. At the same time, challenges such as an aging population and resource constraints are prompting the exploration of new care delivery models. Several emerging trends reflect these pressures, all underpinned by the critical role of data, which is key to optimizing healthcare delivery, improving patient outcomes, and creating a more connected, efficient, and sustainable system.

Summary and outlook: A digital, data-driven future

Digitization, and a deeper integration of data into everyday care are bringing rapid change to the healthcare sector. As healthcare systems adapt to evolving patient expectations and increasing demand for sustainability, the focus will shift toward enhancing accessibility and efficiency through digital tools.

Medical costs are projected to rise at a global average rate of in 2025, highlighting the urgent need for cost-effective solutions. Key trends will include the widespread adoption of digital home monitoring, AI-powered personalized treatments, the integration of behavioral and mental health services into primary care, and a stronger emphasis on prevention to reduce long-term healthcare costs.

As medical technologies continue to drive up costs, nearly of insurers report that new medical technologies are the top cost driver globally. This trend is particularly significant in the Americas () and Asia Pacific (). To mitigate these rising costs, there is a growing push to make healthcare more cost-efficient without compromising patient care. The integration of digital tools, such as telehealth, will play a crucial role in this transition. added telehealth and wellbeing services to their portfolios in 2024, an increase from in 2023, underscoring the growing importance of remote care in improving healthcare access and efficiency.

At the same time, demand for healthcare services is expected to rise significantly. With of insurers predicting higher or significantly higher medical trends globally, healthcare systems will need to adopt more sustainable, tech-driven care models to meet these demands. This is especially evident in regions like Asia Pacific, where of insurers anticipate increased medical trends.

These trends will converge this year to create a more connected, data-driven healthcare system that prioritizes both patients wellbeing and system sustainability. The widespread adoption of digital health tools, personalized AI treatments, and integrated mental health services will help address rising costs and increasing demand, making healthcare more accessible and efficient.

Advanced technologies, data ecosystems, and artificial intelligence (AI) in government are driving transformation in tax and customs. How can authorities in these two areas make best use of these developments to enhance customer experiences, ensure compliance, and protect against cybersecurity threats?

Tax and customs authorities have entered what鈥檚 known as the 鈥淩elationship Era,鈥� where they are redefining how they engage with customers. Inspired by the seamless digital experiences offered by the private sector, different customer groups now expect the same level of service when they鈥檙e paying their taxes. Tax authorities must rise to this challenge by providing frictionless taxpayer experiences that build trust in the tax system. This shift will be driven by a more interconnected tax system that places customers at its core.

The predicts that paying taxes will become an increasingly seamless experience that鈥檚 integrated into daily life. As we move forward, it is imperative for tax agencies to embrace change and innovate continuously to meet the growing demands of today鈥檚 taxpayers.

Five key trends shaping the future of tax and customs

The trends we explore here are closely intertwined and constantly evolving.

Responding to complexity and change

In conclusion, embracing advanced technologies, data analytics, and innovative strategies will better equip tax and customs authorities to navigate the complexities of the global trade environment and enhance customer satisfaction. We also expect these technologies, along with data ecosystems and AI, to help agencies ensure robust compliance and cybersecurity measures.

How will digitalization accelerate transformation in welfare services throughout 2025?

According to an OECD report, utilizing artificial intelligence (AI). However, despite significant strides towards digitalization, welfare organizations are not fully realizing the potential of AI to improve user-centricity, foster innovation through collaboration, and speed up accessibility advancements. They have also yet to make the best use of increasingly sophisticated AI鈥攐r what鈥檚 known as agentic AI. A focus on these areas underpins the key trends for welfare services in 2025. 

While the use of AI may have contributed to what the describes as a 50% increase in the delivery of social benefits, the fact remains that there is huge untapped potential for digital tools to help create a more inclusive, sustainable and effective welfare system for future generations.  Our work with welfare agencies worldwide has enabled us both to identify the trends shaping welfare services this year and to help organizations seize the opportunities these trends present, notably in advancing the use of AI.

2025 鈥� a year of maturity or uncertainty?

Will 2025 be the year of maturity for agentic AI in social welfare? Being a crucial touchpoint for public sector interactions, welfare organizations stand to benefit immensely from the potential of AI. However, the journey towards maturity involves strong executive support, robust technological infrastructure, a solid data foundation, a trained workforce, and a clear AI strategy.

Even with all of these building blocks in place, welfare organizations must also acknowledge鈥攁nd address鈥攖he associated challenges of AI. These include ethical considerations such as the potential for algorithmic bias, which might lead to unfair treatment of certain groups. Data protection is also vitally important: Clear governance and transparency are needed for all tools handling sensitive personal information. And organizations must consider how greater usage of AI can be compatible with their commitments on sustainability: .  

Finally, while agentic and other forms of AI can contribute much to the transformation of welfare services, they should not be regarded as the only solution for digitalization within the welfare sector. Our work with Chief Digital Officers in welfare organizations often reveals a need to prioritize other pressing issues 鈥� for example outdated legacy systems, and gaps in the mobile friendliness of some public sector websites. All of these have their place in an ongoing story of digital transformation, and all can contribute in different ways to helping agencies deliver more for less.

World leaders meet in Paris this week to discuss the future of AI. We need a shared vision for responsible use: both to maximize the benefits for society, and reduce AI鈥檚 potential for online harms.

While the emergence of advanced and highly capable artificial intelligence (AI) models and systems is likely to lead to huge benefits for society, it is also likely they will lead to large increases in online crime through the malicious misuse of AI tools, products and services, as well as potentially accidental harms (mistakes or unintended consequences from non-malicious actors). Of course, many different technologies can be exploited to cause harms as well as for their positive benefits, but often these harms are limited in their impact by requiring a high level of expertise, access to the technology, and money. In the case of generative AI however, these factors are much less constraining. We are now seeing the development of generative AI tools which are free or very cheap, very widely available online, and which have the potential to catalyse several forms of harm, including financial crime, cyber-attacks, and online targeting of individuals or groups (cyberstalking, harassment, or political disinformation). The barrier to entry for committing such crimes is now merely access to a laptop and wifi.

AI鈥檚 specific strengths in handling unstructured and variable datasets, data pattern recognition, replication at scale, tailoring content to the individual level, and predictive analytics, which are already being used by businesses to predict demand, anticipate user trends and identify gaps in offerings, make it uniquely valuable for all sorts of tasks, and also uniquely capable at being used as a tool to commit online crime. AI can facilitate the development of off-the-shelf 鈥渃rime-as-a-service鈥� products which vastly reduce the 鈥渂arriers to entry鈥�, for those so inclined. AI software can be used to conceal the perpetrator鈥檚 identity and location, making it hard to investigate. AI models can also themselves be targets for crimes such as hacking, through prompt injection for instance, and data manipulation or poisoning, causing them to make mistakes, or to react in specific ways given the right trigger. These types of manipulation would be of particular concern in areas of critical national infrastructure and autonomous weapons systems.

New potential for abuses of trust

Some of the most concerning types of harm which generative AI tools, such as social 鈥渂ots鈥�, may facilitate are crimes of persuasion and influence: exploiting an individual鈥檚 psychology or personal circumstances or actually affecting someone鈥檚 mental state to convince them to act in a way they may not have done otherwise, possibly using deepfakes (audio and visual media purporting to show a person or event that in reality never occurred), misinformation and disinformation. These effects could be exploited for deception, phishing, radicalisation and encouragement of social unrest. Early experiences show that our natural tendency to anthropomorphize means that people can become emotionally attached to AI-generated bots, divulge personal information to them, and that they may create an echo-chamber effect which normalizes harmful behaviours such as sex crimes. Even more subtly, the development and awareness of AI products generally creates an environment where people may expect or imply AI-related crimes, even if no AI was involved 鈥� for example threatening to use specific tools, making someone believe a certain effect was possible using AI even if it is not feasible, or making people believe a genuine video or photo was faked. It is also important to note that this type of criminality need not only be targeted at an individual. The capabilities of generative AI can be implemented just as easily for a large target audience, while still being personalized to the individual user to encourage their engagement.

There is also the potential for emergent criminal behaviours, as AI agents become more sophisticated and interact in increasingly complex ways. They may autonomously commit crimes going beyond the user鈥檚 initial expectations or moral compass. An AI system has no innate understanding of ethics, pain, truth, or compassion, and is without human limitations of strength, tiredness and pace. It may propose or take actions which are unacceptable, too complex to comprehend, or too fast to prevent. As AI agents start to be entrusted with acting autonomously on our behalf, we will need to incorporate increasing levels of safeguards to prevent them from over-reaching; but because deploying bots is cheaper and easier than implementing effective governance to control their actions, it is likely to be very difficult to monitor and mitigate all risks and impacts.

How can governments and business leaders take action?

As senior leaders gather in Paris to debate AI safety standards 鈥� amid a complex multi-national arms-race of AI development 鈥� they need urgently to discuss and agree measures to prevent AI-enabled harms occurring. These measures could be defined and coordinated across state boundaries, implementing governance in a similar way to global civil aviation, which is effectively governed by the International Civil Aviation Organization (ICAO), with state-level measures that ensure adherence to these global standards. The United Nations is well-positioned and prepared to coordinate the required oversight. Such measures need to be thought of in three layers:

1. Technical measures to prevent harm: removing any datasets from training data which contain harmful content; vetting datasets; fine-tuning models using reinforcement learning techniques to avoid harmful outputs; adversarial testing and evaluations; stress-testing to identify potential vulnerabilities such as prompt injection; developing explainable AI models; and guardrails to prevent certain types of output generation.
2. Organizational approaches to deter harm: minimum safety standards; terms and conditions; user verification; content moderation and screening (including AI tools to automate content identification and removal); watermarking, labelling and tracking metadata; tagging verifiable data; correcting or flagging fake news; user behavior analysis; blocking, alerts and reporting mechanisms; education, training and awareness; restricted access; governance policies including ethics; and developing appropriate and proportionate law, guidance and regulation.
3. Law enforcement responses to harm: monitoring and intelligence-gathering; detection tools (including AI tools to automate detection); investigatory processes; agreements with technology companies to access evidence; identifying high-risk individuals and communities; accessing technical skillsets; increasing capacity to address the growth in online harms; deploying counter-influence AI tools to mitigate the effects (for example redirecting to counter-radicalization resources); and working with technology companies to respond to emerging criminal behaviors.

In conclusion, generative AI tools can provide great benefits; but could also lead to exponential increases in online harms. Technology companies, governments, and law enforcement agencies are working together to anticipate, understand and prevent such harms occurring, but ultimately everyone will need to be conscious and responsible in their use of AI.

Even though digital commerce has significant potential for high profitability, there are still a variety of e-tailers who have been struggling with strong headwinds. For most e-tailers, the primary strategy is to increase the profitability at scale. However, this strategy comes with several challenges.

• Buying more traffic can lower the quality of this traffic and reduce conversion rates.
• Competitors are driving the same thesis and will compete for the same clients.
• The recent economic climate has slowed down growth in most sectors.
• Perhaps most importantly, many costs will not decrease with scale but rather increase. Pick & pack, fulfillment, reverse logistics, and return costs will likely grow at the same pace or more than turnover growth.

If selling more isn鈥檛 an option, selling smarter is essential! You don’t make a profit on the first items you sell but on the last ones. Therefore, you must optimize and sell your entire inventory from the most cost-effective location with as few price reductions as possible.

When speaking with clients, many don鈥檛 initially see any major issues with Order management. However, after exploring capabilities and alternative ways of working, we often find numerous opportunities for optimizations within distributed order management. In this blog, we will share our perspective on how to:

• Take control of your sellable inventory
• Optimize sourcing & fulfillment.
• Improve and drive CX journeys.
• Increase Sustainability without interfering with the margin.

Take control of your sellable inventory

To drive successful digital commerce, you need to ensure you know what products you can sell. The importance of this can vary based on the cost of products, seasonal products/fashion cycles, return rates, the number of fulfillment locations, etc., but it will affect all e-tailers in some way.

It sounds simple, but can your company calculate what is available to sell if you include e-com inventory, warehouse inventory, store inventory, in-transit replenishment, returns (planned, in transit, in warehouse processes), and products in basket/e-basket?

What is then your available-to-sell stock balance?

Furthermore, the system must calculate an accurate sellable inventory in near real-time and simultaneously support several different channels with available-to-promise (ATP) checks in microseconds. On top of this, you must add some intelligence to the inventory with certain business rules. For example, how can you override the safety stock rules if certain criteria are met to sell the very last product at hand in stores? By increasing the accuracy of the stock levels, you can significantly lower your safety stock and hence sell more of your inventory without jeopardizing the risk of overselling.

Once you have better control of your inventory, you will significantly reduce underselling/overselling, thereby improving the NPS and reducing customer service calls. Make the sellable inventory a fact-based business decision rather than an assumed number.

Optimize Sourcing & Fulfillment

Once we know where we have these products, we can start to evaluate the most cost-effective way to fulfill these orders or allow customers to pick them up from stores or surplus inventory locations. How can you use these insights to optimize sales across all channels? Apply rules around your inventory so that you can take calculated risks in the sellable inventory balances you calculate and thus sell the last items, which often means 100% profit margins as the costs have already been carried by the first products.

Through transparent ATP (available-to-promise) controls, we can guide customers toward the most cost-effective, and sustainable fulfillment options. Split shipments can be managed and aggregated to achieve a single delivery to the customer. Furthermore, the pandemic and the subsequent component shortage and shaky supply chain have shown that flexibility is required in the fulfillment process.

When sourcing fails and we have oversold an item, ensure the best action is automated. For example, prioritize orders to critical channels like Amazon to avoid penalties for unfulfilled promises.

Improve and drive CX Journeys

There is a misconception that OMS processes start when an order is placed. Many areas of pre-purchase can improve significantly by improving your order management processes. Transparency will create opportunities both for better customer experience as well as guidance for improved margins and reduced cost of fulfillment.

There are several ways to influence fulfillment costs without jeopardizing the customer’s journey. For example, can we influence customers to place orders that can be delivered from the same distribution center, thereby reducing the number of deliveries? Can we drive different strategies for membership/membership levels, i.e., only offer free shipment/returns for club members or member levels?

Historically there has been a fear of touching the free delivery/return policies since that will have a significant impact on the conversion rate. Recently, more and more e-tailers have changed policies, but our investigation shows that there is still a lot to do in this area to offer more personalized and optimized fulfillment and return strategies.

Increased Sustainability without interfering with the margin

Sustainability comes in many shapes and forms. Optimizing the order management processes can significantly improve various areas that impact sustainability. Examples include:

• Reducing CO2 emissions due to optimized fulfillment and giving customers the option to choose delivery not only by speed but also by cost as well as emission footprint.
• Using these optimizations to reduce waste due to fewer returns and consolidated orders.

Adapting to new models in sourcing and fulfillment should be a configuration/automation of processes rather than a project.

乌鸦传媒鈥檚 Expertise: At 乌鸦传媒, we bring extensive experience in helping clients in retail, grocery, and consumer product sectors achieve their digital commerce goals with tailored strategies and solutions.

They come from different worlds. Now they鈥檙e meeting and sparks are flying. 

Artificial Intelligence (AI) has evolved through two distinct pathways in parallel over the decades: deterministic AI and statistical AI. These two approaches have historically shaped how AI systems learn, reason, and make decisions, each bringing its own strengths and limitations. One offers clarity and structure, while the other allows patterns to be learned from data and applied into the future. Each approach has a fundamentally different answer to the way AI systems learn, reason, and make decisions. Now, a new field of AI research aims to fuse these two forms of AI into one. If it works, things are about to get very interesting. 

Deterministic AI: the early era 

In the early days of AI research, the limitations were often computing power and memory. This led to AI being dominated by deterministic approaches, specifically symbolic AI. Symbolic AI uses formal logic and explicit rules. These rules-based systems allowed decisions to be made automatically based on intelligence; the intelligence was learned but was simple in nature and needed expert input. 

The most famous of these early systems were expert systems in the 1970s and 1980s. These systems used if-then rules to mimic human decision-making, offering clear, explainable reasoning. For instance, for a given input in a factory setting, a set of actions are defined and taken based on the input. Deterministic AI provides clarity and precision, as every rule is explicitly defined. It is therefore very easy to understand how the system is working and see why the outcome was achieved. The challenge with these systems is their lack of flexibility and dependence on pre-defined knowledge. The answer to this problem? statistical AI. 

The rise of statistical AI鈥� 

Statistical AI had always existed, but it was not easy to use. Statistics were often done statically, meaning the analyses that businesses relied on were based on a snapshot of the past. The idea is that if you use statistics, you are understanding the past to inform the future (often through policy changes) rather than machine learning, which is predicting the future from the past and acting accordingly. The emergence of fast computing changed the game regarding statistical AI. Now people could build and experiment with more complex models and much larger datasets. AI began to move into areas of increasing complexity and uncertainty. 

Another noteworthy advancement was the increasing availability of machine learning libraries like ML++. These packages meant that machine learning models did not have to be built from scratch every time. A developer could now build a model on data and check whether machine learning could solve a business problem faster than ever before. In this way, AI proof of concepts emerged, allowing for testing鈥痶he feasibility of an idea quickly. 

However, the move away from expert knowledge being explicitly programmed into AI came with new challenges, especially around the explainability of the AI models built. This has now developed into an entire field of data science, and the more complex the model, the harder it is to understand (i.e., to explain the decision based on logical reasoning in a way the receiver can understand) and explain what the model is doing (i.e., explain the way the decision traces through the model, based on the functionality of the algorithm). The term 鈥溾�� came into use, highlighting that we don鈥檛 know what happens in the 鈥渂lack box鈥� and therefore cannot inspect it.  

In many cases this might not be problematic, and a few ways developed to understand the models and trace back on the predictions they have made. In some cases (like with computer vision models), the visualization of model activation (i.e., which nodes in the model are triggered most or which areas of the image are most critical to the decision-making) can provide enough 鈥渆xplanation鈥� for us to decide on the quality of that model for the task at hand. 

The birth of hybrid AI: combining deterministic and statistical approaches 

As AI matured, the need to combine the precision and interpretability of deterministic AI with the flexibility and adaptability of statistical AI became clear. Hybrid AI systems (Neurosymbolic AI) have emerged that could handle complex tasks with both structure and uncertainty. 

Reasoning and logical inference: The essence of deterministic AI 

One mechanism of structuring data for use by deterministic AI is Knowledge Graphs (KGs),which can represent entities (such as people, places, or concepts) and then model these relationships in a structured, graph-based format. These graphs embody the symbolic AI tradition by providing a deterministic structure that can be reasoned through logically. For example, if John lives in London and Brian lives with John, then the fact that Brian must also live in London can be derived, while not being explicitly available through the data. Checking model quality automatically through logical constraints, inferring new and hidden knowledge, but also finding inconsistencies and flaws of information models, it is all within reach. The ability of knowledge graphs to handle complex and vast amounts of knowledge, in a structured way, can be very valuable. We can also look to include elements of statistical AI into deterministic AI, and in this way create hybrid AI.鈥赌�&苍产蝉辫;

One example of this is a decision tree which can be used in conjunction with a knowledge graph to to add new elements into the graph, based on the statistical likelihood of a link being there. While a knowledge graph flags individual anomalies, the decision tree evaluates them collectively, enabling a more nuanced analysis. For example in a transaction for a credit card company, we can combine several factors such as the transaction amount being abnormally high, in an unusual location and into a new bank account. The probability of these individually can be 20%, 30% & 10% respectively but combined the likelihood of fraud is now 49.6%, which is greater than any of the individual factors alone. By leveraging the structured relationships and context in the knowledge graph, the decision tree identifies patterns that might go unnoticed, such as the compounded risk of anomalies occurring together. This integration enhances the accuracy of fraud detection by applying reasoning to the interconnected data in the knowledge graph, prioritizing cases for investigation with greater precision. The hybrid approach provides interpretability and scalability, as investigators can trace decisions back to both the individual features in the knowledge graph and the decision tree’s logic, improving transparency and trust. These methods can be immensely powerful and showcase hybrid AI in action.鈥赌�&苍产蝉辫;

Generative AI: The latest buzz in statistical AI. 

Generative AI (Gen AI), which includes models like , Meta鈥檚 Llama 3, Mistral AI, etc., is the latest advancement in statistical AI and is the outcome of further advancements in computing power and availability. Huge models are built at great cost to be able to handle many different tasks, having learned on a huge corpus of information. They can generate text, images, and other forms of content. 

However, being statistical in nature, these models can get things wrong, in machine learning, a model cannot predict the correct outcome 100% of the time. In Gen AI we have these incorrect predictions as well but instead of an incorrect prediction of churn for example, the incorrect prediction of a word leads to the next word being incorrect and entirely plausible sentences being created from an initial incorrect prediction. As you can imagine this can give mean a passage of incorrect text is created in a snowball effect from one incorrect prediction. Therefore, incorrect predictions can be more sinister in Gen AI than in other statistical AI methods. In Gen AI we call these effects hallucinations. To explore this subject further, consider reading this blog we previously published from the AI Lab. The chance of error means they are not as reliable as deterministic AI. But deterministic AI can be coupled with Gen AI to make them more reliable. Here we see the emergence of for instance, where graph technology is used with Gen AI to improve information retrieval from free text questions. This is another example of hybrid AI and is one being .鈥赌�&苍产蝉辫;

An LLM-generated knowledge graph built using GPT-4 Turbo:  

Conclusion: The future of hybrid AI鈥� 

The journey of AI can be told as the parallel journeys of deterministic AI and statistical AI. However, the promise of AI throughout this whole journey has always been in the bridging of these two worlds, where AI not only holds and uses knowledge with reasoning but learns patterns from that data and applies those learnings to the real-world knowledge it has. The ability to abstract, reason, plan, predict and explain comes within reach, thanks to the combination of various types of AI models and methods of knowledge representation. This is where AI is heading, solving the challenges in both statistical and deterministic AI. The future of AI is hybrid AI enabling more intelligent, robust, and trustworthy systems.鈥�&苍产蝉辫;

About AI Futures Lab 

We are the AI Futures Lab, expert partners that help you confidently visualize and pursue a better, sustainable, and trusted AI-enabled future. We do this by understanding, pre-empting, and harnessing emerging trends and technologies. Ultimately, making possible trustworthy and reliable AI that triggers your imagination, enhances your productivity, and increases your efficiency. We will support you with the business challenges you know about and the emerging ones you will need to know to succeed in the future. 

 We create blogs, like this one, Points of View (POVs), and demos around these focus areas to start a conversation about how AI will impact us in the future. For more information on the AI Lab and more of the work we have done, visit this page: AI Lab.鈥�&苍产蝉辫;