R&D Value and Challenges

In early 2015 when Winterhawk was first born, we were at heart a Business Consultancy. We quickly realized that to differentiate ourselves, we needed to innovate.

We specialize in SAP (System Applications and Products in Data Processing) a product suite that is used by more than 1/2 million companies worldwide. The SAP Partner Eco System has 25,000 partners. How does as small UK Headquartered development and consulting company get noticed when there are 25,000 other partners of SAP, that includes $ Billion competitors like Accenture, Deloitte and IBM?

Answer : Investment in Research, Development and Innovation. Inherently within that, taking calculated risks.

Fact 1 : The rapid pace of digitization causes SAP customer to seek efficiencies, they want to achieve customer loyalty, they need to increase their competitiveness and this requires them to determine strategies accordingly. Whether it’s better managing human-machine interactions, advanced visualizations, enhanced user experiences that enables fact-based decision-making.

Fact 2 : The world is changing. We have wars, we have health emergencies, we have supply chain challenges, we have a growing threat to data (Cyber Attacks, Fraud..) and at the same time, the bar is going up and up, as Governments and Industry bodies create new Regulations, and they then continually change and evolve. Add to that “bad actors”, individuals and companies that you should not do business with presents a separate need to continually access and be aware of Sanctions.

As it shows underlined above in red (from our company home page) – one of the things our R&D team does is create “Packages” through Innovation within the many SAP Platforms.

It should first be understood that SAP is not one single system. SAP has spawned to be many systems – literally 100’s and 100’s. The ecosystem is vast.

As an example – whilst Core Financials for a customer might reside in SAP S/4 HANA or SAP ECC, a customer many likely have SAP Ariba (for Procurement), SAP Fieldglass (for Vendor Management), SAP Signavio (Analytics), SAP Risk and Assurance Management (For Business Controls) – the list is almost endless.

Some of the challenges for Winterhawk ? 

• Multiple Systems = Multiple Integration Paths – with that comes greater risk and uncertainty.
• An ever changing technology = SAP change their platforms, their architecture, they add and remove products – all of the time!
• New features being added (A.I, ML, RPA) = exciting but also prone to failure, and significantly higher commercial risks. Quite often a technology simply isn’t ready and mature enough.
• Data = It can be surprisingly hard to find, often customers themselves do not know where all their data is held. If you do find it, can it be used, can it be interpreted, translated and used.

Above is a screenshot from our front page – as mentioned earlier, topics of RPA, ML and A.I are “hot topics” and a business like SAP greatly encourages it’s partners to Innovate and Develop in those areas – so that they can stay ahead of (their) competition – Oracle, Microsoft, Infor, Workday etc.

If Winterhawk did not develop and innovate, what would happen? 

1). We would be less appealing to customers (as our competition is)

2). We would be of less value to our customers – SAP by their own admission do not want to create innovation within their systems, they see that as the role of the partners.

3). We would have less traction with SAP themselves – who naturally follow the businesses that are driving innovation.

Innovation & Development – Winterhawk, the early Innovators

This slide was presented by SAP at an event in Amsterdam in 2019 – showing our, then, early vision for the use of RPA.

As it mentions above, a screenshot taken from our Innovation page – Regulations are a huge challenge for customers.

As a business, and let’s say you are a Mining Company, and you operate Globally, how do you know all the Regulations that you need to adhere to in 100’s of different jurisdictions and geographies? Even if you did work that out at a point in time, how do you ensure you are continually up to date with new Regulations or changes to ones you already do know about? How? It’s really not going to be easy.

Perhaps slightly easier if you are a $ billion organization who can afford a team of lawyers, accountants, risk specialists and administrators to scour the internet, but even then, what might you miss? However 90% of the 500,000 customers using SAP are not in that category, they are mid-market customers who do not have near unlimited financial resources.

and …. that really is just the start, fine, you have the Regulation – Now what ?

If I’m the customer, how do I interpret a 500 page document, written by lawyers, and turn that into a set of Business Controls and Testing that my company can understand and operate?

Answer :

A). You need specialists who are skilled in SAP

B). You need specialists who have functional knowledge across all aspects of business (Order to Cash, Finance, Procurement, Manufacturing etc)

C). You need specialists who understand Governance, Risk and Compliance, Internal Controls and Security and the wider requirements and expectations of Internal and External Audit

D). You need specialists  who understand the SAP Technology stack and who can drive development to achieve highly complex and challenging outcomes which by nature of the work, need to adhere to regulations that carry significant fines, penalties and publicity for the customer should they come up short.

So what makes Winterhawk special? What makes our team special? 

Our leadership team have spent their careers working “client side” rather than being career consultants for companies like Accenture, Deloitte, IBM – they understand the challenges of being the client – working with budgets, budget cycles and funding rounds, organizational changes, image and reputational factors, brand risks, C-level expectations, internal and external audit expectations. They have also spent entire careers in SAP. They have built up trusted relationships with SAP. They have previously developed SAP products. They talk at SAP events around the world (sought after speakers too). They have reputations for Innovation and being Thought Leaders in what is a very niche area – an area where Winterhawk is the only globally trading company in the world, who is focused solely on SAP and solely on SAP Risk solutions.

Q&A – Development, Research & Innovation

When did you start and when will you finish?

The thinking behind it goes back to the inception of Winterhawk in 2015. We see this as a continuous R&D because SAP is continuing to evolve. More and more focus is being put on partner Innovation. We simply cannot stand still. Regulations dont stand still – new ones come out all of the time – and that requires Research – and while we have not talked about it here – and huge amount of R&D time goes into research (so in answer – it won’t finish, it will only grow)

Is it still unresolved?

We are dealing with new technology (SAP evolving to enable partners to attempt to build A.I, ML, RPA, add Predictive Analytic Logic), it takes a lot of time, effort and it is trial and error.

As recent context, we spent 3 days in a European City with partners from all over the world, 20+ people in a room, working on a single A.I example, which at the end of 3 days, we determined was “unsafe to proceed” with (20 people, x10 hours x3 days = 600 hours / 75 days).

Unsafe because we could not see how reliable the results would be. As we could only achieve “85%” consistently reliable results which fell significantly below our minimum threshold. In the regulatory world, in a world where our clients can and do get fined 100s of millions of dollars, 85% is not good enough. A.I is so new, all the more so in SAP – but you will have a huge technological advantage if you get it right and are early to market, but it’s high risk and high effort.

Early R&D effort

When was the gap in knowledge identified?

We knew we needed to innovate from the start of Winterhawk.

We knew there were huge gaps in SAP products that Partners could add value in.

We had to build a team capable of Developing solutions / packages. A significant financial outlay.

We had to build internal systems and infrastructure. Another significant outlay.

The market gap – the need however was clear.

SAP on a almost daily basis are opening up new ways to work with their technology stack – which is fantastic, but it means it never stands still, and the learning curve is continuous.

SAP was born in 1972 – its very first low level A.I started to appear in around 2015 but it’s only in the last 2-3 years its become something that partners could develop upon, and in the Risk space, it’s bleeding edge.

Regulation Challenges

Real life examples:

• Failure to comply with sanctions or obtain the correct licence, in the case of export controls, can lead to significant fines. In 2022 alone, US Office of Foreign Assets Control (OFAC) ’s enforcement penalties hit a record of $1.5 billion. View this guide to U.S. OFAC sanctions for more detail on whether your business is in scope and how not to breach OFAC sanctions.
• Screening for sanctions risk is a mandatory requirement for regulated entities across all sectors and is a critical step in the know your customer or KYC process. Moreover, penalties for breaching sanctions can be severe. In 2021 the US Treasury Office of Foreign Assets Control meted out fines amounting to over US $20 million, 10 of the 20 penalties were levied at corporates. In 2020 the penalties amounted to US $23 million and only two were financial institutions
• AML link – in 2020, HSBC’s share price plummeted due to the allegations of money laundering, which resulted in a fine of £63.9m by the UK’s financial regulator.

The Typical Objectives of Regulations

Development

Winterhawk began collaboration with SAP in 2019 with an aim to help organizations adopt A.I, Machine Learning and Robotic advancements across their core business platforms, applications and processes (SAP ERP and specifically SAP BTP a new SAP platform) related to the “GRC” stack of products. Our aim was (and still is) to create AI-enabled solutions and use cases that can enhance an organization’s investment in SAP technology by improving business performance and employee productivity. Winterhawk will expand its capabilities, enhancing the way users interact with SAP applications.

Vision/Goals

• Improve Governance and Regulatory Assurance
• Streamline operations
• Enhance customer experience
• Drive sustainable growth
• Build upon Winterhawks experience, helping organizations transform their businesses with SAP software
• Embed A.I into industry-specific content (SAP Risk & Assurance Management and SAP Process Control)
• Leverage A.I to help customers accelerate from legacy SAP Systems
• Auto generating of code, documentation, and test scripts
• Built chatbot features
• Enhance Reporting and Analysis leveraging A.I and ML

Winterhawk Development Cycle – High level (more at the bottom)

Phase I – Ongoing

An ongoing journey to create Unique – first to market – Industry Vertical Packages for SAP Risk Solutions

Note : As new regulations, frameworks and standards are created and updated, so to will all of these developments needs continuous work – Phase I will never “complete” – and should be understood as an ongoing initiative, this does not stop.

Phase II – Ongoing

Expansion of phase I to further include- creating SAP apps – leveraging next generation A.I capabilities in SAP – leveraging predictive analytics – leveraging Data Tags – Creating Sanction / Watch Lists for SAP clients, integrated into SAP S/4 and SAP Risk & Assurance Management  – adding Sustainability regulations and controls

Note : As new regulations, frameworks and standards are created there is no end date to Phase II, it will be a continuous EXPAND – and again, should be understood as an ongoing initiative, this does not stop.

NOTE : Phase I and Phase II happen in parallel. 

Phase III

Expansion of phases I and II to further include – if technically feasible – wider predictive analytics for Sustainability Controls, leveraging A.I, ML, RPA for more SAP Solutions (inc SAP Signavio, SAP WalkMe)

A.I Challenges

Oh…… we would write an entire book on the challenges! We as humans build our world models based on our experiences and senses. These models grow more complex as we learn. However, some decision-making based on these models can be simplified and implemented into computer science. We’ve done this before with rules and models in simple programming.

But with AI, we’re dealing with software learning from experiences, or in our case, from data. The models we develop through these algorithms, will likely only capture a limited complexity of the models that we have in mind.

The models we use in business processes may not be as complex as human thought, but they can process vast amounts of data that we can’t retain in our heads. They can effectively analyze and correlate this data, using algorithms for tasks like clustering or projection. This allows us to understand situations, such as a customer’s status or a business process, based on a large volume of diverse data.

The models quickly process this information and provide valuable insights, helping us identify problems or decide on the appropriate action at a given time. They function like a brain that pools all past experiences.

• Reliability is crucial for AI in SAP.  There have been many instances where AI models provide inaccurate or outdated information.
• The quality of results largely depends on the data input;
• If your data is very heterogeneous, it may not be suitable for certain algorithms. This could lead to an inability to capture outliers. Therefore, it’s essential to understand how the data should look, how much data you need, and where the quality signals lie. Selecting the best data for your specific organization or even for a part of your organization. This will best reflect the occurrence of certain analyses in the system.
• Then you have data security. We don’t want to share our intellectual property externally, and many large language models are situated in an external cloud. We need to ensure that our data is converted, encrypted, and cleansed in a way that it can be used in these models without providing any access to the source data.
• Currently, SAP S/4HANA Cloud offers over 25 use cases based on AI technology. Many of these are built on our SAP BTP platform. Winterhawk have decided to focus on solutions within this BTP architecture and specifically in Fraud Detection, which inherently is not an exact science and comes with unstructured data. We’re enhancing these capabilities with generative AI, which can seek to identify the structural elements of an unstructured document. It is a constantly evolving landscape, new, cutting edge, and will lead to (many) failures as well as, we hope, successes.
• A.I calls for a meticulous examination of code validation intricacies

Machine Learning

Code Validation – the tasks of checking and validating code present substantial challenges.

Example

Thousands of lines of code are generated in seconds, each infused with complex logic and algorithms. Can we review this machine-generated code with confidence? Can we? To mitigate, we need;

• Clear annotations
• Test-driven design
• Modularization by breaking down AI-generated code into manageable parts
• Structural tests
• Scan-tests
• Integrating built-in self-tests

Likely Points of Failure

Developing an A.I project and building an A.I model is experimental in nature and may require a long trial-and-error process. AI models try to solve probabilistic business problems which means the outcomes may not be the same for each use case. Once again, Data is the key resource of every AI project. Working with outdated, insufficient, or biased data can lead to garbage-in-garbage-out situations, failure of the project, and wasting resources.

A Snag List of Challenges when Developing A.I in SAP < a beginners guide !>

Complex Integration Processes

AI integration into SAP projects often involves complex processes due to the intricate nature of both technologies. SAP systems are known for their robustness and complexity, and incorporating AI functionalities seamlessly can be challenging. Ensuring that AI aligns with existing SAP architectures without disrupting core functionalities poses a significant technical integration hurdle.

SAP System Robustness

SAP systems are renowned for their robustness and complexity, designed to manage and streamline intricate business processes. Integrating AI into such environments necessitates careful consideration of SAP’s unique architecture. AI solutions need to complement SAP’s existing modules and functions without causing conflicts or compromising system stability.

Alignment with SAP Architecture

The diversity of SAP landscapes across organizations poses a challenge in creating AI solutions that align seamlessly with varying SAP architectures. Different SAP modules, versions, and customizations may exist within an organization, requiring AI integrations to be adaptable and compatible across this diversity. Ensuring that the AI seamlessly fits into these environments is crucial for successful integration.

Data Structures and Models

SAP projects deal with extensive datasets structured in a specific manner to support business processes. AI integration demands a deep understanding of these data structures. Creating AI models that effectively leverage SAP data without causing disruptions or inconsistencies requires meticulous data mapping and preprocessing. Addressing the uniqueness of SAP data models becomes a critical aspect (and risk) in any technical integration process.

Real-time Processing Requirements

Many SAP projects require real-time data processing to support dynamic business operations (examples SAP Business Integrity Screening – used to detect Fraud, and Enterprise Threat Detection, used to prevent cyber attacks – both leverage predictive analytic models). Integrating AI functionalities in real-time scenarios adds a layer of complexity. Ensuring that AI algorithms operate seamlessly within the SAP real-time processing framework demands optimization without compromising responsiveness.

Customization Challenges

SAP projects often involve extensive customization to meet specific business requirements. The challenge in AI integration lies in accommodating these customizations. AI solutions must be flexible and adaptable to accommodate various customization scenarios without necessitating extensive modifications to existing SAP technical configurations.

User Interfaces and Experience

Integrating AI into SAP projects involves considerations not only at the backend but also at the frontend. Designing user interfaces that effectively incorporate AI functionalities while maintaining a cohesive and user-friendly experience is another challenge. Ensuring that users can seamlessly interact with AI features within SAP interfaces requires a careful balance between technical functionality & simplicity.

Integration Testing

Rigorous testing is a crucial aspect of AI integration in SAP projects. Ensuring that AI modules work harmoniously with existing SAP functionalities and that the integrated system performs reliably under various scenarios demands comprehensive integration testing. This involves validating data flows, system responses, and overall system performance.

Data Compatibility and Quality

The success of AI in SAP projects heavily relies on data quality and compatibility. SAP projects deal with vast amounts of structured and unstructured data and ensuring that this data is AI-ready involves rigorous data preprocessing. Issues such as data silos, inconsistent data formats, and data quality concerns can hinder the effectiveness of AI applications within SAP environments.

In the realm of AI integration into SAP projects, one of the critical challenges revolves around ensuring data compatibility and maintaining high data quality.

Heterogeneous Data Sources

SAP projects often involve the integration of data from various sources (SAP and non-SAP), including different SAP modules, external databases, and legacy systems. These data sources may employ diverse data formats, structures, and standards. Ensuring compatibility among these heterogeneous sources is a significant technical challenge. Data must be harmonized and transformed to a standardized format that AI algorithms can effectively process.

Data Preprocessing and Cleansing

Raw data from SAP systems may contain inconsistencies, errors, and missing values. Prior to AI integration, extensive preprocessing and cleansing are necessary. This involves identifying and rectifying data anomalies, handling missing data appropriately, and ensuring that the input data for AI models is of high quality. The effectiveness of AI applications is directly impacted by the cleanliness and accuracy of the underlying data.

Semantic Understanding of SAP Data

SAP data models are inherently complex, often requiring a deep semantic understanding. AI algorithms need to comprehend the intricate relationships between different data entities within SAP systems. Achieving this semantic understanding involves mapping data elements, recognizing dependencies, and creating AI models that can effectively navigate the nuanced structure of SAP data.

Temporal and Sequential Data Considerations

SAP projects frequently involve temporal and sequential data, where the order and timing of events are crucial. Incorporating such data into AI models requires specialized approaches. Ensuring that AI algorithms can effectively capture and utilize temporal dependencies within SAP data is essential for applications such as predictive maintenance, forecasting, and trend analysis.

Master Data Management

Master data, which includes information about customers, products, and business partners, is a cornerstone of SAP projects. Ensuring the consistency and quality of master data is paramount for accurate AI predictions and insights. Challenges arise in maintaining master data integrity, especially when multiple systems or departments contribute to its creation and modification.

Integration with External Data

AI applications within SAP projects may need to leverage external data sources for comprehensive analysis. Integrating and synchronizing this external data with internal SAP data adds another layer of complexity. Ensuring seamless compatibility between SAP and external data formats, APIs, and security protocols is crucial for holistic AI-driven decision-making.

Data Security and Privacy

SAP projects often involve sensitive business data, and ensuring data security and privacy is of utmost importance. Integrating AI without compromising data security involves implementing robust encryption, access controls, and compliance with regulatory requirements. Balancing the need for AI-driven insights with stringent data protection measures is a continuous challenge.

Adaptability to Data Changes

SAP data structures may undergo changes due to system upgrades, customizations, or evolving business requirements. AI models need to be adaptable to such changes without requiring extensive retraining or reconfiguration. Ensuring that the AI integration remains resilient to data schema modifications is crucial for long-term sustainability.

Change Management

AI introduces changes in how users interact with SAP systems. Managing this change effectively requires a comprehensive change management strategy. Users may resist alterations to their familiar workflows, and addressing these concerns through clear communication and support mechanisms is essential for fostering acceptance.

Version Compatibility

SAP systems are subject to updates and version changes. AI integrations need to remain compatible with evolving SAP versions. Ensuring that AI models and algorithms can seamlessly transition across different SAP versions without causing disruptions or requiring extensive reconfiguration is a continuous technical and business challenge.

Transparency in Decision-Making

AI algorithms can sometimes operate as black boxes, making it challenging for users to understand the decision-making processes. Ensuring transparency in AI-driven decisions and providing insights into how AI contributes to outcomes enhances user trust. This transparency is crucial for gaining user acceptance of AI recommendations or automation within SAP projects.

User Involvement in AI Implementation

Involving end-users in the AI integration process fosters a sense of ownership and inclusion. Soliciting user feedback, addressing concerns, and incorporating user insights into the AI implementation strategy can significantly contribute to user acceptance. Users who feel valued and involved are more likely to embrace AI changes.

Tailored User Interfaces

AI often introduces new features and functionalities that necessitate adjustments to user interfaces. Designing interfaces that are intuitive and user-friendly is crucial. User acceptance is enhanced when the AI capabilities seamlessly integrate into the existing SAP interfaces, minimizing the need for extensive retraining.

Providing Use Cases and Success Stories

Sharing real-world use cases and success stories helps users understand the practical benefits of AI in SAP projects. Demonstrating how AI contributes to improved efficiency, decision-making, and overall project success instills confidence and encourages user acceptance.

Comprehensive Training Programs

Effective training programs play a pivotal role in user acceptance. Users need to be equipped with the knowledge and skills required to leverage AI functionalities within SAP systems. Training should cover not only the technical aspects but also the practical applications and benefits of AI in their specific roles.

Parallel Processing

Parallel processing capabilities contribute significantly to both scalability and performance optimization. AI algorithms that can be parallelized efficiently can take advantage of multi-core processors and distributed computing environments, enhancing overall system performance and scalability within SAP projects.

Model Complexity Considerations

AI models within SAP projects vary in complexity based on the tasks they perform. Striking the right balance between model complexity and performance is crucial. Complexity should be tailored to the specific requirements of SAP applications, ensuring that the models can scale without compromising performance.

Adaptability to Changing Workloads

Scalability involves accommodating variations in workloads. AI solutions within SAP projects should be built to be adaptable to changing demands, scaling up or down based on fluctuations in data processing requirements. This adaptability ensures optimal resource utilization and responsiveness to evolving business needs.

Caching and Preprocessing Strategies

Performance optimization often involves the strategic use of caching and preprocessing techniques. Caching frequently accessed data and preprocessing data before feeding it into AI models can significantly enhance processing speed and responsiveness, contributing to improved overall performance in SAP projects.

Yes …. it’s a long list (and that was just the tip of the ice berg!)

The integration of artificial intelligence into SAP projects presents a unique set of technical and business challenges in order to realize the potential benefits of these technologies. Overcoming complexities in integration processes, ensuring data compatibility, addressing interoperability issues, and navigating user acceptance challenges are crucial for successful A.I implementation in SAP projects.

Deeper Dive (the first 10 years)

2022 through at least 2032+

Developing new business solutions SAP RAM, SAP Process Control (PC), SAP Risk Management (RM), SAP Cloud Identity Access Governance (IAG) and SAP Signavio 

The aim being to get these certified by SAP and added to the SAP Store

A.I Vision

Make agile decisions, unlocking insights, and automated tasks with AI

Use AI that is trained on an industry and company data, accessible through SAP RAM, PC, RM etc. 

Run responsible AI built on leading ethics and data privacy standards while maintaining full governance and life cycle management.

AI & ML via SAP RAM & SAP IAG – The Art of the Possible

• Increase performance across a range of financial activities
• Guard against fraud with AI-assisted anomaly detection
• Simplify financial close with intelligent intra company reconciliation
• Build AI into SAP RAM, PC, RM etc with a library of pretrained models
• Manage the AI model life-cycle in one central place
• Deploy and run AI models at scale without compromising data privacy
• Industry AI – Leverage A.I and ML to work in industry packaged content to manage complexity, overcome challenges, and modernize a business.
• Automate and optimize complex industry processes
• Stay ahead of the competition with predictive, data-driven practices
• Disrupt disruption by enabling industry convergence
• Achieve Process Automation & Continuous Control Monitoring

Use Cases

Optimized Supply Chain: A.I automation within SAP can help improve demand forecasting and (in our case, help detect patterns of Fraud). A.I-driven supply chain forecasting management can reduce errors by 20–50%, cut warehousing costs by 5–10%.

Procurement Excellence: A.I automation can streamline procurement processes by assessing supplier performance, timely onboarding and offboarding of new suppliers, expedite invoice processing and accuracy, and detection of suspicious trends.

Sales and Marketing Precision: A.I in SAP can provide valuable insights into customer behaviour and preferences, enabling personalized marketing campaigns. This has a direct impact on return on investment (ROI). In fact, in one study, A.I solutions boosted revenue growth for 67% of the companies surveyed and reduced costs for 79% of them.

Human Resources Transformation: A.I can help your human resources team acquire talent, engage employees, and remove human bias in hiring, and far wider ensure organizations meet regulatory obligations (Child labour, Human Rights, Equal Pay etc).

Innovation Acceleration: By analyzing vast datasets, AI can identify patterns and trends, providing critical insights for informed decision-making. A.I can also identify opportunities for innovation and process improvement.

Success! 

Relating to each of the following financial years

• 1^st Feb 2023 – 31^st Jan 2024.
• 1^st Feb 2024 – 31^st Jan 2025.
• 1^st Feb 2025 – 31^st Jan 2026.

Significant focus was placed upon Retail, Mining, Manufacturing and Cyber

Mining was the biggest focus in 2023-2024 (but also continued through 2024 and 2025)

Cyber was the biggest focus in 2024-2025

Retail was the biggest focus in 2025 

Success stories – these have each been added to the SAP Storefront – which have to go through significant gates to be approved by SAP

MINING PACK
https://www.sap.com/products/financial-management/partners/winterhawk-consulting-ltd-mining-for-risk-and-assurance-management.html

CYBER PACK
https://www.sap.com/products/financial-management/partners/winterhawk-consulting-ltd-cybersecurity-for-risk-and-assurance-management.html

MANUFACTURING PACK
https://www.sap.com/products/financial-management/partners/winterhawk-consulting-ltd-manufacturing-for-risk-and-assurance-management.html

In the case of Retail it is broken out into no less than 20 different verticals (see further below graphics) – this area has been especially challenging and has been delayed. 

Andrew Sawyer, Operations Director and Head of Innovation at Winterhawk said:

It’s been a tough but already rewarding road with 3 innovation packs now accepted by SAP for their SAP store – but when working with AI services like SAP AI Core and SAP BTP here are a few of the development challenges (which related to each of the packs and occurred multiple times often across them all):

Service Subscription Errors:
Issues with subscribing to AI applications or services in the BTP cockpit, sometimes due to regional limitations.

API Execution Failures:
Errors when calling AI APIs, often encountered during the “try out” phase in SAP Cloud Integration or due to incorrect configuration.

Quota Exceeded Errors:
Running into service limits, such as routes quota exceeded errors on a trial account.

Integration and Configuration Issues:
Problems with integrating different systems and data sources, which require careful exception handling and often lead to API call failures if not managed properly.

As we found, this is something that can set you back months and months and when you are Winterhawk rather than Accenture, we don’t have a team of 100 people to figure it out, and nor do we have a hotline to SAP that gets swiftly answered!

SAP BTP AI Project Failures
Beyond technical glitches, many AI projects fail to deliver tangible business outcomes due to strategic and data-related challenges. In our case, it’s all about the data, and with regulations, if you take something like Sanctions, the data is to be blunt a mess, it needs formatting and what one distributor provides doesn’t remotely match up to the next. Which brings me to the next topic, data quality …. 

Poor Data Quality:
AI models are highly dependent on the quality of the data they are trained on (“garbage in, garbage out”). Inaccurate, biased, or incomplete data will lead to flawed AI outputs.

Insufficient Integration Planning & SAP Strategic Product Directional Changes : 
Failure to plan how AI solutions will integrate with existing SAP and non-SAP systems results in data silos, latency, and unreliable data pipelines. Our biggest challenge on this Winterhawk development initiative is SAP continually changing its products, adding new, removing old, flipping from Private to Public Cloud, On Premise v Subscription, it all has huge impact on the development work – at the most basic level just imagine working on a 2 year development of Retail Regulations for SAP Process Control only to then hear that SAP have decided to shutdown that product or decided to no longer support it after 2030. Then what? (and yes, that has literally happened to us)

Internal Silos:
A disconnect between data science teams and the core business units can lead to solutions that do not align with day-to-day business processes, hindering user adoption. Inadequate Training and Change Management: Customers may resist new AI tools if they do not understand how the technology improves their work or if they lack sufficient training. Algorithmic Bias: If the training data is biased, the AI model will replicate that bias, leading to real-world consequences like discriminatory loan decisions or misidentified individuals. Lack of Ethical Governance: Overlooking compliance and ethical standards in data use can lead to data breaches and erosion of customer trust.

RPA Failures
Unrealistic Expectations: Expecting immediate, overnight success or underestimating the effort required for development, testing, and fine-tuning can lead to perceived failure. Automating the Wrong Tasks: RPA excels at high-volume, repetitive tasks with structured data and clear rules. Automating complex processes involving frequent human judgment, unstructured data, or constant changes is a common pitfall. We are certainly finding with SAP, and with RPA that more fails that works at the first, second and third times of asking. Is it SAP and its infrastructure and products not really being truly ready, is it all of us learning on the fly, or is it a mix of them all, its hard to tell. 

Lack of Proper Project Management & Governance: Without a dedicated project team, clear sponsorship, and a Center of Excellence (CoE), our development efforts can lack direction and oversight. We have certainly been guilty of this. In an ideal world we would add more Project Management and Governance and have everything built out into Business Cases …… but …. we are (a) trying to learn as we go (b) we dont have a big budget (c) we have customer work that takes priority because that pays the bills … 

Technical issues related to the SAP BTP environment are continually reported in the SAP Community and support channels – that process is SLOW, VERY VERY VERY SLOW. Again, back to my earlier comments, if you are Accenture, I am sure you get a fast response, if you are Winterhawk you are lucky to get any response. 

Here are some other examples

Connectivity and Destination Errors: Issues connecting the Desktop Agents and to SAP back-end systems can occur. 

Agent Communication Timeouts: The Desktop Agent may fail to download project packages due to “WS MESSAGE TIMEOUT” errors, which can be intermittent or constant.

Authentication and Entitlement Issues

Missing Entitlement: Users may be unable to access the RPA service or activate their account if the necessary entitlements are not granted in the BTP global account settings.

Agent Activation Failures: The agent may register successfully but fail to activate the tenant, sometimes caused by network or firewall policies interfering with SSL inspections.

Project Integrity Errors: Failures can occur if a project has an integrity issue, for instance, due to conflicts between agent versions and project design or corrupted project files.

Scheduling Conflicts: Running multiple unattended jobs consecutively without proper configuration can cause failures, as a running background agent might prevent a second job from executing correctly.
Troubleshooting and Prevention

Consult SAP Support & Documentation: SAP provides knowledge base articles and troubleshooting guides for common technical issues. Refer to the SAP Support Knowledge Base and SAP Help Portal for specific error resolutions.

Focus on Error Handling: We needed to incorporate robust error and timeout management within the automation workflows themselves to log exceptions and handle unexpected pop-ups or system delays gracefully.

Innovation Vision – Development of Risk Packages

Enabling accelerated maturity progression, lowering risks, providing increased business assurance – leveraging A.I, RPA, ML.

When we first announced this we are asked if someone else could do this?

Answer : Extremely difficult :

• This is a niche and highly specialized area. Winterhawk is the only global consultancy 100% dedicated to SAP and SAP GRC solutions – the only one.
• There is significant development time investment and commercial risk.
• An expanded Advisory Board and wider team of experts needing to be tapped into to gain a depth of industry insights, risks and challenges.
• Deep coding skills in SAP.
• Deep experience in SAP Risk applications.
• Specialist skills in regulatory Acts and how to build them in SAP applications.
• Ability to automate technical build steps
• Existing knowledge of A.I, ML, Robotic techniques and ongoing learning (needing to continually evolve as SAP systems and roadmaps moves very quickly with A.I, ML enhancements / changes)

Resolving challenges through Innovation

Steve Hewison, MD and Founder at Winterhawk said: “Over recent years we’ve seen the challenges clients are facing coming to terms with increasing regulatory requirements and expectations, including for example ESG reporting.

Cost is a major factor, coming off the back of several challenging COVID years, high inflation, recession, an energy crisis and spiking wage bills. Organizations are seeking value from their technology purchases more now than ever before, and they want to see a very clear ROI.

Data can be in so many different places within an organization, and then with suppliers you have a whole raft of different data and reporting challenges. Each industry has its own nuances and a myriad of individual regulations and standards which an organization needs to adhere to.

Our approach, is to build solutions, leveraging A.I, ML and Robotic to enable our customers to gain immediate value from SAP solutions, rather than having to build from the bottom up. Providing readymade industry specific packages, that provide both a technological advancement as well as an immediate acceleration value to clients of any size who are running SAP.”

Andrew Sawyer, Operations Director and Head of Innovation at Winterhawk said: “It’s exciting work, we are developing solutions which will change the lives of our customers. Regulations might not sound that much fun but every organization in the world has to adhere to them, that’s the reality. When you have much of your entire business running digitally through systems like SAP – you need control over your finances and suppliers, and while SAP Risk solutions provide a platform, without innovation from partners, they are cannot solve the customer challenge. What we have developed and are continuing to develop, are truly unique regulatory packages, enabling customers to “Plug and Play” all their relevant regulations, with the controls and testing they will needed, to successfully manage their organization. It’s a game changer for someone who’s the head of Internal Controls providing an extra layer of brand and reputation protection for the company”

Dr Neil Patrick, SAP Director said : ““I’ve known Winterhawk since I joined SAP almost a decade ago, and have always been supported by them, impressed by their professionalism, and equally impressed by their innovation. Winterhawk have seized the opportunity to proactively develop new business outcomes for their customers, industries, lines of businesses. This is a game changer for accelerating customer value in SAP Risk & Assurance Management while enabling VARs to drive new recurring revenue. It is truly unique what they are doing here. I’m really excited by the impact, especially for our mid-market customers, SAP are right behind this very exciting and totally unique development”

SAP Dr Neil Patrick presenting Winterhawk Innovations at an SAP Conference

ONGOING

• Expanded Industries
• Expanded Sustainability
• Sanctions & Watch lists

It’s difficult to provide an exact number of regulations that exist worldwide. The number of regulations constantly changes as new laws and policies are enacted – it’s safe to say 1,000s and 1,000s. 

One example, development using A.I & ML , an SAP “APP” for Sanctions and then Regulations, and then Integrated with SAP S/4 HANA and then later with SAP Risk & Assurance Management. A first in kind, groundbreaking solution – that will provide customers with a by the minute position on Sanctions and Regulations around the world (or whatever jurisdictions they select), fully and seamlessly integrated into their core SAP and SAP GRC  systems.

Transforming how the world  (and in our specific case, customers running SAP) consumes Regulation.

Digitizing regulation at source to democratise regulatory data. We are standardizing how regulation is organised, presented, and consumed by transforming regulation as written today into machine-readable and consumable forms of data – Integrated into core SAP

Challenges – Regulators – Customers – Winterhawk

Taking Retail as an example

Retail isn’t just Retail – within Retail there are many different verticals that break out, each one having its own development path (see last slide)