WEBVTT
00:00:00.160 --> 00:00:01.360
Chapter 3.
00:00:01.760 --> 00:00:12.480
The Intersection of Data and Artificial Intelligence In this chapter we explore the intersection of data and artificial intelligence, focusing on how information becomes intelligence.
00:00:12.720 --> 00:00:22.719
Building on the foundations set out in Chapter 1, Data, and Chapter 2, AI, we look at how data and intelligence form a continuous value chain.
00:00:22.960 --> 00:00:24.960
Data fuels the AI engine.
00:00:25.199 --> 00:00:28.399
AI in turn unlocks the latent value of data.
00:00:28.640 --> 00:00:35.200
Continuous learning closes the loop, driving accuracy, adaptability, and insight over time.
00:00:35.679 --> 00:00:45.679
Governance connects these worlds, ensuring that as enterprise intelligence grows, it remains explainable, auditable, and aligned with organizational trust.
00:00:46.079 --> 00:00:55.759
Finally, we consider the strategic and economic implications of bringing data and AI together, where data and intelligence meet.
00:00:56.000 --> 00:01:07.920
In chapter one, we traced the evolution of data, how it became the foundation of enterprise information management, and how its structure, stewardship, and accessibility generate business value.
00:01:08.799 --> 00:01:21.760
Chapter two took a closer look at artificial intelligence as the engine of automation and intelligence, tracing its technological evolution beyond the hype cycle toward real-world agentic AI deployment.
00:01:22.000 --> 00:01:27.040
Throughout this book, our central thesis is that data and AI are symbiotic.
00:01:27.200 --> 00:01:34.000
Data gives AI the context and potential to learn, while AI transforms data in actionable insights.
00:01:34.239 --> 00:01:38.079
Together, they are driving innovation across modern enterprises.
00:01:38.319 --> 00:01:41.840
High performing AI requires high quality data.
00:01:42.079 --> 00:01:48.000
This is data that is well governed, structured, context rich, and secure.
00:01:48.239 --> 00:01:50.319
Not all data is created equal.
00:01:50.560 --> 00:01:56.719
Different data sets have different requirements, particularly as organizations balance public and private data.
00:01:56.959 --> 00:02:08.400
Public data sets train the large language models, LLMs, that underpin tools like ChatGPT, but for enterprises, private, customized data is the critical differentiator.
00:02:08.639 --> 00:02:16.879
Strategies that maintain the privacy and sovereignty of this data while enabling AI to learn from it are essential for competitive advantage.
00:02:17.039 --> 00:02:24.800
It's at this intersection where enterprise data meets intelligent systems that the true potential of enterprise AI is realized.
00:02:25.039 --> 00:02:29.840
Here, context becomes capability and information becomes insight.
00:02:30.159 --> 00:02:36.719
Organizations that can harness this relationship responsibly will define the next era of digital performance.
00:02:37.039 --> 00:02:41.120
Can AI replace data and enterprise information management?
00:02:41.439 --> 00:02:48.800
As AI adoption increases, a common question is whether AI can replace existing data and information management solutions.
00:02:49.039 --> 00:02:50.639
The short answer is no.
00:02:50.800 --> 00:02:56.560
However, it is the intersection of information management and AI that drives outcomes.
00:02:56.879 --> 00:02:59.280
One can't act without the other.
00:02:59.599 --> 00:03:15.680
AI automates data-specific actions, like data extraction and classification, whereas information management provides secure, organized content, and importantly, the foundational structure and rules for governance and compliance that AI does not.
00:03:15.919 --> 00:03:23.520
While data and information management solutions provide the fuel for AI, AI is also transforming information management.
00:03:23.759 --> 00:03:26.879
It does this through automation.
00:03:27.280 --> 00:03:36.240
AI automation tags documents, runs, extracts, summarizes reports, and reduces human error across key workflows.
00:03:37.199 --> 00:03:38.159
Insights.
00:03:38.560 --> 00:03:47.280
AI delivers valuable insights on the content that is being managed, including deriving key insights, sentiment, and other important notes.
00:03:47.520 --> 00:03:48.960
Search and retrieval.
00:03:49.360 --> 00:03:57.360
AI in combination with the metadata in content management makes search interfaces more accurate, efficient, and easier to use.
00:03:57.680 --> 00:04:00.960
Information management is the gatekeeper for trusted data.
00:04:01.199 --> 00:04:05.120
Data quality defines the credibility of every AI decision.
00:04:05.439 --> 00:04:08.240
The two disciplines are deeply interconnected.
00:04:08.479 --> 00:04:17.680
Effective AI relies on governed, high-integrity data, while information management gains new speed and intelligence through AI-driven automation.
00:04:17.920 --> 00:04:24.240
Integrating the two ensures consistency, compliance, and context across the information life cycle.
00:04:24.480 --> 00:04:27.519
We explore this in greater depth in chapter 5.
00:04:27.839 --> 00:04:35.759
AI can enhance how organizations manage content, but it cannot replace the discipline and governance that make information trustworthy.
00:04:36.079 --> 00:04:50.000
This interdependency between AI and information management is illustrated in the following case study about a global foods producer that is applying AI to its business information to modernize operations and improve performance.
00:04:50.319 --> 00:04:53.600
Case study A Global Foods Manufacturer.
00:04:53.839 --> 00:05:02.160
A foods manufacturer with operations around the world has implemented several strategies to position itself as an industry leader in AI adoption.
00:05:02.399 --> 00:05:06.879
What follows are excerpts of an interview with the company's EIM director.
00:05:07.120 --> 00:05:13.439
As part of our transformation project, we're exploring how artificial intelligence can help us modernize operations.
00:05:13.680 --> 00:05:17.680
Today, roughly 10% of our data is stored in the cloud.
00:05:17.839 --> 00:05:25.439
It's not a big number yet, and the cloud solutions we've used so far have been private to ensure the security of our proprietary information.
00:05:25.839 --> 00:05:36.240
But the technology is advancing quickly, and we're becoming more open to public cloud adoption, provided we can ensure strong governance and maintain ownership of our data.
00:05:36.560 --> 00:05:41.279
AI is becoming central to how we manage and extract value from our information.
00:05:41.439 --> 00:05:49.519
We're using AI-driven systems to pull insight from our operational data, helping management run our plants more efficiently.
00:05:49.839 --> 00:05:57.600
The results are tangible, higher quality products, more sustainable practices, and measurable improvements to the bottom line.
00:05:57.920 --> 00:06:01.519
AI has also inspired entirely new ways of working.
00:06:01.759 --> 00:06:08.319
We're piloting drone-based crop monitoring, an extension of a practice we've used for years with satellite imagery.
00:06:08.560 --> 00:06:13.920
Satellites helped us estimate crop health, but they don't perform well under cloud cover.
00:06:14.160 --> 00:06:23.600
Drones, on the other hand, can be programmed to fly over entire fields, capture high-resolution images, and feed that data directly into our AI models.
00:06:23.839 --> 00:06:33.759
Once processed, the models predict crop performance, identify stress or disease, and even recommend specific irrigation or fertilizer adjustments.
00:06:34.000 --> 00:06:44.639
That same insight is then integrated into automated fertilizer spreaders, which apply the right amount of treatment in the right places, reducing waste and improving yield.
00:06:44.879 --> 00:06:47.920
We are also advancing into predictive agriculture.
00:06:48.160 --> 00:06:56.800
By combining AI models with decades of historical weather and crop data, we can forecast growing conditions two to three years out in specific regions.
00:06:57.040 --> 00:07:02.240
These models aren't perfect, but they're increasingly accurate and incredibly useful for planning.
00:07:02.480 --> 00:07:04.639
Every region of the world is different.
00:07:04.879 --> 00:07:08.079
It's soil, weather, crops, and farmers.
00:07:08.319 --> 00:07:13.360
Our challenge is to adapt to all of them, and AI helps us do that at scale.
00:07:13.600 --> 00:07:22.399
The technology lets us understand local conditions in real time and make decisions that improve productivity, sustainability, and resilience.
00:07:22.560 --> 00:07:26.959
What used to take weeks of manual analysis now happens continuously.
00:07:27.279 --> 00:07:34.800
AI has become not just a tool for insight, but a partner in how we grow, produce, and feed the world.
00:07:35.040 --> 00:07:37.920
The data and AI Value Chain.
00:07:38.160 --> 00:07:43.360
Unlocking value from data begins with understanding the data and AI value chain.
00:07:43.519 --> 00:07:50.639
This process starts with data generation and collection, where access to enterprise data becomes a fundamental enabler.
00:07:50.879 --> 00:07:57.600
Once collected, data must be integrated, cleaned, and governed to ensure quality and reliability.
00:07:58.000 --> 00:08:08.959
Organizations that have invested in strong information management practices are better positioned to accelerate AI adoption because they have already done the foundational work of enabling their data.
00:08:09.120 --> 00:08:11.759
However, data alone is not enough.
00:08:11.920 --> 00:08:16.000
Without structured processes and workflows, data is not actionable.
00:08:16.240 --> 00:08:24.160
AI becomes valuable only when applied to real business challenges, integrating into these workflows to generate measurable outcomes.
00:08:24.480 --> 00:08:29.600
This is where AI model training, fine-tuning, and validation comes into play.
00:08:29.920 --> 00:08:44.879
LLMs are initially trained on public data sets, but enterprises can extend their value by fine-tuning them with private data or by using retrieval augmented generation that connect AI to internal knowledge sources.
00:08:45.120 --> 00:08:50.399
The right strategy depends on the organization's goals, resources, and maturity level.
00:08:50.559 --> 00:08:56.399
However, regardless of approach, data quality and model governance are critical requirements.
00:08:56.639 --> 00:09:01.840
The final and often overlooked step in the value chain is the feedback loop.
00:09:02.159 --> 00:09:09.919
Many organizations rush to deploy AI capabilities without establishing mechanisms for continuous learning and improvement.
00:09:10.159 --> 00:09:14.000
This is where the true value emerges, especially with AI.
00:09:14.240 --> 00:09:21.039
Iterative fine-tuning allows model accuracy to improve over time and drive more impactful results.
00:09:21.279 --> 00:09:26.720
To put the data in AI value chain into context, consider an example from manufacturing.
00:09:27.039 --> 00:09:32.480
Sensors on the factory floor collect data, then feed this data into the AI models.
00:09:32.639 --> 00:09:37.120
Based on this data, the models make recommendations to optimize performance.
00:09:37.279 --> 00:09:43.120
This in turn generates new data sets that can continually and iteratively be improved.
00:09:43.440 --> 00:09:52.559
This approach is demonstrated in the following case study, which describes how Noor Brimps keeps the wheels rolling with predictive maintenance powered by actionable insights.
00:09:53.039 --> 00:09:55.360
Case study Noor Brimps.
00:09:55.919 --> 00:10:02.960
Based in Munich, the Noor Brimps Group is the world's leading manufacturer of braking systems for rail and commercial vehicles.
00:10:03.279 --> 00:10:11.679
For more than 110 years now, the company has pioneered the development, production, marketing, and servicing of state-of-the-art braking systems.
00:10:12.000 --> 00:10:28.720
Noorbrimps' ICOM, Intelligent Condition Oriented Maintenance platform brings digitization to the rail business, connecting wireless-enabled sensors aboard trains to a back office cloud-based network using an IoT, Internet of Things model.
00:10:29.200 --> 00:10:34.399
This platform transmits detailed data that can help predict repair and replacement needs.
00:10:34.559 --> 00:10:44.240
The ICOM platform required a powerful and user-friendly analytics component to enable the analysis of the data received to help users make data-driven decisions.
00:10:44.480 --> 00:10:50.720
The ability to make predictive data-driven decisions results in more efficient and cost-effective repairs.
00:10:51.039 --> 00:10:55.759
With data being continually collected, the volumes across a fleet are significant.
00:10:56.000 --> 00:11:05.039
Customers now have the ability to visualize the data through interactive graphical dashboards, reducing the reliance on IT to create new reports.
00:11:05.279 --> 00:11:19.840
For example, they can provide heat maps of condition-based events, such as overheating breaks on a specific incline, helping customers put measures in place to reduce component failures, extending component life, and ultimately saving money.
00:11:20.159 --> 00:11:22.559
Data as the fuel for AI.
00:11:23.600 --> 00:11:26.639
Data is the fuel that powers the AI engine.
00:11:26.879 --> 00:11:34.320
The quantity, quality, and diversity of data matters far more than the complexity of the AI models themselves.
00:11:34.480 --> 00:11:40.480
High quality diverse data sets give AI systems the context they need to learn effectively.
00:11:40.720 --> 00:11:52.000
Simple AI can deliver impressive results on high quality and diverse datasets, whereas complex AI cannot deliver the same results on low quality and homogeneous data sets.
00:11:52.240 --> 00:11:57.840
As described in chapter one, both structured and unstructured data fuel this mix.
00:11:58.159 --> 00:12:04.559
Unlocking this data safely and responsibly is the key to meaningful AI adoption in business context.
00:12:04.879 --> 00:12:14.080
For most organizations, success depends not on training massive public models, but on leveraging private data strategically within existing frameworks.
00:12:14.320 --> 00:12:19.919
And once this private data is unlocked for AI, protecting that AI becomes critical.
00:12:20.240 --> 00:12:24.879
This is the key to the concept of sovereignty that we will discuss later in the book.
00:12:25.120 --> 00:12:33.679
Differentiating between public and private data sets and providing the appropriate protections for those private data sets is an urgent priority.
00:12:34.000 --> 00:12:41.679
AI doesn't merely consume data, it interprets, enriches, and organizes it for use across the enterprise.
00:12:41.919 --> 00:12:45.679
In this sense, AI and data form a two-way relationship.
00:12:46.000 --> 00:12:54.080
AI uses data to learn, but it also enhances the value of data by improving its structure, integrity, and accessibility.
00:12:54.320 --> 00:12:56.399
The Continuous Feedback Loop.
00:12:56.639 --> 00:13:01.440
Every effective enterprise AI system utilizes a continuous feedback loop.
00:13:01.679 --> 00:13:03.519
Data trains AI models.
00:13:03.759 --> 00:13:11.519
AI produces insights, and those insights generate new data that refines both the model and the underlying data sets.
00:13:11.840 --> 00:13:16.480
Intelligence improves not in a straight line, but in cycles of learning.
00:13:16.720 --> 00:13:21.360
For instance, recommendation systems constantly learn from user behavior.
00:13:21.600 --> 00:13:26.159
Each interaction creates new data that helps the system make better predictions.
00:13:26.399 --> 00:13:33.279
Over time, this iterative refinement increases accuracy, personalization, and efficiency.
00:13:33.600 --> 00:13:37.360
Think of this in the context of your favorite online shopping website.
00:13:37.600 --> 00:13:44.879
Every click, purchase, or pause creates new signals that reshape the system's understanding of user intent.
00:13:45.120 --> 00:13:49.919
The next set of recommendations reflects what the model has learned since the last one.
00:13:50.159 --> 00:14:00.960
It's trained to provide you with relevant recommendations, but as you continue shopping, it uses that data to generate new insights and over time improve the quality of the recommendations.
00:14:01.440 --> 00:14:06.080
Equally important to this process are observability and monitoring.
00:14:06.320 --> 00:14:09.840
The loop must be governed and models evolved responsibly.
00:14:10.000 --> 00:14:19.519
Continuous oversight of a model performance and data flows ensures that AI systems remain reliable, explainable, and aligned with business objectives.
00:14:19.759 --> 00:14:25.840
As we will explore later in the book, the operational management of AI systems cannot be an afterthought.
00:14:26.000 --> 00:14:31.440
It must be contemplated as a strategic capability that underpins long-term success.
00:14:31.759 --> 00:14:33.919
Governance at the intersection.
00:14:34.320 --> 00:14:38.720
Governance lies at the heart of the data in enterprise AI intersection.
00:14:38.879 --> 00:14:50.320
On the data side, governance focus on privacy, lineage, access control, and compliance with regulations, such as GDPR, General Data Protection Regulation.
00:14:50.559 --> 00:14:57.679
On the AI side, governance emphasizes fairness, transparency, accountability, and explainability.
00:14:58.000 --> 00:15:05.039
These two domains are now coming together under shared principles such as ethics, auditability, and trust.
00:15:05.360 --> 00:15:21.360
Emerging AI trust frameworks and international standards like ISO slash IEC 42001 for AI management and ISO slash IEC 38505 for data governance illustrate this convergence.
00:15:21.600 --> 00:15:28.799
As these frameworks mature, they will shape how organizations design, deploy, and monitor AI responsibly.
00:15:29.039 --> 00:15:34.240
We take a deeper dive into data and AI governance in chapters five and six respectively.
00:15:34.559 --> 00:15:37.840
Integrating data and AI creates competitive advantage.
00:15:38.080 --> 00:15:42.720
Governing them responsibly is what turns it into lasting economic value.
00:15:43.039 --> 00:15:45.600
Strategic and economic implications.
00:15:46.399 --> 00:15:53.120
Finally, integration of data and AI creates both strategic advantage and economic opportunity.
00:15:53.360 --> 00:16:02.399
Organizations that align these capabilities effectively are better equipped to innovate, optimize operations, and differentiate in competitive markets.
00:16:02.639 --> 00:16:13.200
With so much anticipation from boards and executive leadership on the potential for enterprise AI, it's easy to understand some of the market disappointment around the pace of change and impact.
00:16:13.440 --> 00:16:18.639
This has put a spotlight on early AI pilots in enterprises and their relative success.
00:16:18.879 --> 00:16:29.120
However, it's worth noting that many early AI pilots have underdelivered because they relied on publicly trained models without contextualizing them with enterprise data.
00:16:30.000 --> 00:16:37.759
Enterprise leaders need to understand that their competitive advantage lies in safely and securely unlocking this data.
00:16:38.080 --> 00:16:49.840
The next phase of success lies in taking a data-centric AI approach that prioritizes improving data quality and process design over building ever more complex AI models.
00:16:50.320 --> 00:16:55.120
Good data and sound processes lead to reliable AI outcomes.
00:16:55.440 --> 00:17:05.519
While large-scale computing remains necessary for training foundational models, most enterprises can achieve meaningful value through smaller, targeted deployments.
00:17:06.319 --> 00:17:12.079
Understanding your data requirements help determine what level of compute investment is truly needed.
00:17:12.319 --> 00:17:17.200
Preventing overspending and aligning AI initiatives with real business value.
00:17:17.440 --> 00:17:24.960
This also helps to reduce concerns around AI for leaders and employees who might still be working to fully understand the technology.
00:17:25.200 --> 00:17:32.160
Discover how ITAC Software AG is using intelligence to enable smart factories in the following case study.
00:17:32.400 --> 00:17:35.519
Case study ITAC Software AG.
00:17:36.160 --> 00:17:46.160
Since its founding, ITAC, Internet Technologies and Consulting Software AG, has been specializing in providing Internet technologies for the manufacturing industry.
00:17:46.480 --> 00:18:00.319
The manufacturer of standard software and products for cross-company IT applications is an industry-leading system and solution provider of manufacturing execution systems, MES, for the entire supply chain.
00:18:00.640 --> 00:18:18.799
To offer its customers the greatest possible transparency and decision-making capability for production control, and to meet growing demands related to the Internet of Things, IoT, ITAC wanted to integrate business intelligence, BI, and analytics software into their MES suite.
00:18:19.039 --> 00:18:25.359
Doing so would support customer demands for manufacturing intelligence, quality control, and traceability.
00:18:25.599 --> 00:18:37.200
In addition to rapid, effective implementation and seamless integration, ITAC required the customization of reports, analysis, and dashboards with full interactivity and security.
00:18:37.440 --> 00:18:45.759
All these needed to be web-based, offer transparent personalization for various applications, and be available through different channels.
00:18:46.160 --> 00:18:59.039
ITAC now has the BI operational and analytical capabilities it needs to support customer demands for greater intelligence, quality control, and traceability throughout the entire manufacturing process.
00:18:59.359 --> 00:19:09.680
The solution ensures transparency and metrics management and supports product lifecycle management, budget control, and quality assurance, as well as field activity management.
00:19:09.920 --> 00:19:19.519
The company's clients can access and analyze large amounts of data centrally with extensible support for future expansion, delivering competitive advantage.
00:19:19.839 --> 00:19:27.119
As we've covered in this chapter, data and AI are inseparable partners, with data as the fuel for the AI engine.
00:19:27.440 --> 00:19:32.960
AI without data is directionless, and data without AI is not actionable.
00:19:33.200 --> 00:19:38.079
Together, they form the foundation of intelligent enterprise decision making.
00:19:38.640 --> 00:19:46.480
As organizations increasingly move toward AI-driven decision frameworks, strong governance and strategic alignment become essential.
00:19:46.720 --> 00:19:58.480
The intersection of data and AI represents not only an operational shift, but also an innovation frontier, redefining how organizations think, decide, and compete.
00:19:58.799 --> 00:20:06.960
This convergence marks the beginning of a new chapter in digital transformation, one where information truly becomes intelligence.
00:20:07.279 --> 00:20:10.000
The Fast Five Download one.
00:20:10.960 --> 00:20:13.680
Prioritize data quality and governance.
00:20:13.920 --> 00:20:19.359
Establish data readiness as an organizational mandate, not a project deliverable.
00:20:19.759 --> 00:20:31.359
Direct your teams to perform comprehensive data audits and implement governance policies that ensure accuracy, security, and accessibility across all critical information assets.
00:20:31.519 --> 00:20:36.480
Make data quality a board level priority to maximize AI effectiveness.
00:20:36.799 --> 00:20:40.960
Two, integrate AI into real business workflows.
00:20:41.279 --> 00:20:48.079
Embed AI in high impact business processes to identify two to three key operational areas.
00:20:48.319 --> 00:20:56.559
Example, customer support, supply chain optimization, risk management, where it can deliver immediate benefits.
00:20:56.880 --> 00:21:05.359
Task business and technical leaders with deploying AI solutions that leverage proprietary data to address real business challenges.
00:21:05.680 --> 00:21:10.559
Three, establish continuous feedback loops for AI improvement.
00:21:10.799 --> 00:21:12.960
AI performance is never static.
00:21:13.119 --> 00:21:15.759
It requires continual monitoring and training.
00:21:16.000 --> 00:21:25.920
Institute an organizational policy for ongoing AI model performance monitoring, including user feedback loops and automated retraining using new data.
00:21:26.240 --> 00:21:33.680
Assign accountability for this process to ensure models remain accurate, personalized, and aligned with business goals.
00:21:34.000 --> 00:21:34.559
4.
00:21:34.960 --> 00:21:38.720
Align data and AI governance for trust and compliance.
00:21:39.279 --> 00:21:54.240
Bring data in AI under a single governance framework, a point across functional task force that unites privacy, security, compliance, and ethical oversight to create consistent standards for how intelligence is built and applied.
00:21:54.480 --> 00:21:59.680
Adopt or benchmark against emerging standards, such as ISO slash IEC.
00:22:00.319 --> 00:22:09.119
42001 and 38505 to proactively manage legal, reputational, and operational risks.
00:22:09.519 --> 00:22:10.079
5.
00:22:10.480 --> 00:22:13.599
Take a data-centric approach to AI investment.
00:22:14.160 --> 00:22:18.160
Tie all investment decisions to data value and business outcomes.
00:22:18.319 --> 00:22:28.640
Before approving new AI projects, require business units to articulate how the initiative unlocks value from enterprise data and delivers measurable business results.
00:22:28.960 --> 00:22:36.240
Limit investment in large scale AI models unless justified by unique data assets and a clear path to ROI.
00:00:00.160 --> 00:00:01.360
Chapter 3.
00:00:01.760 --> 00:00:12.480
The Intersection of Data and Artificial Intelligence In this chapter we explore the intersection of data and artificial intelligence, focusing on how information becomes intelligence.
00:00:12.720 --> 00:00:22.719
Building on the foundations set out in Chapter 1, Data, and Chapter 2, AI, we look at how data and intelligence form a continuous value chain.
00:00:22.960 --> 00:00:24.960
Data fuels the AI engine.
00:00:25.199 --> 00:00:28.399
AI in turn unlocks the latent value of data.
00:00:28.640 --> 00:00:35.200
Continuous learning closes the loop, driving accuracy, adaptability, and insight over time.
00:00:35.679 --> 00:00:45.679
Governance connects these worlds, ensuring that as enterprise intelligence grows, it remains explainable, auditable, and aligned with organizational trust.
00:00:46.079 --> 00:00:55.759
Finally, we consider the strategic and economic implications of bringing data and AI together, where data and intelligence meet.
00:00:56.000 --> 00:01:07.920
In chapter one, we traced the evolution of data, how it became the foundation of enterprise information management, and how its structure, stewardship, and accessibility generate business value.
00:01:08.799 --> 00:01:21.760
Chapter two took a closer look at artificial intelligence as the engine of automation and intelligence, tracing its technological evolution beyond the hype cycle toward real-world agentic AI deployment.
00:01:22.000 --> 00:01:27.040
Throughout this book, our central thesis is that data and AI are symbiotic.
00:01:27.200 --> 00:01:34.000
Data gives AI the context and potential to learn, while AI transforms data in actionable insights.
00:01:34.239 --> 00:01:38.079
Together, they are driving innovation across modern enterprises.
00:01:38.319 --> 00:01:41.840
High performing AI requires high quality data.
00:01:42.079 --> 00:01:48.000
This is data that is well governed, structured, context rich, and secure.
00:01:48.239 --> 00:01:50.319
Not all data is created equal.
00:01:50.560 --> 00:01:56.719
Different data sets have different requirements, particularly as organizations balance public and private data.
00:01:56.959 --> 00:02:08.400
Public data sets train the large language models, LLMs, that underpin tools like ChatGPT, but for enterprises, private, customized data is the critical differentiator.
00:02:08.639 --> 00:02:16.879
Strategies that maintain the privacy and sovereignty of this data while enabling AI to learn from it are essential for competitive advantage.
00:02:17.039 --> 00:02:24.800
It's at this intersection where enterprise data meets intelligent systems that the true potential of enterprise AI is realized.
00:02:25.039 --> 00:02:29.840
Here, context becomes capability and information becomes insight.
00:02:30.159 --> 00:02:36.719
Organizations that can harness this relationship responsibly will define the next era of digital performance.
00:02:37.039 --> 00:02:41.120
Can AI replace data and enterprise information management?
00:02:41.439 --> 00:02:48.800
As AI adoption increases, a common question is whether AI can replace existing data and information management solutions.
00:02:49.039 --> 00:02:50.639
The short answer is no.
00:02:50.800 --> 00:02:56.560
However, it is the intersection of information management and AI that drives outcomes.
00:02:56.879 --> 00:02:59.280
One can't act without the other.
00:02:59.599 --> 00:03:15.680
AI automates data-specific actions, like data extraction and classification, whereas information management provides secure, organized content, and importantly, the foundational structure and rules for governance and compliance that AI does not.
00:03:15.919 --> 00:03:23.520
While data and information management solutions provide the fuel for AI, AI is also transforming information management.
00:03:23.759 --> 00:03:26.879
It does this through automation.
00:03:27.280 --> 00:03:36.240
AI automation tags documents, runs, extracts, summarizes reports, and reduces human error across key workflows.
00:03:37.199 --> 00:03:38.159
Insights.
00:03:38.560 --> 00:03:47.280
AI delivers valuable insights on the content that is being managed, including deriving key insights, sentiment, and other important notes.
00:03:47.520 --> 00:03:48.960
Search and retrieval.
00:03:49.360 --> 00:03:57.360
AI in combination with the metadata in content management makes search interfaces more accurate, efficient, and easier to use.
00:03:57.680 --> 00:04:00.960
Information management is the gatekeeper for trusted data.
00:04:01.199 --> 00:04:05.120
Data quality defines the credibility of every AI decision.
00:04:05.439 --> 00:04:08.240
The two disciplines are deeply interconnected.
00:04:08.479 --> 00:04:17.680
Effective AI relies on governed, high-integrity data, while information management gains new speed and intelligence through AI-driven automation.
00:04:17.920 --> 00:04:24.240
Integrating the two ensures consistency, compliance, and context across the information life cycle.
00:04:24.480 --> 00:04:27.519
We explore this in greater depth in chapter 5.
00:04:27.839 --> 00:04:35.759
AI can enhance how organizations manage content, but it cannot replace the discipline and governance that make information trustworthy.
00:04:36.079 --> 00:04:50.000
This interdependency between AI and information management is illustrated in the following case study about a global foods producer that is applying AI to its business information to modernize operations and improve performance.
00:04:50.319 --> 00:04:53.600
Case study A Global Foods Manufacturer.
00:04:53.839 --> 00:05:02.160
A foods manufacturer with operations around the world has implemented several strategies to position itself as an industry leader in AI adoption.
00:05:02.399 --> 00:05:06.879
What follows are excerpts of an interview with the company's EIM director.
00:05:07.120 --> 00:05:13.439
As part of our transformation project, we're exploring how artificial intelligence can help us modernize operations.
00:05:13.680 --> 00:05:17.680
Today, roughly 10% of our data is stored in the cloud.
00:05:17.839 --> 00:05:25.439
It's not a big number yet, and the cloud solutions we've used so far have been private to ensure the security of our proprietary information.
00:05:25.839 --> 00:05:36.240
But the technology is advancing quickly, and we're becoming more open to public cloud adoption, provided we can ensure strong governance and maintain ownership of our data.
00:05:36.560 --> 00:05:41.279
AI is becoming central to how we manage and extract value from our information.
00:05:41.439 --> 00:05:49.519
We're using AI-driven systems to pull insight from our operational data, helping management run our plants more efficiently.
00:05:49.839 --> 00:05:57.600
The results are tangible, higher quality products, more sustainable practices, and measurable improvements to the bottom line.
00:05:57.920 --> 00:06:01.519
AI has also inspired entirely new ways of working.
00:06:01.759 --> 00:06:08.319
We're piloting drone-based crop monitoring, an extension of a practice we've used for years with satellite imagery.
00:06:08.560 --> 00:06:13.920
Satellites helped us estimate crop health, but they don't perform well under cloud cover.
00:06:14.160 --> 00:06:23.600
Drones, on the other hand, can be programmed to fly over entire fields, capture high-resolution images, and feed that data directly into our AI models.
00:06:23.839 --> 00:06:33.759
Once processed, the models predict crop performance, identify stress or disease, and even recommend specific irrigation or fertilizer adjustments.
00:06:34.000 --> 00:06:44.639
That same insight is then integrated into automated fertilizer spreaders, which apply the right amount of treatment in the right places, reducing waste and improving yield.
00:06:44.879 --> 00:06:47.920
We are also advancing into predictive agriculture.
00:06:48.160 --> 00:06:56.800
By combining AI models with decades of historical weather and crop data, we can forecast growing conditions two to three years out in specific regions.
00:06:57.040 --> 00:07:02.240
These models aren't perfect, but they're increasingly accurate and incredibly useful for planning.
00:07:02.480 --> 00:07:04.639
Every region of the world is different.
00:07:04.879 --> 00:07:08.079
It's soil, weather, crops, and farmers.
00:07:08.319 --> 00:07:13.360
Our challenge is to adapt to all of them, and AI helps us do that at scale.
00:07:13.600 --> 00:07:22.399
The technology lets us understand local conditions in real time and make decisions that improve productivity, sustainability, and resilience.
00:07:22.560 --> 00:07:26.959
What used to take weeks of manual analysis now happens continuously.
00:07:27.279 --> 00:07:34.800
AI has become not just a tool for insight, but a partner in how we grow, produce, and feed the world.
00:07:35.040 --> 00:07:37.920
The data and AI Value Chain.
00:07:38.160 --> 00:07:43.360
Unlocking value from data begins with understanding the data and AI value chain.
00:07:43.519 --> 00:07:50.639
This process starts with data generation and collection, where access to enterprise data becomes a fundamental enabler.
00:07:50.879 --> 00:07:57.600
Once collected, data must be integrated, cleaned, and governed to ensure quality and reliability.
00:07:58.000 --> 00:08:08.959
Organizations that have invested in strong information management practices are better positioned to accelerate AI adoption because they have already done the foundational work of enabling their data.
00:08:09.120 --> 00:08:11.759
However, data alone is not enough.
00:08:11.920 --> 00:08:16.000
Without structured processes and workflows, data is not actionable.
00:08:16.240 --> 00:08:24.160
AI becomes valuable only when applied to real business challenges, integrating into these workflows to generate measurable outcomes.
00:08:24.480 --> 00:08:29.600
This is where AI model training, fine-tuning, and validation comes into play.
00:08:29.920 --> 00:08:44.879
LLMs are initially trained on public data sets, but enterprises can extend their value by fine-tuning them with private data or by using retrieval augmented generation that connect AI to internal knowledge sources.
00:08:45.120 --> 00:08:50.399
The right strategy depends on the organization's goals, resources, and maturity level.
00:08:50.559 --> 00:08:56.399
However, regardless of approach, data quality and model governance are critical requirements.
00:08:56.639 --> 00:09:01.840
The final and often overlooked step in the value chain is the feedback loop.
00:09:02.159 --> 00:09:09.919
Many organizations rush to deploy AI capabilities without establishing mechanisms for continuous learning and improvement.
00:09:10.159 --> 00:09:14.000
This is where the true value emerges, especially with AI.
00:09:14.240 --> 00:09:21.039
Iterative fine-tuning allows model accuracy to improve over time and drive more impactful results.
00:09:21.279 --> 00:09:26.720
To put the data in AI value chain into context, consider an example from manufacturing.
00:09:27.039 --> 00:09:32.480
Sensors on the factory floor collect data, then feed this data into the AI models.
00:09:32.639 --> 00:09:37.120
Based on this data, the models make recommendations to optimize performance.
00:09:37.279 --> 00:09:43.120
This in turn generates new data sets that can continually and iteratively be improved.
00:09:43.440 --> 00:09:52.559
This approach is demonstrated in the following case study, which describes how Noor Brimps keeps the wheels rolling with predictive maintenance powered by actionable insights.
00:09:53.039 --> 00:09:55.360
Case study Noor Brimps.
00:09:55.919 --> 00:10:02.960
Based in Munich, the Noor Brimps Group is the world's leading manufacturer of braking systems for rail and commercial vehicles.
00:10:03.279 --> 00:10:11.679
For more than 110 years now, the company has pioneered the development, production, marketing, and servicing of state-of-the-art braking systems.
00:10:12.000 --> 00:10:28.720
Noorbrimps' ICOM, Intelligent Condition Oriented Maintenance platform brings digitization to the rail business, connecting wireless-enabled sensors aboard trains to a back office cloud-based network using an IoT, Internet of Things model.
00:10:29.200 --> 00:10:34.399
This platform transmits detailed data that can help predict repair and replacement needs.
00:10:34.559 --> 00:10:44.240
The ICOM platform required a powerful and user-friendly analytics component to enable the analysis of the data received to help users make data-driven decisions.
00:10:44.480 --> 00:10:50.720
The ability to make predictive data-driven decisions results in more efficient and cost-effective repairs.
00:10:51.039 --> 00:10:55.759
With data being continually collected, the volumes across a fleet are significant.
00:10:56.000 --> 00:11:05.039
Customers now have the ability to visualize the data through interactive graphical dashboards, reducing the reliance on IT to create new reports.
00:11:05.279 --> 00:11:19.840
For example, they can provide heat maps of condition-based events, such as overheating breaks on a specific incline, helping customers put measures in place to reduce component failures, extending component life, and ultimately saving money.
00:11:20.159 --> 00:11:22.559
Data as the fuel for AI.
00:11:23.600 --> 00:11:26.639
Data is the fuel that powers the AI engine.
00:11:26.879 --> 00:11:34.320
The quantity, quality, and diversity of data matters far more than the complexity of the AI models themselves.
00:11:34.480 --> 00:11:40.480
High quality diverse data sets give AI systems the context they need to learn effectively.
00:11:40.720 --> 00:11:52.000
Simple AI can deliver impressive results on high quality and diverse datasets, whereas complex AI cannot deliver the same results on low quality and homogeneous data sets.
00:11:52.240 --> 00:11:57.840
As described in chapter one, both structured and unstructured data fuel this mix.
00:11:58.159 --> 00:12:04.559
Unlocking this data safely and responsibly is the key to meaningful AI adoption in business context.
00:12:04.879 --> 00:12:14.080
For most organizations, success depends not on training massive public models, but on leveraging private data strategically within existing frameworks.
00:12:14.320 --> 00:12:19.919
And once this private data is unlocked for AI, protecting that AI becomes critical.
00:12:20.240 --> 00:12:24.879
This is the key to the concept of sovereignty that we will discuss later in the book.
00:12:25.120 --> 00:12:33.679
Differentiating between public and private data sets and providing the appropriate protections for those private data sets is an urgent priority.
00:12:34.000 --> 00:12:41.679
AI doesn't merely consume data, it interprets, enriches, and organizes it for use across the enterprise.
00:12:41.919 --> 00:12:45.679
In this sense, AI and data form a two-way relationship.
00:12:46.000 --> 00:12:54.080
AI uses data to learn, but it also enhances the value of data by improving its structure, integrity, and accessibility.
00:12:54.320 --> 00:12:56.399
The Continuous Feedback Loop.
00:12:56.639 --> 00:13:01.440
Every effective enterprise AI system utilizes a continuous feedback loop.
00:13:01.679 --> 00:13:03.519
Data trains AI models.
00:13:03.759 --> 00:13:11.519
AI produces insights, and those insights generate new data that refines both the model and the underlying data sets.
00:13:11.840 --> 00:13:16.480
Intelligence improves not in a straight line, but in cycles of learning.
00:13:16.720 --> 00:13:21.360
For instance, recommendation systems constantly learn from user behavior.
00:13:21.600 --> 00:13:26.159
Each interaction creates new data that helps the system make better predictions.
00:13:26.399 --> 00:13:33.279
Over time, this iterative refinement increases accuracy, personalization, and efficiency.
00:13:33.600 --> 00:13:37.360
Think of this in the context of your favorite online shopping website.
00:13:37.600 --> 00:13:44.879
Every click, purchase, or pause creates new signals that reshape the system's understanding of user intent.
00:13:45.120 --> 00:13:49.919
The next set of recommendations reflects what the model has learned since the last one.
00:13:50.159 --> 00:14:00.960
It's trained to provide you with relevant recommendations, but as you continue shopping, it uses that data to generate new insights and over time improve the quality of the recommendations.
00:14:01.440 --> 00:14:06.080
Equally important to this process are observability and monitoring.
00:14:06.320 --> 00:14:09.840
The loop must be governed and models evolved responsibly.
00:14:10.000 --> 00:14:19.519
Continuous oversight of a model performance and data flows ensures that AI systems remain reliable, explainable, and aligned with business objectives.
00:14:19.759 --> 00:14:25.840
As we will explore later in the book, the operational management of AI systems cannot be an afterthought.
00:14:26.000 --> 00:14:31.440
It must be contemplated as a strategic capability that underpins long-term success.
00:14:31.759 --> 00:14:33.919
Governance at the intersection.
00:14:34.320 --> 00:14:38.720
Governance lies at the heart of the data in enterprise AI intersection.
00:14:38.879 --> 00:14:50.320
On the data side, governance focus on privacy, lineage, access control, and compliance with regulations, such as GDPR, General Data Protection Regulation.
00:14:50.559 --> 00:14:57.679
On the AI side, governance emphasizes fairness, transparency, accountability, and explainability.
00:14:58.000 --> 00:15:05.039
These two domains are now coming together under shared principles such as ethics, auditability, and trust.
00:15:05.360 --> 00:15:21.360
Emerging AI trust frameworks and international standards like ISO slash IEC 42001 for AI management and ISO slash IEC 38505 for data governance illustrate this convergence.
00:15:21.600 --> 00:15:28.799
As these frameworks mature, they will shape how organizations design, deploy, and monitor AI responsibly.
00:15:29.039 --> 00:15:34.240
We take a deeper dive into data and AI governance in chapters five and six respectively.
00:15:34.559 --> 00:15:37.840
Integrating data and AI creates competitive advantage.
00:15:38.080 --> 00:15:42.720
Governing them responsibly is what turns it into lasting economic value.
00:15:43.039 --> 00:15:45.600
Strategic and economic implications.
00:15:46.399 --> 00:15:53.120
Finally, integration of data and AI creates both strategic advantage and economic opportunity.
00:15:53.360 --> 00:16:02.399
Organizations that align these capabilities effectively are better equipped to innovate, optimize operations, and differentiate in competitive markets.
00:16:02.639 --> 00:16:13.200
With so much anticipation from boards and executive leadership on the potential for enterprise AI, it's easy to understand some of the market disappointment around the pace of change and impact.
00:16:13.440 --> 00:16:18.639
This has put a spotlight on early AI pilots in enterprises and their relative success.
00:16:18.879 --> 00:16:29.120
However, it's worth noting that many early AI pilots have underdelivered because they relied on publicly trained models without contextualizing them with enterprise data.
00:16:30.000 --> 00:16:37.759
Enterprise leaders need to understand that their competitive advantage lies in safely and securely unlocking this data.
00:16:38.080 --> 00:16:49.840
The next phase of success lies in taking a data-centric AI approach that prioritizes improving data quality and process design over building ever more complex AI models.
00:16:50.320 --> 00:16:55.120
Good data and sound processes lead to reliable AI outcomes.
00:16:55.440 --> 00:17:05.519
While large-scale computing remains necessary for training foundational models, most enterprises can achieve meaningful value through smaller, targeted deployments.
00:17:06.319 --> 00:17:12.079
Understanding your data requirements help determine what level of compute investment is truly needed.
00:17:12.319 --> 00:17:17.200
Preventing overspending and aligning AI initiatives with real business value.
00:17:17.440 --> 00:17:24.960
This also helps to reduce concerns around AI for leaders and employees who might still be working to fully understand the technology.
00:17:25.200 --> 00:17:32.160
Discover how ITAC Software AG is using intelligence to enable smart factories in the following case study.
00:17:32.400 --> 00:17:35.519
Case study ITAC Software AG.
00:17:36.160 --> 00:17:46.160
Since its founding, ITAC, Internet Technologies and Consulting Software AG, has been specializing in providing Internet technologies for the manufacturing industry.
00:17:46.480 --> 00:18:00.319
The manufacturer of standard software and products for cross-company IT applications is an industry-leading system and solution provider of manufacturing execution systems, MES, for the entire supply chain.
00:18:00.640 --> 00:18:18.799
To offer its customers the greatest possible transparency and decision-making capability for production control, and to meet growing demands related to the Internet of Things, IoT, ITAC wanted to integrate business intelligence, BI, and analytics software into their MES suite.
00:18:19.039 --> 00:18:25.359
Doing so would support customer demands for manufacturing intelligence, quality control, and traceability.
00:18:25.599 --> 00:18:37.200
In addition to rapid, effective implementation and seamless integration, ITAC required the customization of reports, analysis, and dashboards with full interactivity and security.
00:18:37.440 --> 00:18:45.759
All these needed to be web-based, offer transparent personalization for various applications, and be available through different channels.
00:18:46.160 --> 00:18:59.039
ITAC now has the BI operational and analytical capabilities it needs to support customer demands for greater intelligence, quality control, and traceability throughout the entire manufacturing process.
00:18:59.359 --> 00:19:09.680
The solution ensures transparency and metrics management and supports product lifecycle management, budget control, and quality assurance, as well as field activity management.
00:19:09.920 --> 00:19:19.519
The company's clients can access and analyze large amounts of data centrally with extensible support for future expansion, delivering competitive advantage.
00:19:19.839 --> 00:19:27.119
As we've covered in this chapter, data and AI are inseparable partners, with data as the fuel for the AI engine.
00:19:27.440 --> 00:19:32.960
AI without data is directionless, and data without AI is not actionable.
00:19:33.200 --> 00:19:38.079
Together, they form the foundation of intelligent enterprise decision making.
00:19:38.640 --> 00:19:46.480
As organizations increasingly move toward AI-driven decision frameworks, strong governance and strategic alignment become essential.
00:19:46.720 --> 00:19:58.480
The intersection of data and AI represents not only an operational shift, but also an innovation frontier, redefining how organizations think, decide, and compete.
00:19:58.799 --> 00:20:06.960
This convergence marks the beginning of a new chapter in digital transformation, one where information truly becomes intelligence.
00:20:07.279 --> 00:20:10.000
The Fast Five Download one.
00:20:10.960 --> 00:20:13.680
Prioritize data quality and governance.
00:20:13.920 --> 00:20:19.359
Establish data readiness as an organizational mandate, not a project deliverable.
00:20:19.759 --> 00:20:31.359
Direct your teams to perform comprehensive data audits and implement governance policies that ensure accuracy, security, and accessibility across all critical information assets.
00:20:31.519 --> 00:20:36.480
Make data quality a board level priority to maximize AI effectiveness.
00:20:36.799 --> 00:20:40.960
Two, integrate AI into real business workflows.
00:20:41.279 --> 00:20:48.079
Embed AI in high impact business processes to identify two to three key operational areas.
00:20:48.319 --> 00:20:56.559
Example, customer support, supply chain optimization, risk management, where it can deliver immediate benefits.
00:20:56.880 --> 00:21:05.359
Task business and technical leaders with deploying AI solutions that leverage proprietary data to address real business challenges.
00:21:05.680 --> 00:21:10.559
Three, establish continuous feedback loops for AI improvement.
00:21:10.799 --> 00:21:12.960
AI performance is never static.
00:21:13.119 --> 00:21:15.759
It requires continual monitoring and training.
00:21:16.000 --> 00:21:25.920
Institute an organizational policy for ongoing AI model performance monitoring, including user feedback loops and automated retraining using new data.
00:21:26.240 --> 00:21:33.680
Assign accountability for this process to ensure models remain accurate, personalized, and aligned with business goals.
00:21:34.000 --> 00:21:34.559
4.
00:21:34.960 --> 00:21:38.720
Align data and AI governance for trust and compliance.
00:21:39.279 --> 00:21:54.240
Bring data in AI under a single governance framework, a point across functional task force that unites privacy, security, compliance, and ethical oversight to create consistent standards for how intelligence is built and applied.
00:21:54.480 --> 00:21:59.680
Adopt or benchmark against emerging standards, such as ISO slash IEC.
00:22:00.319 --> 00:22:09.119
42001 and 38505 to proactively manage legal, reputational, and operational risks.
00:22:09.519 --> 00:22:10.079
5.
00:22:10.480 --> 00:22:13.599
Take a data-centric approach to AI investment.
00:22:14.160 --> 00:22:18.160
Tie all investment decisions to data value and business outcomes.
00:22:18.319 --> 00:22:28.640
Before approving new AI projects, require business units to articulate how the initiative unlocks value from enterprise data and delivers measurable business results.
00:22:28.960 --> 00:22:36.240
Limit investment in large scale AI models unless justified by unique data assets and a clear path to ROI.
