SILOED! Business Unit vs IT Department ??

Our drive to build ever-more-complex systems for digital transformation is hampered by a classic anti-pattern: organizational silos. These silos, reflected in our system architecture, impede customer value delivery. They stem from our habit of structuring teams by domain, not by customer value flow.

As Mel Conway so astutely observed, "organizations which design systems ... are constrained to produce designs which are copies of the communication structures of these organizations". In other words, our system architecture is a mirror of our organizational structure. And when that structure is fragmented into silos, our systems suffer accordingly.

The consequences of these silos are far-reaching. They lead to a proliferation of hyper-specialized teams, each with their own technical expertise but often lacking in business and domain knowledge. They create a culture of fragmentation, where data is hoarded and access is restricted, rather than being treated as a product that can be leveraged across the organization. And they ultimately hinder our ability to innovate, as we struggle to integrate disparate systems and teams.

But inefficiency is just the tip of the iceberg. Silos are becoming a major roadblock on the digital transformation highway. As we pour resources into cutting-edge tech like cloud and AI, we're inadvertently creating fertile ground for new silos to sprout and solidify. Consider Microsoft Cloud 's hefty investments in regional expansion, data centers, and cloud/AI initiatives (think $17 billion regionally between 2023 and 2024, with data center spending exceeding $9 billion!). Without careful planning, these investments could simply reinforce existing silos. It's high time we confront the silo stranglehold and break free.

In this article series, we'll explore the problem of silos in depth, and examine the strategies we can use to break them down and create a more integrated, customer-centric approach to system design.

This time around, we'll discuss:

SILOED! Business Unit vs IT Department ?? (A true classic.)

Imagine a chess game where neither player can see the other’s board. That’s the reality for many organizations attempting AI implementation. Business units and IT, supposed allies, are stuck in a blind game, making disconnected moves without understanding each other’s goals or limitations.

A recent study found a communication chasm: 82% of business leaders struggle to translate their needs into technical terms for AI projects. Meanwhile, 43% of IT professionals believe their business counterparts wouldn’t know a neural network from a network cable. This gap isn’t a minor inconvenience; it’s a recipe for project disaster. Unrealistic expectations and missed deadlines lead to abandoned AI dreams.

Project ownership is another hurdle. AI initiatives become a multi-million dollar hot potato, tossed around without a champion. A whopping 61% of AI projects fail due to this ownership ambiguity.

Lost in Translation

Remember the Babel fish from The Hitchhiker’s Guide to the Galaxy? We need a corporate version for AI projects. Businesses speak in KPIs, ROIs, and market share, while IT responds with APIs, ML algorithms, and scalability metrics. It’s like a conversation between someone speaking Shakespearean sonnets and another using Boolean logic. Misunderstandings abound. Business leaders demand science fiction solutions, and IT struggles to explain AI’s value in C-suite terms.

Traditionally, projects have clear ownership. Marketing owns campaigns, finance owns budgets. But AI is like international waters — everyone wants the benefits, but no one claims responsibility. This lack of ownership creates inefficiency. Without clear leadership, AI initiatives become organizational orphans, passed around with dwindling enthusiasm. The result? Initial excitement followed by a slow descent into project purgatory.

The Multi-Disciplinary Imperative: Breaking Down Silos

The root of these communication and ownership problems lies in traditional siloed corporate structures. AI is inherently multidisciplinary. It requires a blend of technical expertise, business understanding, ethical considerations, and domain knowledge. Trying to squeeze AI into existing structures is a square peg in a round hole.

To bridge the gap, we need education on both sides. Business leaders should gain foundational knowledge of AI, and IT professionals should develop business acumen. “AI literacy” programs can create a shared understanding of how AI drives business value.

For ownership, creative structures are needed. Some companies experiment with AI Centers of Excellence — cross-functional teams acting as internal AI consultants. Others appoint Chief AI Officers to bridge the business-technology gap.

The key is to create structures that acknowledge the unique, cross-cutting nature of AI. It’s not just another IT project, nor purely a business initiative. It’s a new approach to problem-solving and value creation.

The High Cost of Low-Quality Data

We’ve all heard the adage “garbage in, garbage out.” In AI, it translates to “slightly imperfect data in, disastrously biased results out.” Imagine a misplaced decimal causing your AI to make nonsensical conclusions, like a butterfly effect gone wrong. A staggering $3.1 trillion is lost annually due to poor data quality — enough to treat everyone on Earth to coffee, with biscotti on the side!

The root of the problem? A classic turf war. Business units hoard data like dragons guarding treasure, while IT wrestles with building infrastructure on an unstable foundation. Without collaboration, your AI is as reliable as a weather forecast in a hurricane.

The $3.1 trillion price tag isn’t theoretical. It represents lost productivity, missed opportunities, and decisions based on faulty information. Think of it like baking a cake with expired ingredients. Training an AI model on incomplete or biased data isn’t just ineffective; it’s harmful. Imagine a healthcare AI diagnosing patients based on a dataset that underrepresents certain groups, or a financial AI making loan decisions based on skewed historical data. Not only are resources wasted, but societal inequalities are perpetuated.

The Human Firewall

The fear of AI goes beyond job loss. It’s about a fundamental shift in how we approach work, decision-making, and creativity. Here are some key anxieties:

• Loss of Control: Managers accustomed to making decisions based on experience suddenly face algorithms that process information faster and more precisely. It’s like a chess grandmaster forced to play Deep Blue.
• Skill Obsolescence: Workers worry their hard-earned skills will become obsolete. The rapid pace of technological change means the shelf life of skills is shrinking.
• Ethical Concerns: As AI influences decisions impacting lives, concerns about fairness and transparency grow. Who’s accountable when an AI makes a bad call?
• The Black Box Problem: Many AI systems, especially deep learning models, have opaque decision-making processes. Professionals used to explaining their reasoning are uncomfortable relying on uninterpretable systems.
• Cultural Shift: AI disrupts cultures that value human intuition and experience. Data-driven decisions become the new norm, a seismic shift in organizational DNA.

This resistance is complex, a response to a technology that promises (or threatens) to reshape the world of work.

Halfway Measures Won’t Cut It

Organizations face a classic innovator’s dilemma with AI. The potential benefits are vast: increased efficiency, improved decision-making, novel products and services, and a competitive edge. However, fully embracing AI may necessitate disrupting power structures, altering processes, and potentially displacing employees.

Imagine being the captain of a ship. You recognize the potential of a new, efficient engine technology, but to use it, you need to stop the ship, retrain the crew, and convince passengers it’s safe and worth the disruption. All while competing with other ships — every moment spent retooling gives your competitors an edge.

This dilemma often leads to a halfway approach: limited, non-disruptive AI investments. AI becomes a fancy add-on, not a core reimagining of business operations.

The Way Forward

How do we navigate this data dilemma and human resistance? Here are some ideas:

• Data Governance as a Team Sport: Make data quality an organization-wide priority. Create cross-functional teams, implement data governance frameworks, and foster a culture that values good data.
• Metadata: Your Data Map: Invest in robust metadata management. Think of it as a detailed map of your data landscape, helping you understand data lineage, quality, and relevance for AI projects.
• Continuous Data Quality Monitoring: Implement automated systems to constantly assess data quality.
• AI Education for All: Educate everyone, from C-suite executives to frontline workers, on AI basics — its potential, limitations, and implications for their roles.
• Focus on Augmentation, Not Replacement: Position AI initiatives as tools to augment human capabilities, not replace them. Show how AI can handle routine tasks, freeing humans for more strategic and creative duties.

Let’s address the elephant in the room: job displacement fears. AI may automate some tasks, but it will also create new opportunities. The key is to view AI as an augmentation, freeing humans to focus on creative problem-solving, strategic thinking, and even deciding where to order lunch.

Invest in reskilling and upskilling programs. Show your team you’re preparing them for the future. It’s like evolution, but faster and with more online courses.

Creating a Common Language (aka "Taxonomy")

To bridge the communication gap, we need a common language — an AI Rosetta Stone. Consider creating an “AI Translator” role. These individuals translate between business-speak and tech-talk, ensuring everyone understands each other.

But don’t stop at translation. Create immersive experiences like “Day in the Life” programs where business leaders shadow IT professionals and vice versa. This fosters empathy and understanding.

The Ownership Orchestra

Let's move beyond the "IT project" or "business initiative" binary. AI transcends departments like a quantum particle defying classical physics. It's time to embrace the corporate equivalent of Schr?dinger's cat – a project that's simultaneously owned by everyone and no one until you open the box (or in this case, the Jira board).

Enter the revolutionary concept of distributed ownership. Picture this: a cross-functional steering committee that's less "bored" room and more War Room for your AI initiatives. We're talking representatives from IT, relevant business units, legal, and ethics all sitting at the same table, preferably round to avoid any "head of the table" squabbles.

Think of them as the Avengers of your AI project, assembling to save your organization from the threats of siloed thinking and fragmented ownership:

1. Iron Man (IT)

Your tech wizards bring the suit of armor to the fight. They're the ones who can turn the business's wildest AI dreams into functioning reality.

Case Study: At TechCorp, the IT team, led by their CTO "Tony," developed a natural language processing model that could understand and respond to customer queries in 17 languages. This wasn't just a tech showcase; it was a direct response to the business's global expansion strategy.

2. Captain America (Business Leaders)

These are your strategic visionaries, the ones who can see the big picture and rally the troops.

Case Study: At RetailGiant, the VP of Operations "Steve" identified an opportunity to use AI for inventory optimization. His understanding of the business need guided the IT team in developing a system that reduced overstock by 23% and understocking by 18% in just six months.

3. Black Widow (Legal)

Stealthy, sharp, and always three steps ahead of potential threats. Your legal eagles aren't here to say no; they're here to find a way to yes that doesn't end in a class-action lawsuit.

Case Study: At FinTech Innovators, the legal team, spearheaded by "Natasha," worked proactively with IT to develop an AI-driven fraud detection system that improved accuracy by 34% while ensuring compliance with GDPR, CCPA, and other data protection regulations.

4. Vision (Ethics Officer)

The moral compass that keeps your AI initiatives from wandering into the dark side.

Case Study: At HealthCare Forward, the ethics officer "Vision" collaborated with IT and business leaders to develop an AI triage system for the ER. The system not only improved wait times by 27% but also ensured equitable treatment across all demographic groups, avoiding the pitfalls of biased datasets that have plagued similar systems.

A Game of High-Stakes Musical Chairs

But here's where it gets interesting: rotate that leadership faster than a game of musical chairs at a programmers' party. Maybe your IT lead takes point during the development phase, wielding their technical knowledge like Thor's hammer. Then, as you move into deployment, the baton passes to your business lead, who can navigate the treacherous waters of change management and user adoption.

Case Study: At EdTech Frontier, their AI-powered personalized learning platform saw a rocky start when IT led the entire project. The system worked flawlessly from a technical standpoint but saw dismal adoption rates. The tide turned when they shifted leadership to the Head of Curriculum for the deployment phase. She understood the teachers' pain points and was able to position the AI as a classroom aide rather than a replacement, boosting adoption from 23% to 89% in one semester.

This rotation isn't just a bureaucratic shuffle; it's a strategic move to ensure accountability across the project lifecycle. It's like a corporate version of the Ocean's Eleven heist – each phase of the project needs a different specialist in the driver's seat.

Skin in the Game

But let's not forget, with great power comes great responsibility (wrong superhero universe, but the point stands). Each leader needs skin in the game.

Case Study: At IndustrialTech Solutions, they implemented a novel approach: each rotating leader's performance bonus was tied not just to their phase of the project, but to the overall success of the AI initiative. Suddenly, the IT lead was deeply invested in user adoption, and the business lead found herself brushing up on data architecture. The result? An AI predictive maintenance system that not only functioned flawlessly but was enthusiastically embraced across the organization, leading to a 47% reduction in unplanned downtime.

Remember, this isn't about diluting responsibility; it's about creating a symbiotic ecosystem where the success of the AI initiative is everyone's success. It's turning your project into a corporate Voltron, where each part is powerful, but together, they're unstoppable.

Make AI Tangible

Move beyond vague promises of “increased efficiency.” Develop an AI Value Dashboard for each project. This dashboard showcases how AI is impacting metrics the business cares about, like reduced stockouts or faster customer service resolution times.

Capture the human stories behind these metrics. How has AI changed your employees’ work? These narratives personalize the impact of AI.

Start small, dream big. Think of it as a trust trapeze act. Begin with low-risk, high-visibility pilot projects, like an AI for meeting room bookings. As you build confidence, gradually increase the complexity and impact of your AI initiatives.

Maintain transparency throughout this journey. Celebrate successes, but also be open about failures. Every misstep is a learning opportunity. Build trust in the AI and the integration process.

Composing the Future of Work

Let’s tackle job displacement head-on.

Change is scary, and AI can be downright terrifying. A recent study found that 54% of business leaders fear AI displacing jobs. Imagine a sales manager facing an AI that generates leads better than their top performer, or a marketing director watching AI churn out copy that outshines Hemingway. It’s enough to make anyone nervous.

Instead of fearing AI, let’s reimagine work with AI support. Initiate a company-wide “AI Augmentation Audit”. This audit identifies tasks ripe for AI assistance and helps employees see AI as a collaborator.

Develop personalized “AI Collaboration Pathways” that blend technical skills with uniquely human skills like emotional intelligence and creative problem-solving. Create an “AI Future Lab” where employees experiment with emerging AI technologies.

Ethical considerations are crucial. Establish an “AI Ethics Review Board” with internal and external experts. This board approves, rejects, or modifies AI projects based on ethical considerations.

Develop a clear, public-facing “AI Ethics Charter” outlining your principles on data privacy, algorithmic bias, and transparency. Make ethics a key component of your AI training programs.

Iterate, Integrate, Innovate

AI integration is an ongoing journey. Establish regular “AI Alignment Summits” where business units and IT collaborate. Create cross-functional “AI Innovation Squads” to explore new ways to leverage AI. Implement an “AI Suggestion Box” where anyone can submit ideas.

As we stand at the precipice of the AI revolution, the marriage of business and technology is imperative. Each thread, from visionary business leaders to coding virtuosos, is essential. The future belongs to those who can bridge the gap between conception and creation.

The choice is yours. Will you be the architect of synergy or the harbinger of discord? Find your “why” in this AI journey, be it efficiency, innovation - or simply creating something greater than the sum of its parts.

Learn more about the way of the IT Architect:

IT’s not magic, it’s architecture: Leading, Aligning, and Innovating IT & Business with Principles…

Like the images in this article? You can do that, too:

The DALL-E Cookbook For Great AI Art: For Artists. For Enthusiasts.

Launch into the series with "SILOED! Data vs Product":

https://www.dhirubhai.net/pulse/siloed-data-vs-product-mohammed-brueckner-kgcxe/