@tgroenwals

Most AI programs still aren’t hitting the P&L.

The ones that are all made the same early decision.

They stopped treating AI as a board-level response and started treating it as an operating model redesign.

88% of enterprises now use AI in at least one function. Yet only 39% report EBIT impact (McKinsey, 2025).

This isn’t an AI problem. It’s a rollout problem.

Here’s what’s actually happening inside most enterprises:

The Broken Playbook: How Most Enterprises Roll Out AI

1/ The Board Mandates It

→ Competitive pressure hits the boardroom
→ “We need an AI strategy” lands with the CEO
→…

The CIO role is being rewritten across retail.

It used to be:
• Infrastructure
• Systems
• Stability
• Cost control

Now?

It’s shifting towards:
• Commercial impact
• Customer experience
• Product and platform thinking
• Enabling growth

Technology is no longer sitting on the sidelines.

It’s embedded in how the business operates.

Which means tech leaders are now being judged differently.

Not on:
System uptime

But on:
Business outcome realisation

Revenue enablement
Speed to market
Customer experience
Operational performance

That’s a big shift.

Enterprise Architecture - From Cost Center to Value Engine

https://buff.ly/s2twgl2

Govern transformation like a project.
You’ll get project outputs.

A CEO said this to me recently.
“Rob, we’re spending millions each month,
but margin hasn’t moved in two years.”

The room full of senior leaders went still.
The CFO looked up at the on-screen dashboard.
Every status box was green.

She turned to the project sponsor and said:
"...and so how has this changed our numbers?"

Projects were completing on schedule.
Teams were celebrating milestones.
The business had little to show for it.

That comment stayed with me.
Because it’s a common pattern
in medium and large organisations.

Look at the dominoes in this image.

That is the most accurate representation of AI ROI I have seen all year.
Miss one domino, the chain stops.
And most companies miss the same one.

This 10-step framework from Vaibhav is sharp.
Save it. Send it to your team.

Then notice the pattern most readers will miss.

Steps 1 through 5 are head work.
Define ROI. Choose problems. Build the business case. Start small. Measure what matters.
This is the part that gets the budget approved.

Steps 6 through 10 are body work.
Embed AI into workflows. Drive adoption. Fix bottlenecks. Scale. Optimize.

𝐈𝐟 𝐲𝐨𝐮𝐫 𝐝𝐚𝐭𝐚 𝐬𝐭𝐫𝐚𝐭𝐞𝐠𝐲 𝐬𝐭𝐚𝐫𝐭𝐬 𝐰𝐢𝐭𝐡 "𝐰𝐡𝐢𝐜𝐡 𝐭𝐨𝐨𝐥," 𝐢𝐭'𝐬 𝐚𝐥𝐫𝐞𝐚𝐝𝐲 𝐰𝐫𝐨𝐧𝐠.

It should start with which problem.

Mesh, Fabric, and Lakehouse get pitched like rival platforms. They're not. They answer three completely different questions.

Use this to decide 👇

𝐈𝐬 𝐲𝐨𝐮𝐫 𝐩𝐫𝐨𝐛𝐥𝐞𝐦 𝐩𝐞𝐨𝐩𝐥𝐞? → 𝐃𝐚𝐭𝐚 𝐌𝐞𝐬𝐡
→ Central data team is the bottleneck
→ Domains can't move fast enough
→ Need: domain ownership, data-as-product, federated governance, SLA contracts…

The most misleading AI metric might be adoption.

Lots of people using AI doesn’t automatically mean your business is improving.

“We rolled out AI to 5,000 employees.” Great. What changed?

This is the question many organisations still struggle to answer.

Too often, AI measurement starts and ends with adoption metrics:
Number of users, prompts per day, licences activated.

Those metrics matter to start, but you need to keep an eye on bigger outcomes.
AI isn’t just a productivity tool.
It can change customer experience, revenue generation, operational speed and employee experience.

Hello everyone, leopardracer here.

I moved everything to a headless Mac Mini M4, the base $599 model with 16GB of RAM. Started with smaller Qwen models for message classification and context compression. Then I found a way to run a 35 billion parameter model on 16 gigs. 17 tokens per second. Zero swap. Everyone said you need 32GB minimum. I’ve been running it for about a week now.
Google dropped Gemma 4 under Apache 2.0. Within hours I benchmarked it against my Qwen setup. Classification went from 8.5 seconds to 1.9 seconds. By evening the swap was done. Five files changed.
This is all of it. How I fit a 35B model on a $600 machine, what local models actually do when you’re running an AI agent 24/7 (spoiler: it’s way more than classification), how model routing works across three tiers, and why Gemma 4 made me swap the brain overnight.
https://pbs.twimg.com/media/HFtDh1RXkAA3zqy.png
My Mac Mini runs as a headless AI automation server.

Two ways to get AI wrong:
chase every shiny object, or play it so safe you sleepwalk into irrelevance.

Some organisations pile into agents and GenAI and hope it changes everything.

Others stay in their comfort zone.
Rules, reporting and incremental improvements.

Both create hidden risk.

If you only chase big bets, results get fragile fast.
If you only back safe bets, you slowly fall behind.

The AI strategies that hold up over time look more like portfolios.

You have dependable layers:
▶️ Clear rules that protect compliance.

50% of SMEs are stuck in Phase 1.

Another 30% are stuck in Phase 2.

That's 80% of companies that bought AI tools and got nothing back.

Here's why:

They skipped the ladder.

AI adoption isn't a tool decision.

It's a 4-phase climb.

Skip a phase and the gain disappears.

Phase 1: Foundation
Executive ownership of outcomes. Strategy tied to business goals.
Honest readiness audit before tools are bought.

Phase 2: Enablement
Clean data. Governance guardrails.
2 to 3 high-impact pilots running in parallel.

“Let’s define a data strategy.”

Translation: let’s argue about tools for 6 months.

Here’s the thing nobody wants to admit:

Most data strategies are just about data stacks.
Not actual strategy.

If you want a strategy that survives the next reorg, AI hype wave, or exec sponsor change - start here:

1. Problem: What are we solving?
2. Users: Who are you serving?
3. UVP: What's our data team's secret sauce?
4. Solution: What are we building and what are we NOT building?
5. Distribution: How will we get this in front of users without chasing them down?
   6.

AI does not fail because the technology is wrong.
It fails because the sequence is wrong.

Two companies can run the same AI program with the same vendors and the same use cases.
One scales across the enterprise.
The other shuts down in eighteen months.

The difference is decided in week one.

I have watched this pattern play out at four Fortune 500s.
The companies that succeed are not smarter.
They are not better resourced.
They make better choices earlier.

This image lays out fourteen cells.
The right column compounds into transformation.
The left column compounds into a graveyard.

Most analysts won’t lose their job to AI.

But many “SQL + dashboard” analysts will.

Right now the data world is split into two camps:

→ “AI will replace analysts.”
→ “Analysts are more important than ever.”

Both sides are partly right.

The real story:

The analyst role isn’t dying.
It’s evolving.

Knowing SQL and building dashboards was never the real value.

And with AI, it’s definitely not enough anymore.

The future analyst needs to:

• understand business context
• ask better questions
• structure messy problems
• drive decisions, not just reports…

Most companies still think AI is a technology problem.
It’s not.

PwC’s 2026 predictions remind us once again:
Only ~20% of value with AI comes from the tech.
The other 80% comes from how work is redesigned.

Yet most teams do the opposite.
I’ve seen teams with incredible tools still stuck with no measurable impact.
And others with basic tools driving real gains.

The difference?
They changed how the work gets done.

Buying feels like progress.
Pilots feel like progress.
But months later… nothing actually changes.

Redesigning work? That’s slower, messier and more uncomfortable.

Most data teams never make it to the promised land.

They’re either:

→ stuck in a bottleneck where everything goes through the data team
→ lost in a jungle of tools, chaos, and one-off solutions
→ trapped in a fortress where nothing escapes the compliance checklist

But here’s the twist:

You can set outcome-driven goals in any of these places.

In fact, you have to - otherwise, you’ll never move.

1️⃣ In the bottleneck?

Focus on reliability and trust.

• 90% of critical reports delivered by 9am
• 0% of bugs discovered by stakeholders

2️⃣ In the jungle?

Your autonomous AI might get you in jail. Even if you gave it legal guidelines.
The only thing that makes the act of hacking legal, is an authorized scope.
Mistakes against the scope may result in severe damages, financial losses and prosecution.

You are liable for the models you run. A single hallucination can ruin your career.

In the example, the model thinks the system decodes the URL and moves up a directory, which would make it in scope. However, most systems will not do this.

Here's what EVERYONE using autonomous AI for testing should do:…

Most AI initiatives look impressive at the demo.

Six months later they're quietly falling apart.

Same story every time.

The team moved fast.
Built something that looked remarkable on a slide.
Leadership approved the budget.

Then it hit the real organization.

Messy data nobody had cleaned.
No clear owner when something produced a bad output.
No way to measure whether it was actually working.
No governance to catch problems before they compounded.

The technology worked fine.

The foundation was never there.

That's not an AI failure.

I asked a client last month how many decisions their AI agent makes per day.

They said “maybe a few.”

We counted. It was over 600.

Which data source to query. How to interpret the metric. What context to include in the answer. Whether the output was confident enough to show a customer.

Every one of those is a decision. Every one of them had zero governance around it.

Nobody owned any of them. No documentation, no named reviewer, no exception process.

So I drew the levels of decision rights an AI agent actually has, and what humans need to own at each level.

“Help! I’m drowning in ad-hoc requests!”

Stakeholders ask for dashboards. You deliver.
Then they ask for another dashboard.
Then a slightly different cut.
Then export it to Excel.

You’re not building strategy. You’re running a reporting helpdesk.

Here’s a 9-question canvas that flips the script:

1. Problem
   What’s the real business pain?
   If you can’t name it in plain English, don’t build anything yet.

2. Users
   Who are you solving it for?
   No, “the business” is not a user.

3. UVP
   Why should users come to your team instead of just guessing, using gut feel, or pulling GA numbers?

AI isn’t a single technology you adopt.
It’s a set of building blocks you choose from.

AI feels sudden.
But what looks like an overnight breakthrough is really decades of layers quietly stacking up.

Thanks to my friend Luís Rodrigues for the image inspiration, it captures that story better than most explanations I’ve seen.

Here’s how we got here…

⏭️ At the bottom: Rules and logic...sometimes rebranded as Classical AI
Rigid rules. No learning.
Think eligibility checks in insurance or compliance rules in banking.
Useful, but fragile.

AI risk isn’t where you think it is.

It’s not in the boardroom.
It’s not in the policy document.

It’s in the everyday decisions no one sees.

A prompt typed into a tool.
A file uploaded with best intentions but without thinking it through.
A shortcut taken to save time.
That’s where things go wrong.

Most organisations focus on governance first.
And that makes sense.

But governance alone doesn’t change behaviour.

What works is building guardrails into how work gets done.

Four layers matter:
1/ Governance
Who owns the risk and what is acceptable…

Old security waits for alerts.
New security understands attack paths.

The shift is simple: stop chasing noise and start focusing on context, behavior, and identity before damage happens. Modern defense isn’t about more tools
— it’s about smarter visibility and proactive protection.

Great visual breakdown by Inga S.

👉 Follow Cyber Rescue Alliance for more practical cybersecurity insights that cut through the noise.

Companies winning with AI aren’t better at choosing tools.

They’re better at redesigning how work gets done.

PwC’s 2026 AI predictions put a hard number on it:
~20% of value comes from technology.
~80% comes from how work is redesigned.

Most enterprises have it reversed.

They’re spending 80% of their energy evaluating tools and running pilots, and 20% asking the only question that matters:

“How does this change the way we operate?”

Here’s the 80/20 AI strategy framework:

1/ Start With the Workflow…

I’ve spent over two decades in cybersecurity.

Alongside CISOs.
Advising boards.
Sitting in incident rooms.
And leading as a CISO myself.

And in all that time, the same feedback follows this profession everywhere.

“Too technical.”
“Speak business language.”
“The board won’t understand that.”

I’ve heard it so often it has become background noise.

And I want to challenge it.

Not reject it.
Challenge it.

And this isn’t just my view.

In mentoring conversations with CISOs
and those aspiring to step into the role,
this comes up again and again.

“Let’s define a data strategy.”

Translation: let’s argue about tools for 6 months.

Here’s the thing nobody wants to admit:

Most data strategies are just about data stacks.
Not actual strategy.

If you want a strategy that survives the next reorg, AI hype wave, or exec sponsor change - start here:

1. Problem: What are we solving?
2. Users: Who are you serving?
3. UVP: What's our data team's secret sauce?
4. Solution: What are we building and what are we NOT building?
5. Distribution: How will we get this in front of users without chasing them down?
   6.

The best data pipeline is not the most complex one. It is the right one for your workload.

Teams often copy architectures they see online, then wonder why costs rise, latency increases, or maintenance becomes painful.

The smarter question is not “What is popular?”
It is “What pattern fits my data, scale, and business need?”

Here are 8 data pipeline patterns every data team should understand 👇

1. ETL (Extract, Transform, Load)
   Transform first, then load clean data into storage.

2. ELT (Extract, Load, Transform)
   Load raw data first, transform inside modern warehouses.

Your data team delivered last quarter.

Leadership cannot name a single thing that changed because of it.

That is not a people problem. That is a process problem.

And it starts before the first line of code gets written.

Before the build:
— Does anyone know what decision this is supposed to support?

After delivery:
— Did the person receiving it understand what they were looking at?

Three months later:
— Has anything in the business actually changed because of this?

Most teams skip all three.

Then wonder why leadership stops paying attention.

If Enterprise Architecture becomes an approval gate, it slows the business down.

If it becomes a decision partner, it helps the business scale with confidence.

EA should not be where decisions go to wait, it should be where decisions go to get smarter.

The value of EA is not saying “yes” or “no.” It is helping leaders understand:

1. What risks are being created
2. What capabilities are being duplicated
3. What costs will scale over time
4. What decisions may limit future agility

That is how EA moves from governance theater to business value.

Data isn't the hard part.
Understanding each other is.

Ontology. Lineage. Semantic layers. Vector databases.

I've been in data for over 15 years, and sometimes even I feel like I'm decoding a foreign language.
We've turned simple ideas into jargon that makes non-data people tune out.

Here's what these terms actually mean and why they matter for AI:

▶️ Ontology
A shared definition of your core business concepts and how they relate.
It gives AI clear concepts to reason about instead of guessing.

▶️ Entity
A real world thing like a customer, product or event.

“Let’s define a data strategy.”

Translation: let’s argue about tools for 6 months.

Here’s the thing I don't like to admit openly:

I've been a data leader for almost 20 years.

In my first 6 years, I believed that data strategy was mostly about data stacks.

Over those years, I realized (the hard way):

If you want a strategy that survives the next reorg, AI hype wave, or exec sponsor change - start here:

1. Problem: What are we solving?
2. Users: Who are you serving?
3. UVP: What's our data team's secret sauce?
4. Solution: What are we building and what are we NOT building?
   5.

Most AI initiatives in Private Equity don’t fail because of the technology.
They fail because of 𝐡𝐨𝐰 𝐰𝐨𝐫𝐤 𝐢𝐬 𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐝.

After looking at multiple portfolio companies, a pattern emerges: a use case is identified, a tool is implemented, a team is “enabled”… and then nothing really changes.
Why?
𝐀𝐈 𝐢𝐬 𝐚𝐩𝐩𝐥𝐢𝐞𝐝 𝐭𝐨 𝐭𝐚𝐬𝐤𝐬. 𝐁𝐮𝐭 𝐯𝐚𝐥𝐮𝐞 𝐢𝐬 𝐜𝐫𝐞𝐚𝐭𝐞𝐝 𝐢𝐧 𝐰𝐨𝐫𝐤𝐟𝐥𝐨𝐰𝐬.

Take a simple example: you automate document extraction. Great.
But approvals are still manual, data isn’t reused downstream, and decisions remain inconsistent.

The real reason your data team is underperforming?

You're using them like a dashboard concierge service.

And treating strategy like a synonym for “data stack.”

Here’s how most companies use their data team:

→ To build dashboards no one uses
→ To build dbt models that collect dust
→ To chase tool-of-the-month hype
→ To enforce rules no one understands

And leaders wonder why the data team feels burned out and the business still isn’t “data-driven.”

Data teams are often treated like:

• Internal service desks taking tickets
• Janitors cleaning up tech debt
• Compliance cops for naming…

The most dangerous phase of an AI program is right after the win.

The demo worked.
Stakeholders are excited.
Everyone wants to move on to the next shiny thing.

That’s when problems start if the long term hasn’t been factored in.

AI solutions succeed when they're safe & reliable.
That means:
→ Clear decisions on what it should and should not do
→ Humans who stay accountable
→ Time (& budget) set aside for fixes, tuning, and judgement calls
→ Acceptance that edge cases happen

AI requires an ongoing operating model, not a one off investment.

Most businesses do not have an AI idea problem.

They have an execution problem.

I once stepped into a major AI programme with money behind it, internal attention, and plenty of confidence around it.

What it did not have was a clearly defined product.

There was already momentum.
Senior interest.
Big expectations.
Plenty of talking.

But when you stripped it back, the basics were missing.

No clear definition of the problem.
No shared understanding of the workflow.
No real agreement on what needed to be built first.

This is where many teams lose time.

The security terms look technical.  
The gaps behind them are governance.

Week four of MIT Applied Agentic AI for Organisational Transformation moved into security architecture.

Programme governance has to absorb every term on this list — before deployment, not after.

Fourteen terms. Half AI-native. Half classic cybersecurity. Every single one carries a governance owner — or it doesn't.

That absence is the story.

Take Phased Deployment.

The definition is clean — AI is introduced gradually, starting with low-risk applications before scaling to higher-stakes ones.

Everyone wants AI magic.
Few want to invest in the plumbing.

We live in a moment where the promise of AI
is louder than the reality of data.

I see it every week with teams.

The excitement is real. The budgets are flowing. The pilots look impressive.
But when you lift the lid a different story shows up.

Data that lives in twelve places.
Metrics that mean one thing in finance and another in operations.
Ownership that sounds like “we think it sits with them”.
Governance that feels optional.

And still we keep sprinting toward AI, hoping it will smooth over the cracks.
It never does.

The biggest myth in data teams is that data workflow is a straight line.

Data → Model → Insight → Decision

That’s not how real systems work, it only looks like that on slides.

But in reality, nothing is linear.

Before you even get to insights, you’re already stuck in definitions, broken data sources, conflicting metrics, and teams that don’t agree on what “truth” means.

Then comes modeling, where assumptions change mid-way.

Then governance, access, and compliance slow everything down.

And even after launch, adoption fails if the business wasn’t aligned on the problem from the start.

Data Leader reports to the CEO → company cares a lot about data

Data Leader reports to the COO → company cares about data

Data Leader reports to the CIO → company does not care much about data

Data Leader doesn't report to any CxO → company doesn't give a 💩 about data

Someone once said this to me and it made me laugh.

In 5 companies, I saw 5 reporting lines:

• CTO
• CFO
• CEO
• CMO
• CPO

No one knows where the Data Leader belongs.

The truth is: there is no perfect answer.

But there IS a better way to make the decision.

Everyone talks about AI models.
Very few talk about AI systems.

When you look under the hood of most AI today, you rarely find just a large language model.

You find layers.
Context. Memory. Retrieval. Tools. Autonomy.

This diagram shows the progression.

▶️ Large Language Models (LLMs) are the foundation.
They are massive neural networks trained to understand and generate human language.
They take an input and produce a response based on patterns learned during training.
Powerful, but limited to what they already know.

At the M&A Summit, organized by Melle Eijckelhoff, founder of the M&A Community Belgium, and Nancy De Beule of PwC, I had the pleasure to lead a workshop on the IT dimensions of M&A.

After years advising and supporting complex transactions, one pattern is consistent: IT is rarely the bottleneck on paper, yet often the decisive factor in value realization.
Deal readiness, carve‑outs, and integrations succeed or fail based on technology choices made far earlier than most teams expect.

🛡️ 𝗖𝘆𝗯𝗲𝗿𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀 & 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱𝘀: 𝗪𝗵𝗶𝗰𝗵 𝗢𝗻𝗲 𝗙𝗶𝘁𝘀 𝗬𝗼𝘂𝗿 𝗢𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻?

Navigating cybersecurity can be overwhelming—but frameworks and standards provide a structured path to security and compliance.

Here’s a quick breakdown of some of the most widely used frameworks:

🔐 Global & General Frameworks
✔️ ISO 27001 – Information security management across industries
✔️ NIST Framework – Widely adopted for critical infrastructure
✔️ CIS Controls – Practical, prioritized security controls
✔️ COBIT – Governance and IT management…

Two teams. Same M365 Copilot license. Same budget.
Team A deployed to 500 users, ran a satisfaction survey, called it done.

Adoption rate: 12%.

Team B deployed to 40 users, built role-specific prompts grounded in SharePoint data, measured time-to-task reduction every two weeks.
Adoption rate: 74%.

━━━━━━━━━━━━━━━━━━━━━━

⚡ 𝗪𝗛𝗔𝗧 𝗧𝗘𝗔𝗠 𝗕 𝗗𝗜𝗗 𝗧𝗛𝗔𝗧 𝗧𝗘𝗔𝗠 𝗔 𝗦𝗞𝗜𝗣𝗣𝗘𝗗:

1️⃣ 𝗠𝗮𝗽𝗽𝗲𝗱 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝘁𝗼 𝗮𝗰𝘁𝘂𝗮𝗹 𝗷𝗼𝗯 𝘁𝗮𝘀𝗸𝘀 → not generic "productivity"
2️⃣ 𝗕𝘂𝗶𝗹𝘁 𝗦𝗵𝗮𝗿𝗲𝗣𝗼𝗶𝗻𝘁-𝗴𝗿𝗼𝘂𝗻𝗱𝗲𝗱 𝗽𝗿𝗼𝗺𝗽𝘁𝘀 → per department, per role
3️⃣ 𝗠𝗲𝗮𝘀𝘂𝗿𝗲𝗱 𝗼𝘂𝘁𝗽𝘂𝘁 𝗾𝘂𝗮𝗹𝗶𝘁𝘆 → not just usage frequency
4️⃣ 𝗥𝗲𝗺𝗼𝘃𝗲𝗱 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝘄𝗵𝗲𝗿𝗲 𝗶𝘁 𝘀𝗹𝗼𝘄𝗲𝗱 𝗽𝗲𝗼𝗽𝗹𝗲…

It takes 10 minutes to fix a crash.
It takes 3 days to find a silent data quality error.

Most data architectures fail quietly.
They don't break on launch day.

They break on day 90, when nobody remembers the decision that caused it.

Here’s what that looks like in practice:

INGESTION
✕ Pull everything, filter later
✓ Validate at the edge
Bad data is cheapest to kill at entry. Let it in and it travels everywhere.
✕ No schema contract with the source
✓ Agree on types and nullability upfront
Upstream changes without a contract = your problem, not theirs.

Microsoft is slowly owning the entire enterprise AI Agent market

And most people have only seen 10% of it, here's the rest...

Most people know Microsoft Copilot. Maybe Azure. Maybe GitHub Copilot.

But Microsoft has quietly assembled one of the most complete full-stack AI ecosystems on the planet, from research models to responsible AI guardrails.

Let me break down all 7 layers:

📌 Models

• Azure GPT-5.1, GPT-5.1 Codex, Microsoft Phi-4, MAI-1, KOSMOS-2, MAI-Voice-1, Florence 2
• The foundational intelligence layer powering every product above it

📌 Frameworks…

The fastest way to understand AI governance?
Stop thinking about data. Start thinking about decisions.

There's a lot of noise around governance in the age of AI.
And leaders cut through some of that noise by understanding the split between inputs and outcomes.

Data governance protects the inputs.
It’s about accuracy, privacy, lineage, access and compliance.
It answers: Can we trust the data going into our systems?

AI governance protects the outcomes.
It deals with fairness, explainability, drift and human oversight.
It answers: Can we trust the decisions made on outputs of AI systems?

🚨 Did you know? 
Cybersecurity priorities don’t look the same in the boardroom vs. on the front line.

👔 CEOs’ top concern today: 
➡️ Cyber-enabled fraud 
➡️ AI-related vulnerabilities

🛡️ CISOs, however, are still focused on: 
➡️ Ransomware 
➡️ Supply chain resilience

This growing gap highlights a critical challenge: aligning leadership priorities with operational realities to stay ahead of evolving cyber risks.
 
📊 These insights come from the Global Cybersecurity Outlook 2026 by the World Economic Forum.

Every governance program needs data owners. Almost nobody writes down what a data owner actually does.
 
So you end up with a name on a slide and no real change.
 
I've been pulled into too many governance reviews where the data owner role exists in theory and disappears in practice. Nobody knows what they're accountable for. Nobody knows what they decide. Nobody knows when to call them.
 
Here's the job description that actually works:
 
Accountable for:
 
→ One named metric
→ Its written definition
→ Its calculation logic
 
Decisions they make:
 
→ Approves definition changes
→ Signs off…

For years, I was confused by data org models.

Centralized, hub & spoke, mesh...

Everyone had strong opinions.
But I didn't understand what actually works.

Until it finally clicked for me:

You don’t need a PhD in operating models.

👉 You need four things:

1. Define the right roles.
   Ownership and responsibilities must be clear

2. Place them intentionally.
   Some roles belong in the business, some on the data team.

3. Get the ratios right.
   A lone analyst supporting 12 teams? Burnout incoming.

4. Start centralized.
   It's way easier to decentralize once you have trust, tooling, and maturity.

It took me 6 years to realize:

Data Quality was never the thing holding back my promotion.

All these years grinding to create "perfect data".

Until I finally came to the obvious conclusion:

Executives don't care about data quality.
They only care about revenue.

That gap is where most data work dies.

I once worked with a company where:

→ Customer Acquisition Costs looked too high
→ Marketing seemed underperforming
→ Growth was slowing down

So everyone tried to “optimize marketing campaigns.”

But the real issue?

We were measuring reality wrong.

I’ve audited 100+ data teams.

Most of them had the same problems:

→ Backlogs full of “quick questions”
→ Dashboards nobody really trusts
→ Stakeholders exporting to Excel anyway

Many teams think that these are the reasons:

• “Wrong tools.”
• “Data-illiterate stakeholders.”
• “Not enough people.”
• “Lack of data culture.”

They usually aren't.

The main root cause for their problems?

They either over-optimize for control:

• Perfect models
• Heavy governance
• Centralized everything

Or they over-optimize for agility:…