Most enterprise AI conversations start at the wrong layer.

They start with the model. Which LLM should we use? Which copilot should we deploy? Which agent framework should we choose?

But enterprise AI does not create value because a model can generate an answer. It creates value when intelligence can understand business context, interact with enterprise systems and take the right action within the right controls.

Consider a simple business question: "Why is this supplier invoice blocked, and what should happen next?"…

Most people hear "AI".
And think it's one thing.

There are seven layers. Each one wraps around the last, like Russian nesting dolls.

Image by my friend Will Stewart, MBA , give him a follow.

𝗖𝗹𝗮𝘀𝘀𝗶𝗰𝗮𝗹 𝗔𝗜

Pure logic. “If this, then that”.

Symbolic AI and Expert Systems that follow rigid rules to make decisions.
It doesn't learn, it just follows orders.

𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴

We stopped writing rules and started using math.

Algorithms analyze data to find patterns.
They classify, predict, and optimize based on history, not hard-coded instructions.

𝗡𝗲𝘂𝗿𝗮𝗹 𝗡𝗲𝘁𝘄𝗼𝗿𝗸𝘀…

It's so easy to start from the top, with the fancy shiny stuff around agentic AI and generative AI. And that's also how most people define AI nowadays.

However, there are other things underneath that need to work.

I think it's super important to really understand the bottom thing here, the data foundation, because that's what everything is built on.

If you don't have that in place, the ownership, the definitions, the data quality, the lineage, then everything above it will break. It will start to hallucinate, and you will make decisions based on unsolid information.

Most cyber risk isn’t accepted.
It’s inherited by assumption.

Shadow IT was only the beginning….now the shadows are everywhere.

Used to be someone used an unapproved tool.

Someone spun up a system nobody knew about.

Now it’s bigger.

→ Shadow AI
→ Shadow data
→ Shadow cloud
→ Shadow access
→ Shadow procurement
→ Shadow risk acceptance
→ Shadow automation / APIs

And here’s what leadership needs to understand:

Most shadow activity doesn’t happen because people are trying to create risk.

It happens because the business is trying to move.

A team needs a tool now.

AI models are becoming a commodity.
AI systems are becoming the moat.

McKinsey: 88% of organizations now use AI.
BCG: only 5% are capturing AI value at scale.
MIT’s research was even sharper: 95% of enterprise GenAI pilots show no measurable P&L impact.

The gap is not intelligence.
It is architecture.

Most companies are buying Ferrari engines and installing them
in horse carriages.

Open almost any serious AI product.
You will not find a model working alone.
You will find a system:

Retrieval.
Memory.
Tools.
Permissions.
Workflows.
Monitoring.
Evaluation.
Governance.

The technology almost always gets delivered.

The transformation almost never does.

From what I have seen, most organisations cannot wait to announce a 'go-live'.

But then spend months figuring out why they aren't seeing the benefits that they thought they had paid for.

In almost every one of these situations, this could have been avoided.

Here's how:

1. The outcome is defined before the work begins
   ↳ Agree what success looks like before selecting technology or vendors. Every trade-off should be measured against it.

From PMO to TMO to VMO

PMO
-Focus: Deliver Projects
-On time
-On budget
-Within scope
-Milestone driven
-Measures execution

TMO
-Focus: Drive Transformation
-Enable adoption
-Manage organizational change
-Build new capabilities
-Align stakeholders
-Measures transformation readiness

VMO
-Focus: Realize Value
-Define value upfront
-Track value realization
-Govern investment decisions
-Align strategy to execution
-Measure customer value and business value

The Evolution
PMO: Did we deliver the project?
⬇️
TMO: Did the organization adopt the change?
⬇️
VMO: Did we create the value we promised?

87% of executives are investing in AI without understanding what they're actually buying:

They're throwing millions at "AI solutions" while their teams can't tell machine learning from deep learning.

Here are the 6 AI model types that you need to know:

1/ Machine Learning Models:
↳ Learn from historical data to predict what happens next.
↳ Cut forecasting errors by 40% in sales, inventory, and risk management.
↳ Think: Netflix recommendations that keep users watching longer.

2/ Deep Learning Models:
↳ Brain-inspired systems that find patterns humans miss in massive datasets.

The board does not need more cyber detail
It needs business clarity.

Not because leaders don’t care about security…because they’re hearing it in the wrong language.

Security talks about controls.
Finance talks about exposure.
Operations talks about downtime.
The CEO talks about business risk.

Too many cyber plans stay inside the security function.

The result: security speaks in tools and vulnerabilities while the business is thinking about margin, continuity and growth.

That gap slows decisions.

🧙🏼‍♂️ CROWN gives leadership teams a shared lens for cyber planning.

A great way to test if your governance is working is just to ask: which are the decisions that have been made in the past 90 days, and where are those documented?

If those are not documented anywhere, and you can't trace back who made them and why, then you don't really have a clear operating model around data governance.

Decision authority is the most important part here, because without that everything else will fail.

And if you're not making decisions then you will never be able to derive the tangible outcomes that you would like to.

Your AI governance policy exists.

Your organization's behavior doesn't match it.

McKinsey just put a number on the gap.

Enterprise Responsible AI maturity averages 2.3 out of 4 in 2026, up from 2.0. But fewer than a third of organizations have reached level 3 in the dimensions that matter most: strategy, governance, and agentic AI controls.

Most executives think of "AI governance readiness" as binary: you either have it or you don't. That's the wrong frame.

Readiness is a ladder, and most organizations are standing on a lower rung than the board deck suggests.

2023: Everyone bought AI.
Almost no one changed the work.

For the last three years, most people used AI
↳ To draft faster.
↳ To summarize documents.
↳ To automate repeatable tasks.

They said people are still needed for decisions on the bubble.
Those that required seeing across systems.
Edge cases that needed judgment.

Across BCG, Deloitte, McKinsey and Anthropic, a pattern is emerging.
Automation is not enough.

Not anymore. Agentic AI changes the map.

In most companies, it takes a custom IT project to make 17 different systems talk.

The ENTERPRISE AI GUIDE in 1 page 📌

Leading AI in a business is a big job.

So I put my 9 most popular enterprise AI frameworks in one place:
📌 Get it free here: https://lnkd.in/gwtgqA3H

Start with what AI is:
👉 Types of AI: agents, generative, deep learning, classical ML and rules, and which one solves your problem

Then set strategy and value:
👉 The AI Strategy Wheel: 7 decisions every AI strategy has to make
👉 The AI Value Funnel: where value gets lost between interest and impact
👉 The Total Cost of AI Ownership: the costs your business case doesn't show…

This map of agentic AI is probably one of the best ones I've seen.

But have a look at where the arrow starts.

AI and ML, then deep learning, then gen AI, then agents, then full automation.

Five rings, around sixty labels, and none of them say who owns the data going into the first one.

Governance turns up twice, on two small stickies out near rollback mechanisms and retention policies.

That's more or less how it looks inside the organisations I walk into as well.

Most organisations that say they’re data-driven are not.

They’re data-aware. At best.

✅ 10 signs that separate the real thing from the performance:

1. Leaders change decisions because of data

Not just cite data to support decisions already made. Data that changes a direction is culture. Data that decorates one already set is not.

2. When the data contradicts the instinct, the instinct gets questioned

Not explained away. Investigated.

3. People closest to the data have a path to the decision-makers

Analysis that cannot reach the room where decisions happen is not part of the culture.

Most people using AI daily can't explain 12 basic terms.

Not obscure ones, the ones that show up in every product announcement, every board deck, every job description.

I picked 12 that matter right now.
Not 12 for ML engineers.

12 for anyone who builds, buys, or decides anything involving AI in 2026.

Grouped them into four buckets:

→ The Foundations: LLM, Token, Context Window

The building blocks, what AI is, how it reads, and how much it can hold in its head.

→ How AI Learns: Fine-tuning, RAG, Hallucination

How models get smarter, stay grounded, and where they still fail.

Spot on!

AI, ML, GenAI and LLMs are NOT the same thing.
Here's the difference, in plain English:

So here's the simplest way to think about it...
They're not competing tools.
They're a hierarchy that keeps building on itself:
(See the visual for the full breakdown 👇)

1️⃣ AI - Artificial Intelligence
↳ The broad umbrella. Systems that can think, reason, and act intelligently.
↳ Examples: expert systems, robotics, game AI.

2️⃣ ML - Machine Learning
↳ A subset of AI that learns patterns from data.
↳ Examples: spam detection, recommendations, fraud detection.

Adding a human next to the AI is not oversight.
Regulators do not care that your AI is fast.
They care who signed off.

EU AI Act Article 14. NIST AI Risk Framework.
Colorado AI Act. Texas TRAIGA.

All four require human oversight with three elements.
Context. Authority. Rationale.

Most governance frameworks skip all three.
They add a checkbox. Not a checkpoint.

Here is how to deploy Human-in-the-Loop across the 5 governance layers.

→ Ethics:
Named reviewer with real veto authority.

→ Risk Management: Tiered checkpoints.
High-risk pauses. Low-risk runs.

Your AI can pass every benchmark and still fail the only question that matters.

"𝗦𝗵𝗼𝘄 𝗺𝗲 𝘄𝗵𝗲𝗿𝗲 𝘁𝗵𝗶𝘀 𝗱𝗮𝘁𝗮 𝗰𝗮𝗺𝗲 𝗳𝗿𝗼𝗺."

Months in production.
A room full of smart people.
Nobody could answer.

The foundation was never built.
An auditor just happened to find it first.

The pattern inside regulated companies is predictable.

The board asks about the outer ring:
agents, high-risk decisions, regulated use cases.

The budget follows the board.
The failure lands in the inner ring:
source registry, lineage, ownership.

Where no one was watching.

I call this the 𝗔𝗜 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗚𝗮𝗽.

AI governance is getting the attention.
Data governance is doing the work.

Right now, companies everywhere are rushing to scale AI.

But underneath the excitement is a quieter question:
Can we actually trust the data feeding these systems?

Because AI is only as good as the foundation beneath it.

If nobody knows:

• where the data came from
• who owns it
• whether it’s accurate
• whether it’s current
• whether sensitive information is buried inside it
  ...then AI will scale this confusion.

A complete guide to CLAUDE.md, custom commands, skills, agents, and permissions, and how to set them up properly.

Most Claude Code users treat the .claude folder like a black box. They know it exists. They've seen it appear in their project root. But they've never opened it, let alone understood what every file inside it does.
That's a missed opportunity.
The .claude folder is the control center for how Claude behaves in your project. It holds your instructions, your custom commands, your permission rules, and even Claude's memory across sessions. Once you understand what lives where and why, you can configure Claude Code to behave exactly the way your team needs it to.
This guide walks through the entire anatomy of the folder, from the files you'll use daily to the ones you'll set once and forget.

Two folders, not one

Before diving in, one thing worth knowing upfront: there are actually two .claude directories, not one.

Everyone treats AI governance as one conversation.
It's six (or more), and they're happening in different rooms.

The CEO worries about moving fast without a blow-up they have to explain to the board.

The CFO asks what getting this wrong costs, and what they're actually funding.

The CIO starts counting what's already running, who owns it, and what it can touch.
All three are right. They're just not in the same conversation.

Here's how governance reads from each seat:

🔵 CEO: I want permission to say yes. 72% of CEOs now say they're the main AI decision maker (BCG, AI Radar 2026).

We worry about choosing the wrong AI tool or provider.
But it's generally the most reversible decision on the list.

With an AI strategy, what sometimes arrives is a tool decision, a pilot, and a slide about data quality.

That's a good start, but you're missing some important parts.

An AI strategy that scales covers at least seven dimensions.

Four that usually go missing can be the tough ones:

• Ambition owned at CEO level and tied to the goals the board already tracks.

• An operating model that sets who approves new initiatives and who keeps the live ones running.

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
→…

I started data projects with 3 silly questions.

Today, I skip all of them:

❌ “Which BI tool should we use?”
❌ “Should we build a real-time pipeline?”
❌ “What should our data architecture look like?”

The result:

I built a dashboard factory and lost trust with senior management.

That’s why I now use the same 5-step process to kickstart every data project.

Simple.
Repeatable.
Business-first.

1️⃣ Understand the game

• Where is the company today?
• Where do we want to be in 12 months?
• Which initiatives already exist?

Most AI programmes never reach transformation.
They stop at Level 1.

Everyone wants AI to deliver business value.

But many organisations are measuring success
by faster tasks and lower costs.
They stay at Level 1.

That’s not transformation.
It’s improvement.

AI creates value at three different levels.

Level 1 improves tasks.
People work faster and costs fall.
But the business stays largely the same.

There’s nothing wrong with Level 1.
It's useful, and every organisation
should look for quick wins.

The problem is believing that’s enough.
Believing it’s transformation.

Most companies think the biggest challenge with AI agents is intelligence.

It is not.

The reality?
🚫 Powerful agents with weak security create massive risk
🚫 One permission mistake can expose critical data
🚫 Automation without oversight increases vulnerability
🚫 Trust becomes the real competitive advantage

Here are 7 ways smart companies build secure AI agents:

1. Limit Access by Design
   ↬ Give agents only the permissions they need
   ↬ Role-based access reduces unnecessary risk

2.

Your project closed green.
So why does nothing feel different?

Most leaders have lived this.

The milestones hit. The budget held.
The steering committee applauded.
Yet nothing fundamental changed in the business.

Here's the uncomfortable truth.
You delivered exactly what
you were asked to deliver.

Projects deliver outputs.
Outcomes and benefits only arrive
when the business changes how it works.

That's not my definition.
PRINCE2 and the PMBOK both say it.

And there's a deeper pattern.
Most projects, not all, are focused on optimisation.
They make the existing business more efficient.

This is the version I now print and walk into AI strategy meetings with.

Last week, forty minutes into a diagnostic call, a CDO stopped me.

She told me her entire customer analytics pipeline depended on one Excel file maintained by one person.

This kind of happens on almost every call I have run this year.

The team always wants to start at the AI layer. I keep pulling the conversation down the stack, and we usually do not make it past Operational Reality.

Excel files, Slack threads, one person who knows how it works.

That is where AI in production fails, not at the model layer.

𝗗𝗼𝗻'𝘁 𝘁𝗿𝘂𝘀𝘁 𝘁𝗵𝗲 𝗵𝘆𝗽𝗲. 𝗪𝗲 𝗱𝗼𝗻'𝘁 𝗮𝗹𝘄𝗮𝘆𝘀 𝗻𝗲𝗲𝗱 𝗮 𝗳𝗲𝗱𝗲𝗿𝗮𝘁𝗲𝗱 𝗺𝗼𝗱𝗲𝗹.

Everyone seems to recommend federated governance. Over the last few years, it has become the model everyone talks about. But that doesn't mean it's the right choice for every organization.

The right model depends on your maturity, your organization and the problems you're trying to solve.

A 𝗰𝗲𝗻𝘁𝗿𝗮𝗹𝗶𝘇𝗲𝗱 𝗺𝗼𝗱𝗲𝗹 usually works best when governance is new, regulations are strict, data quality is inconsistent or ownership is still unclear. It helps create common standards, shared definitions and consistent ways of working.

𝗠𝗼𝘀𝘁 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗱𝗼𝗻'𝘁 𝗵𝗮𝘃𝗲 𝗮 𝗱𝗮𝘁𝗮 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆... 𝗧𝗵𝗲𝘆 𝗵𝗮𝘃𝗲 𝗮 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 𝗿𝗼𝗮𝗱𝗺𝗮𝗽.

I often see organizations investing in cloud platforms, AI and modern data stacks before answering a much simpler question.

What business problem are we trying to solve?

A data strategy defines where the company wants to create value with data. Data governance makes that strategy executable.

Without governance, priorities stay unclear.
↳ Ownership is missing.
↳ Data quality depends on individual effort.
↳ Teams move in different directions.

Small businesses can win with AI faster than big companies.

These 6 simple steps will help you start.

Most people think AI belongs to large companies.

Big budgets. Big teams. Lots of meetings where nothing gets decided.

I know this because I've worked 10+ years in corporate.
I've been in situations where we took 6 weeks to approve a tool to collect data to improve a product.

But small businesses have an unfair advantage.

They can move faster.

You do not need a huge AI strategy plan.

You need a clear place to start.

Here is the simple roadmap I would use.

Two firms have access to the same AI.
By 2030, their positions will have diverged.

Same AI. Two different futures.

Most firms point AI at the present.
They apply it to what already exists
then call it transformation.

But the truth is, it's optimisation.
Valuable, yes.
But it's not transformation.

Productivity. Efficiency. Automation.
All real. All worth having. All challenging.

But optimisation has a ceiling.
It makes the current model better, not different.
The game stays the same.

And it follows a familiar curve.
Refine. Plateau. Diminishing returns.

What makes data governance work?

Clear ownership
Business value
Better decisions
Continuous improvement

...and more.
These 10 guiding principles help turn governance into practical, everyday work that people can understand, support, and sustain.

Which principle has made the biggest difference in your organization?

♻️ Share this with someone building or strengthening a data governance program.
Follow here for more George Firican

hashtag
#datagovernance
hashtag
#datamanagement
hashtag
#data

𝐖𝐡𝐚𝐭 𝐬𝐞𝐩𝐚𝐫𝐚𝐭𝐞𝐬 𝐬𝐜𝐚𝐥𝐚𝐛𝐥𝐞 𝐝𝐚𝐭𝐚 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦𝐬 𝐟𝐫𝐨𝐦 𝐞𝐱𝐩𝐞𝐧𝐬𝐢𝐯𝐞 𝐝𝐚𝐭𝐚 𝐩𝐫𝐨𝐣𝐞𝐜𝐭𝐬?

It’s rarely the technology.

It’s the architecture decisions made before the first pipeline is built.

Many organizations invest heavily in cloud platforms, analytics tools, and AI initiatives-yet struggle with fragmented data, inconsistent metrics, and slow delivery.

The reason is simple:

Different business needs require different warehouse design patterns.

𝐇𝐞𝐫𝐞 𝐚𝐫𝐞 6 𝐜𝐨𝐦𝐦𝐨𝐧 𝐃𝐚𝐭𝐚 𝐖𝐚𝐫𝐞𝐡𝐨𝐮𝐬𝐞 𝐃𝐞𝐬𝐢𝐠𝐧 𝐏𝐚𝐭𝐭𝐞𝐫𝐧𝐬 𝐞𝐯𝐞𝐫𝐲 𝐝𝐚𝐭𝐚 𝐥𝐞𝐚𝐝𝐞𝐫 𝐬𝐡𝐨𝐮𝐥𝐝 𝐮𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝:…

Your biggest AI breach probably won’t trigger an alert.

It will happen quietly.

And by the time someone notices, your data may already be somewhere it should never have been.

AI isn’t the biggest risk.

Careless AI usage is.

I’ve seen companies spend millions on cybersecurity…

…then copy confidential information into an AI tool without thinking twice.

No malware.
No phishing email.
No ransomware.

Just one prompt.

Here are the AI risks most leaders ignore:

• Employees paste customer data into public AI tools.

Most people think Agentic AI is just "ChatGPT + tools"

They're wrong.

This diagram shows the full stack.

Five layers. Each one matters.

Most failures happen in layers four and five. Not in the models.

𝟭/ 𝗔𝗜 & 𝗠𝗟 (𝗧𝗵𝗲 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻)
Data becomes decisions here.

Classical learning. Supervised, unsupervised, reinforcement.

𝟮/ 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 (𝗧𝗵𝗲 𝗘𝗻𝗴𝗶𝗻𝗲)
Neural networks. Transformers.

The machinery learns patterns at scale.

𝟯/ 𝗚𝗲𝗻𝗔𝗜 (𝗧𝗵𝗲 𝗖𝗿𝗲𝗮𝘁𝗶𝘃𝗲)
LLMs generate content. RAG pulls context.

Multimodal models handle text, image, audio, video.
The output layer.

$600 Mac mini + Kimi API = personal AI server that runs 24/7. Here's how to turn it into a business

http://x.com/i/article/2066836498851201024

Most people think about AI as a chat window they open when they need something. The developers making real money with AI think about it differently - as infrastructure that runs continuously, handles client requests automatically and generates revenue while they sleep.
The hardware that makes this possible isn't a $10,000 GPU server or a cloud subscription that bills you by the hour.

Most companies I work with have one of these four quadrants and call it data governance.

The most common version has artifacts. A catalog, a few metric cards, a policy document somewhere on SharePoint. Other companies build out roles instead. A CDO with no peer at her level, a Gov Council that meets quarterly, a DPO who is mostly there for the regulator. A few do concepts well, with decision rights on a slide and named ownership on another, both signed off by someone who left the company. Almost everyone has use cases queued up: AI pilots running on data the team does not trust.

Most governance business cases die in the rewrite, before the meeting even happens.

The pitch gets softened in the 48 hours before the meeting, the meeting happens anyway, and the funding question gets pushed to next quarter.

I have watched €2.4M in concrete GDPR exposure get rewritten as "significant regulatory risk" in that window. The pitch did not get worse exactly, but the teeth got pulled.

The image has five rows that decide whether a governance program gets funded or filed. Every row needs a number and a name attached.

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…

A frontier without an ecosystem is not stable

http://x.com/i/article/2065582894790365184

I’ve been thinking a lot about the future of the firm in an AI-driven economy.
This transition is different than any previous platform shift. In the past, we used digital systems to enhance human capital. This is the first time we can create a real cognitive loop between people and digital systems. That is a mind-bender, because it changes how we even conceptualize work inside an enterprise.

Your AI spend per employee keeps climbing.

At what point does it cost less to keep the human in the loop?

CFOs are starting to run that math.

For the past two years, the pitch was simple:

AI replaces headcount. Headcount is expensive.

So AI saves money.

That logic is starting to wobble.

Not because AI does not create value.

Because the cost model is more complicated than the sales deck made it sound.

Here’s what’s actually happening:

1/ The swap is becoming explicit

→ CFOs are prioritizing AI and technology investment while headcount growth slows
→ AI budgets are increasingly…

▶️ The AI ROI Fallacy: Unmeasured Infrastructure in Turnaround Assets

David’s Bridal recently claimed asking for AI ROI is “the wrong question” because AI is now “infrastructure.”

For a legacy retail brand fresh out of multiple bankruptcies, that should set off alarms for their board and private equity sponsors.

Infrastructure still consumes capital. Capital has a cost. Warehouses, ERP upgrades, data platforms, and governance programs all require a business case. AI doesn’t get a pass.

The TravelAI Control Stack - A Unified Portfolio of Enterprise AI Infrastructure Assets
Private equity does not acquire software.

It acquires control.

Enterprise value expands when decision ownership is preserved from board mandate through operating execution.

Most AI companies are building applications.
This is control infrastructure.

I have built and positioned a full portfolio as a cohesive TravelAI Control Stack for scaled travel and hospitality platforms.

• TravelAir.ai - Master control layer and intelligence core. Anchors decision coordination across the entire ecosystem.

The AI problem inside most PE-backed and ambitious service businesses is not lack of interest.

It is too much choice.

Too many tools.
Too many internal experiments.
Too many people saying “we should probably be doing something with AI.”

For founders and Management teams, that makes it easier to delay than to choose. The problem with too much choice is that it pushes teams into tool decisions before they have made a value decision.

I spoke recently to a leader at a PE-backed businesses, and they put it plainly.

Everyone in private equity can now see the AI opportunity across the portfolio. Now the real work is choosing the right workflow, getting access to the right systems, redesigning the process, measuring the output, and getting the team to adopt a new way of operating.

This is where a lot of AI initiatives get stuck. A portfolio company might have 20 plausible use cases: engineering velocity, customer support, finance operations, professional services automation, compliance, reporting, product modernization. But not all of them are good first projects. Some are commercially irrelevant.

A recent conversation with a CHRO highlighted a common concern among HR leaders:

“We don’t have millions of dollars to throw at AI initiatives in HR like big tech companies do.”

This concern is valid. Many organizations do not have access to unlimited AI budgets, extensive product teams, or well-prepared enterprise data. However, implementing AI does not have to begin with a large-scale program.

A valuable lesson from private equity is to focus on value first.