tgroenwals shared this post · 3d ago

Clare Kitching

Most teams pour budget into dashboards and warehouses, then wonder why decisions still crawl.

It's a reminder that data doesn’t create action,
people do.

The highest leverage in any AI or data program isn’t collection or tooling.

It’s translation. The humans who turn rows into reasoning.

Here’s what I see strong leaders doing differently:
1️⃣ Funding insights and change, not just infrastructure

2️⃣ Asking “What will we change?” instead of “What does the dashboard say?”

3️⃣ Measuring the decisions made, the value implemented and not just the data gathered.

Sayed Bin Habib Almost every old company has three things in common: a beautiful dashboard almost nobody looks at, a meeting to discuss what the dashboard means, and a decision that gets made on gut instinct anyway. 3d ago 1 like

Ashutosh Panda Information only becomes valuable once it changes behavior, priorities, or action Clare Kitching 4d ago 3 likes

tgroenwals shared this post · 3d ago

Clare Kitching

Before AI can reason with your data,
your business needs to agree on what that data means and
what you want to achieve.

The data isn't the hard part but understanding each other is.

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.

Leon Jose This makes complicated data terms so easy to understand! Thank you for breaking down the jargon for the rest of us. 3d ago

Ferdinand Macagba The semantic layer point is the one I’d underline.

Most AI data problems are not really “data problems.” They are definition problems. One team says customer, another means account, another means active user, and the AI inherits the confusion.

Before reasoning improves, the business has to agree on the language the system is reasoning over. 3d ago

tgroenwals shared this post · 3d ago

Clare Kitching

AI works best when it’s used like a knife, not a miracle.

AI is starting to look a lot like digital transformation all over again.
Big promises, huge investment, hundreds of pilots. Very little impact.

And I think the problem is surprisingly simple:
Most companies misunderstand what AI actually is.

They treat it like magic when it behaves more like a tool.

I see three mindsets:
🪄 The magic wand
“AI will solve it.”

This is the hope driven approach.
Buy the platform, hire consultants, announce the transformation.

Expect productivity to appear.

Sandipan Bhaumik Absolutely Clare Kitching focusing on a few high-impact use cases is where AI actually delivers value. 5d ago 3 likes

Justin R. Nailing this analogy, Clare. Everyone loves a good Swiss army knife but no one loves trying to actually use one when it’s all in one place. Time to release the blade and get focused. 4d ago 1 like

tgroenwals shared this post · May 13

Clare Kitching

What catches most companies off guard with AI?
The speed with which small costs become massive ones.

There’s been a lot of talk about the surprisingly high cost of AI.
It seems many teams skipped the business case.

AI isn’t cheap, but the reality is nuanced.

Some AI tasks cost cents.
Some cost dollars.
Some run thousands of times a day.
Others only a few times a week.
Some save hours of work.
Some are not worth automating at all.

And the cost isn’t just driven by usage volume.

Michael Lovegrove big impact comes when you build AI with clear value in mind Clare. May 9 1 like

Robert A. Lienhard Clare, a solid recap. Balanced and insightful. Glad you brought this forward. Enjoy a great weekend ahead. 🌟👍. May 9 1 like

tgroenwals shared this post · May 11

Clare Kitching

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.

Andy Lauret ⬆️ The important question is whether AI is actually changing decisions, workflows or business outcomes measurably. Clare Kitching
May 11 1 like

Albert C. Thanks for sharing May 11 1 like

tgroenwals shared this post · May 7

Clare Kitching

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.

John Wernfeldt yes i’ve watched orgs run experiments without the foundations to fund them May 7 4 likes

Tiffany Masson, Psy.D. Strongest AI play is mixing predictable foundations with bold experiments that earn to scale. May 7 1 like

tgroenwals shared this post · Apr 30

Clare Kitching

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.

Dr.Ramyaa Ganesh Strong point—this shows up everywhere.

I’d add that even “redesigning work” isn’t enough on its own.

The real shift is redesigning how decisions get made within that work.

Because in practice, impact shows up when teams are explicit about:
– which decisions are automated vs human-owned
– what level of confidence is required to act
– and how decisions are challenged or escalated

Without that, processes may change…
but outcomes don’t.

That’s where the 80% actually gets realized. Apr 30 1 like

Dipin Kanojia Strong point: AI success depends on rethinking workflows, not tools. Impact comes when organizations redesign decisions, roles, and execution models. Clare Kitching Apr 30 2 likes

tgroenwals shared this post · Apr 27

Clare Kitching

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.

Dr. Monika Mkhitaryan, MHBA The portfolio framing is the right starting point, Clare, and the part most leaders haven't priced sits one layer beneath the diagram.

Each layer in your stack implies a different infrastructure choice. Rules and classical ML run on commodity hardware; the top of the stack does not. Generative AI runs at per-interaction cost, and agentic AI multiplies those calls by an order of magnitude per task, which is how a useful pilot becomes unviable at production scale.

In the enterprise AI projects we are brought into, the silicon beneath the software is the layer that decides which parts of the portfolio actually scale. A purpose-built inference path commonly cuts cost-per-query by more than half against the general-purpose default, and that ratio is what separates strategic capability from pilots that quietly stall. Apr 25 1 like

Batchu V Sarath Chandra Brilliant breakdown. The mistake most orgs make is trying to skip to Layer 5 (GenAI) before they’ve mastered Layer 2 (Machine Learning). If you can't predict churn with ML, you shouldn't be asking GenAI to 'solve' customer retention. AI is a portfolio, but it requires a Single Source of Truth to be effective. In engineering, we call this the Medallion Architecture—build the floor before you buy the furniture. Apr 25 1 like

tgroenwals shared this post · Apr 27

Clare Kitching

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…

Amit Gandhi The system layer is where governance either becomes real or stays theoretical. Filters, validation, default controls built into the technology, that's what governs behaviour at scale when no one is watching. Relying on people alone to enforce policy is a design choice that assumes everyone reads, remembers, and applies it under time pressure, but not really a reasonable assumption. Apr 26 2 likes

Paul Souhuwat A useful reframing Clare.

As AI accelerates how quickly answers are produced, the real challenge becomes maintaining decision discipline—knowing when to trust, when to verify, and where human judgment must remain.

Without that, speed can quietly amplify risk rather than reduce it. Apr 26 1 like

tgroenwals shared this post · Apr 22

Clare Kitching

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.

1 / 4

Rosemary Daly Lineage is the one I see fails most Clare - teams build beautiful semantic layers, then later people have trouble answering -where did this training sample actually come from - and the model’s trust story collapses. Your glossary is the vocabulary. Keeping it alive once the AI is in production is where the assurance work starts. Apr 22 2 likes

Lorphic Most AI failures aren’t intelligence problems, they’re translation problems between messy data and unclear meaning. Apr 22 2 likes

tgroenwals shared this post · Apr 21

Clare Kitching

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.

Des Raj C. Honestly the dangerous moment is when a pilot proves just enough value to get scaled before anyone has defined failure. What gets monitored, who can pause it, when does a human step in, what’s the rollback path when the edge cases stop being edge cases. A lot of teams don’t fail because the model was weak, they fail because the operating discipline showed up too late. Apr 21 3 likes

Mike Reid The real risk starts when early wins make teams forget the need for long term ownership. Apr 21 5 likes

tgroenwals shared this post · Apr 21

Clare Kitching

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.

1 / 2

Brad Wolfe Clare’s cartoon is worth a thousand slide decks.
The data truth most PE-backed LMMs are quietly avoiding: they don’t actually know what their data looks like. Not at the level of specificity that AI deployment requires.
They know roughly where the data lives. They have a general sense that the CRM and ERP don’t talk to each other cleanly. They suspect the revenue recognition logic in the billing system doesn’t match what gets reported. But they haven’t done the forensic work to quantify the gap — because that work is expensive, uncomfortable, and doesn’t show up in a board update as progress.
So the AI pilot launches on top of the mess. It looks impressive for 90 days. Then someone asks a question the model wasn’t designed to handle and the underlying fragmentation surfaces — faster and more expensively than before.
Clare’s boring foundation — clear definitions, reliable data, agreed owners, confidence in outputs — isn’t actually boring. It’s the $2M-$5M remediation investment that determines whether the AI spend that follows it is transformational or just expensive.

Brad Wolfe | wolfepacks.com Apr 9 2 likes

Pseudobytes Strong take this hits the gap most teams try to skip.
You could reply with something like: “Exactly this. AI doesn’t fix broken foundations it amplifies them.The teams getting real ROI aren’t the ones with the best models, they’re the ones who treated data like infrastructure first: defined, owned, and trusted. Everything else is just expensive iteration.” Apr 9 1 like

tgroenwals shared this post · Apr 21

Clare Kitching

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.

Paulo Segala The biggest gap isn't intelligence, it's decision architecture.

Most organizations still evaluate AI readiness based on capability: model performance, data quality, retrieval, and tooling.

But once the system starts influencing real work, the real question shifts to:
Who owns the decision?
Who steps in when the system drifts?
What escalation path is available when speed and control clash?

That’s why moving from LLMs to agents isn't just a technological upgrade. It’s a change in operational authority.

If governance, decision rights, and intervention boundaries aren’t redesigned with that shift, autonomy will outpace accountability. Apr 20 1 like

Antônio Marberger Exceptional breakdown, Clare. The industry is currently obsessed with 'Intelligence' (the LLM), but the real enterprise value lies in the Orchestration. At ai.fainow.com, we see that moving up this stack—from RAG to Agentic systems—fails if the underlying Organization of data is weak. You can't have reliable 'Autonomy' without a rock-solid Operation. Designing for trust, as you mentioned, is effectively designing for predictability. Most organizations want Agents, but they haven't yet mastered the Capture and data quality required for RAG. We have to build the foundation before we can trust the agent to act. Apr 20 1 like

tgroenwals shared this post · Apr 18

Clare Kitching

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?

Ferdinand Macagba This is a useful distinction because a lot of governance conversations still get stuck too low in the stack.

Data governance tells you whether the system was fed responsibly. AI governance tells you whether the system is allowed to influence action responsibly. Once leaders see that split clearly, it gets easier to understand why good data alone does not create safe decisions.

The harder question is usually not “is the data clean?” It is “who is accountable for the outcome if the model is wrong?” Apr 18 1 like

Selva Kumar The ownership problem is the one we run into every time. Legal, risk, engineering all hold a piece but nobody wants to be the accountable party when an autonomous system makes a decision that goes wrong.

With agentic deployments this isn’t abstract. The agent is already acting. The question of who owns the outcome needs to be answered before deployment, not after the first incident. Apr 18 1 like

Public author posts