Perspectives | IT infrastructure insights | ISM - A Kyndryl Company

Notes from the field.

From people who stay after the project closes. No jargon. No hype.
Just what we've learned about keeping things running.

Critical infrastructure.

What backbone infrastructure looks like when it's running the way it should. Eight parts, from foundation to finish.

One team, one pane

Article

1

When five tools all claim to be the source of truth, none of them are. Here’s what consolidation actually looks like.

One team, one pane

The costs that don't look like costs

Article

2

Infrastructure sprawl is expensive in ways that don’t always show up on a single invoice. A methodical look at where the money goes.

The costs that don't look like costs

Visibility that changes the story

Article

3

A dashboard nobody acts on is just decoration. This is about operational intelligence that actually changes what people do next.

Visibility that changes the story

Security that follows identity

Article

4

Access controls in a spreadsheet aren’t access controls. On building security architecture that travels with the user.

Security that follows identity

Operations that don’t require heroics

Article

5

When the same people keep saving the same systems, the system is the problem. Automation that removes the need for late-night phone calls.

Operations that don’t require heroics

Resilience you can prove

Article

6

Backup is not recovery. Recovery is a rehearsed sequence you can demonstrate before the incident, not a plan you read during one.

Resilience you can prove

Where the money goes

Article

7

Ten places infrastructure spend leaks. Not a list of cuts. A map of where the work goes wrong.

Where the money goes

The other side

Article

8

The final article in the series. What changes when the foundation holds: fewer surprises, less noise, more time for the work that matters.

The other side

AI vs AI.

AI gave attackers new capabilities for $20/month. Seven parts on what changed, what holds up, and the decisions that make the difference.

93% confident, 29% sure

Article

1

A solo operator jailbroke an AI chatbot and breached ten Mexican government agencies in a month. 93% of IT leaders say they're secure. 29% are prepared. ISM on what's actually missing.

93% confident, 29% sure

What 27 seconds means for your AI security posture

Article

2

AI-enabled attacks up 89%. Fastest breakout: 27 seconds. They're not hacking companies. They're hacking infrastructure. The dark hoodie hacker lost their job to AI too.

What 27 seconds means for your AI security posture

When the exposure is already credentialed

Article

3

76% of organizations have shadow AI. The threat isn't at the gate. It has a badge you issued. On vendor-embedded models, compromised plugins, and the tools your team invited in.

When the exposure is already credentialed

Sovereignty is an architecture decision

Article

4

Everyone's saying sovereign. None of them are talking about who their cloud vendor actually answers to. On jurisdiction, the CLOUD Act, and architecture that lets you move when the world shifts.

Sovereignty is an architecture decision

The people gap

Article

5

6-month contract. Four careers in one posting. The people who can do this work aren't looking for it. The staffing agencies can't reach them. There's a different way.

The people gap

What to ask your vendor

Article

6

Seven questions. Print them out. Bring them to the meeting. The demo was polished. Then someone asked where inference runs. The room changed temperature.

What to ask your vendor

What it looks like when AI security works

Article

7

Tuesday morning. The phone didn't ring. The customer didn't leave. The Reddit thread didn't get written. That's what good looks like.

What it looks like when AI security works

A different people model.

Staff augmentation isn't setup for modern projects. Hour banks instead of fixed contracts. Fit instead of credentials. Structure that moves work instead of just tracking it.

Written for the people who are over the old model.

No items found.

Making "AI" ROI happen.

The model is only as good as what you feed it. Seven parts on the data work that separates production AI from proof-of-concept.

Protection starts with a map

Article

1

5.2 million Canadians exposed in one breach. The root cause wasn’t a hack. It was a cataloguing failure. What a data catalogue actually does, and what happens without one.

Protection starts with a map

Policy on paper, reality in production

Article

2

Nova Scotia Power was storing social insurance numbers it had no reason to keep. Follow one record through four copies and three environments to see where privacy actually breaks.

Policy on paper, reality in production

Invisible until it matters

Article

3

Governance doesn’t look like anything when it works. The audit is a calendar item. Five questions governance answers before anyone has to ask.

Invisible until it matters

When the model works and the results don't

Article

4

38 million customer accounts breached through a supporting e-commerce database nobody thought to protect. Security as concentric rings, not perimeter walls.

When the model works and the results don't

When the model works and the data isn't ready

Article

5

93% of Canadian business leaders say they’re using AI. 2% are seeing a return. A readiness scorecard you can run on a Monday.

When the model works and the data isn't ready

Six costs AI budgets rarely account for

Article

6

The pilot worked. Production didn’t. Six places the money leaks before AI ships, and how to close each one.

Six costs AI budgets rarely account for

A final word from one of us

Article

7

Not a summary. Not a sales pitch. Just an honest note from someone who has watched organisations wrestle with this stuff for a long time.

A final word from one of us

About us

Where we are, how we work.

Services

The broader view, for your bigger picture

What we're capable of

Careers

Big projects. Big moves. We're looking.

Contact

We're here for you