<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>AIorDie</title>
    <link>https://aiordie.com</link>
    <atom:link href="https://aiordie.com/rss.xml" rel="self" type="application/rss+xml" />
    <description>AIorDie is where Jason Wojahn writes about what stops being true next — arguments about what AI changes, breaks, and reprices.</description>
    <language>en-us</language>
    <lastBuildDate>Thu, 10 Sep 2026 22:25:23 GMT</lastBuildDate>
    <item>
      <title>The Screen Can Shrink. Governance Cannot.</title>
      <link>https://aiordie.com/arguments/the-screen-can-shrink-governance-cannot</link>
      <guid isPermaLink="true">https://aiordie.com/arguments/the-screen-can-shrink-governance-cannot</guid>
      <pubDate>Tue, 08 Sep 2026 12:00:00 GMT</pubDate>
      <category>Argument</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>The strongest objection to the disappearing application is also the reason the execution layer becomes valuable.</description>
      <content:encoded><![CDATA[There is a fair criticism of the argument that AI moves work above the application. Enterprises still need control.

Someone still needs to know:

What happened. Why it happened. Who authorized it. Which policy applied. What state existed before the action. What changed afterward. Whether the action can be reversed.

That criticism is right. But I think it supports the thesis rather than undermining it.

The screen can shrink. Governance cannot.

In fact, as AI gains more authority to act, governance becomes more valuable. The mistake is assuming governance and interface have to live in the same place. They do not.

ServiceNow can remain the authoritative source for configuration and policy. Salesforce can remain the customer system of record. SAP can remain the transactional backbone.

Another layer can still own the interaction and coordinate the work across all three. That creates a more interesting architecture.

Probabilistic intelligence at the top. Deterministic systems underneath. A governed execution layer between intent and action.

The winners may not simply own the workflow. They may own the trusted transition between what the AI believes should happen and what the enterprise allows to happen.

That layer needs:

Identity. Permissions. Policy. State. Observability. Audit. Rollback. Human escalation.

This is why “AI replaces enterprise software” is the wrong argument. AI rearranges where value lives inside the enterprise stack. Some parts become less visible. Others become much more important.

The UI can disappear from a workflow. Accountability cannot. And the more software becomes the work, the more important that distinction becomes.]]></content:encoded>
    </item>
    <item>
      <title>The Workflow Escaped the Application</title>
      <link>https://aiordie.com/arguments/the-workflow-escaped-the-application</link>
      <guid isPermaLink="true">https://aiordie.com/arguments/the-workflow-escaped-the-application</guid>
      <pubDate>Tue, 01 Sep 2026 12:00:00 GMT</pubDate>
      <category>Argument</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>Once intelligence can coordinate across systems, the boundaries of the application stop being the boundaries of the work.</description>
      <content:encoded><![CDATA[The ServiceNow argument got attention. But ServiceNow was never really the point.

The more interesting question is what happens when an application no longer owns the user.

Enterprise work does not actually respect application boundaries. A customer problem may touch Salesforce, ServiceNow, SAP, email, Slack, a data platform, a contract, and a person.

The old software architecture forced the human to become the integration layer.

We logged into one system. Found the information. Moved to another. Interpreted the context. Updated something. Sent a message. Opened another record.

The workflow was technically distributed but cognitively assembled by a person.

AI changes that. An intelligence layer can increasingly assemble context across those systems and coordinate the execution itself.

That means workflow begins escaping the application. And when it does, the platform battle moves upstream.

The question becomes:

Who owns the context? Who understands the policy? Who knows the desired state? Who decides what happens next? Who has permission to make it happen? Who maintains authoritative state? Who can prove what happened afterward?

This is why I do not think the future is simply AI replacing SaaS. That is too crude. The systems underneath the enterprise remain essential. But their strategic role can change.

A system of record can be immensely valuable without being the primary interface. A transaction engine can be indispensable without determining the workflow. An application can survive while its old moat disappears.

For decades we built systems of record. Then systems of engagement.

I think the next fight is over systems of execution. And this time, the boundary may not be an application at all.]]></content:encoded>
    </item>
    <item>
      <title>The application owns the user.</title>
      <link>https://aiordie.com/dead-ideas/dead-idea-the-application-owns-the-user</link>
      <guid isPermaLink="true">https://aiordie.com/dead-ideas/dead-idea-the-application-owns-the-user</guid>
      <pubDate>Tue, 25 Aug 2026 12:00:00 GMT</pubDate>
      <category>Dead Idea</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>The most valuable enterprise application may become invisible to the person using it.</description>
      <content:encoded><![CDATA[For more than 20 years, enterprise software companies competed to become the place where work happens.

CRM. ITSM. ERP. HCM.

Own the record. Own the workflow. Own the screen. Own the user.

I think that chain is breaking.

Increasingly, I will not open five applications to get something done. I will tell an intelligence layer what outcome I want.

It will understand the request. Gather context. Apply rules. Determine the workflow. Work across the necessary systems. Execute.

Salesforce may still hold the customer record. ServiceNow may still hold service data, configuration, permissions, and workflow logic. SAP may still run the transaction.

Those systems remain extremely important. But importance does not guarantee ownership of the interaction.

What dies

The assumption that owning the application automatically means owning the user.

What survives

Systems of record. Authoritative data. Business rules. Permissions. Governance. Transactions. Auditability.

Where the value moves

Upstream. Toward whoever understands intent, context, workflow, policy, and what needs to happen next.

That does not necessarily kill the application. Something more uncomfortable can happen. It becomes plumbing.

Critical. Expensive. Everywhere.

But no longer where the user lives.]]></content:encoded>
    </item>
    <item>
      <title>The SDLC Is Becoming Agent-Executed</title>
      <link>https://aiordie.com/signals/signal-sdlc-is-becoming-agent-executed</link>
      <guid isPermaLink="true">https://aiordie.com/signals/signal-sdlc-is-becoming-agent-executed</guid>
      <pubDate>Tue, 18 Aug 2026 12:00:00 GMT</pubDate>
      <category>Signal</category>
      <category>Work</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>Coding agents are not simply making developers type faster. They are changing what software engineering work looks like.</description>
      <content:encoded><![CDATA[What happened

Watch what is happening in software development.

The first AI coding tools completed lines. Then functions. Then files.

Now agents increasingly plan work, modify systems, run tests, diagnose failures, iterate, and complete larger pieces of the development lifecycle.

Why it matters

The developer’s job starts moving upward.

Direction. Architecture. Context. Judgment. Review.

That should matter far beyond engineering. Software development may simply be one of the first knowledge-work domains where the pattern is visible enough to measure.

Human directed. Agent executed. Human governed.

If that model works in software, expect every other function with structured digital work to study it.

Finance. Operations. Marketing. Service. Legal. IT.

What to watch

Watch whether engineering organizations start redesigning team structure and workflow rather than simply buying more coding assistants. That is when the change becomes structural.]]></content:encoded>
    </item>
    <item>
      <title>Software Is Becoming the Work</title>
      <link>https://aiordie.com/arguments/software-is-becoming-the-work</link>
      <guid isPermaLink="true">https://aiordie.com/arguments/software-is-becoming-the-work</guid>
      <pubDate>Tue, 04 Aug 2026 12:00:00 GMT</pubDate>
      <category>Argument</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>The important shift is not that software got smarter. It is that software can increasingly act.</description>
      <content:encoded><![CDATA[For decades, software helped us understand work.

Databases stored it. Applications organized it. Dashboards displayed it. Analytics interpreted it. Workflow platforms moved it.

People still did it.

AI changes that boundary.

Software can increasingly interpret intent. Understand context. Determine a path. Take action. Evaluate what happened. Continue.

That turns software from something beside the work into something participating directly in the work.

And once that happens, the competitive question changes. It is no longer simply what your company knows. It becomes how quickly and reliably your company can turn what it knows into an outcome.

That sounds subtle. It is not.

It changes software architecture, because intelligence needs access to context, rules, systems, and actions. It changes services, because execution no longer scales only through labor. It changes organizations, because humans do not need to coordinate every intermediate step. It changes measurement, because activity becomes less interesting than completed outcomes. It changes strategy, because owning a feature is less defensible than owning a meaningful execution loop.

I do not think humans disappear from this model. Judgment becomes more important. Direction becomes more important. Accountability becomes more important.

But much of the mechanical distance between decision and execution starts disappearing.

For 30 years, technology helped humans do work. The next era is different.

Software is becoming the work.]]></content:encoded>
    </item>
    <item>
      <title>Productivity is the goal.</title>
      <link>https://aiordie.com/dead-ideas/dead-idea-productivity-is-the-goal</link>
      <guid isPermaLink="true">https://aiordie.com/dead-ideas/dead-idea-productivity-is-the-goal</guid>
      <pubDate>Tue, 21 Jul 2026 12:00:00 GMT</pubDate>
      <category>Dead Idea</category>
      <category>Work</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>Saving ten hours is useful. Removing ten days from the path between decision and outcome is different.</description>
      <content:encoded><![CDATA[Most companies measure AI using productivity.

Hours saved. Tasks completed. Code produced. Tickets resolved.

Those are useful numbers. I am not sure they are the most important ones.

If every employee becomes 20 percent faster but work still waits for:

A meeting. An approval. Another function. A committee. A handoff. A quarterly planning cycle.

Then the company may have created productivity without creating much organizational advantage.

What dies

Productivity as the ultimate measure of AI value.

What survives

Efficiency still matters. Cost still matters. Individual productivity still matters.

Where the value moves

Toward execution throughput. How quickly can the organization move from idea, to decision, to coordinated action, to measurable outcome?

The competitive advantage of the next decade may not be which company has the most intelligence. Everyone will have intelligence.

It may be which company can convert intelligence into action fastest without losing quality, governance, or accountability.]]></content:encoded>
    </item>
    <item>
      <title>AI Is Moving From Assist to Execute</title>
      <link>https://aiordie.com/signals/signal-ai-moving-from-assist-to-execute</link>
      <guid isPermaLink="true">https://aiordie.com/signals/signal-ai-moving-from-assist-to-execute</guid>
      <pubDate>Tue, 07 Jul 2026 12:00:00 GMT</pubDate>
      <category>Signal</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>The language around enterprise AI is changing because the capability boundary is changing.</description>
      <content:encoded><![CDATA[What happened

For the first phase of generative AI, the verbs were mostly familiar.

Write. Summarize. Search. Recommend. Assist.

Now the verbs increasingly sound different.

Plan. Route. Investigate. Resolve. Build. Test. Deploy. Execute.

Why it matters

That change in language reflects a change in what the systems are expected to do.

The first generation of enterprise AI sat beside the workflow. The emerging generation increasingly participates inside it.

This does not mean autonomous systems should suddenly receive unlimited authority. Governance becomes more important as execution increases. But the architectural direction is hard to ignore.

We are moving from human executed, AI assisted — toward human directed, system executed.

What to watch

Watch the verbs vendors use over the next year. They may tell you more about the future of enterprise software than the feature list does.]]></content:encoded>
    </item>
    <item>
      <title>An AI strategy is enough.</title>
      <link>https://aiordie.com/dead-ideas/dead-idea-an-ai-strategy-is-enough</link>
      <guid isPermaLink="true">https://aiordie.com/dead-ideas/dead-idea-an-ai-strategy-is-enough</guid>
      <pubDate>Tue, 19 May 2026 12:00:00 GMT</pubDate>
      <category>Dead Idea</category>
      <category>Organization</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>Most companies do not have an AI problem. They have an operating-model problem that AI is exposing.</description>
      <content:encoded><![CDATA[Almost every large company now has an AI strategy.

Pilots. Models. Copilots. Training. Governance committees. Innovation funds.

Yet surprisingly little about the organization itself changes. That should tell us something.

The technology is advancing faster than the operating model wrapped around it.

Companies were designed for expensive intelligence.

Information moved slowly. Expertise was scarce. Managers coordinated people. Meetings synchronized decisions. Approval chains controlled risk. Departments accumulated specialists.

AI compresses many of those constraints.

But most companies simply insert it into the existing structure. Then they wonder why the result feels incremental.

What dies

The belief that adopting enough AI eventually creates an AI-native company.

What survives

Strategy. Governance. Leadership. Accountability. Human judgment.

Where the value moves

Toward organizations willing to redesign how decisions, workflows, teams, and authority actually operate.

The important question is no longer what our AI strategy is. It is what work should no longer happen the way it happens today.

That is a much harder conversation. It is also where the value is.]]></content:encoded>
    </item>
    <item>
      <title>Headcount is scale.</title>
      <link>https://aiordie.com/dead-ideas/dead-idea-headcount-is-scale</link>
      <guid isPermaLink="true">https://aiordie.com/dead-ideas/dead-idea-headcount-is-scale</guid>
      <pubDate>Tue, 05 May 2026 12:00:00 GMT</pubDate>
      <category>Dead Idea</category>
      <category>Services</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>For services businesses in particular, this assumption is about to get expensive.</description>
      <content:encoded><![CDATA[For most of my career, scaling a technology services company had a fairly predictable relationship with people.

More demand. More people. More delivery capacity. More revenue.

The best operators improved utilization, mix, geography, pricing, and leverage. But the basic equation survived.

AI begins changing the denominator.

If an engineer operating with agents can design, build, test, document, and iterate materially faster, adding a person is no longer the only way to add execution capacity.

That does not mean people stop mattering. I think the opposite happens at the top of the value chain.

Judgment matters more. Architecture matters more. Customer context matters more. Domain expertise matters more.

But repeatable execution becomes increasingly software-mediated.

What dies

Headcount as the default proxy for productive capacity.

What survives

Talented people. Deep expertise. Human accountability. Trust. Customer relationships.

Where the value moves

Toward firms that can combine smaller, more capable teams with reusable IP, agents, orchestration, data, and outcome accountability.

The services company that scales systems starts to look very different from one that only scales staffing. Eventually the valuation should too.]]></content:encoded>
    </item>
    <item>
      <title>AI is a feature.</title>
      <link>https://aiordie.com/dead-ideas/dead-idea-ai-is-a-feature</link>
      <guid isPermaLink="true">https://aiordie.com/dead-ideas/dead-idea-ai-is-a-feature</guid>
      <pubDate>Tue, 21 Apr 2026 12:00:00 GMT</pubDate>
      <category>Dead Idea</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>If your AI strategy ends with adding AI to the product, you may be improving yesterday’s architecture.</description>
      <content:encoded><![CDATA[The first response to every platform shift is predictable. Add the new thing to the old thing.

Mobile application. Cloud version. AI assistant. Copilot. Chat interface.

Some of those additions are valuable. But they can disguise the larger change.

If AI simply makes an existing workflow faster, that is AI-enabled. If AI changes whether the workflow needs to exist at all, that is something else.

That is why I make a distinction between AI-first and AI-native.

AI-first asks where AI can improve this workflow. AI-native asks why this workflow still exists in this form.

That second question is much more disruptive.

What dies

The assumption that AI is simply another feature layer.

What survives

Great products. Great workflow logic. Enterprise data. Security. Governance. Domain expertise.

Where the value moves

Toward systems designed around what AI can execute, rather than interfaces designed around what humans have historically had to do.

If you would not build the workflow the same way today, adding an AI button is probably not enough.]]></content:encoded>
    </item>
    <item>
      <title>A Violent Repricing of Tech Has Started</title>
      <link>https://aiordie.com/arguments/violent-repricing-of-tech-has-started</link>
      <guid isPermaLink="true">https://aiordie.com/arguments/violent-repricing-of-tech-has-started</guid>
      <pubDate>Tue, 07 Apr 2026 12:00:00 GMT</pubDate>
      <category>Argument</category>
      <category>Economics</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>AI is attacking the two economic assumptions that powered technology for the last twenty years: expensive labor and differentiated features.</description>
      <content:encoded><![CDATA[I think a violent repricing of how value is created in technology is underway.

Not because software disappears. Not because people disappear.

Because AI is attacking two economic models simultaneously.

First, it compresses labor. Work that once required a team can increasingly be performed by a smaller team working with software.

Second, it commoditizes features. Capabilities that required years of product development can increasingly be generated, assembled, or reproduced much faster.

That creates a difficult problem.

Services companies traditionally scale revenue by adding people. Software companies traditionally scale revenue by selling differentiated functionality.

What happens when fewer people can deliver more work and features become easier to create?

Value moves. I think it moves toward execution.

Owning the business problem. Owning the context. Owning the workflow. Owning enough authority to make a decision. Being accountable for whether the outcome actually happens.

That creates businesses that look less neatly like software or services. Software takes on more execution. Services embed more software. The boundary starts to disappear.

The dangerous place in technology is a business that scales primarily with people, or sells features without controlling meaningful workflow.

Both models can continue operating. They just may command very different economics.

The question is no longer whether you are a software company or a services company.

The better question is: can you execute?]]></content:encoded>
    </item>
    <item>
      <title>The Execution Layer Is Where the Value Moves</title>
      <link>https://aiordie.com/arguments/execution-layer-is-where-the-value-moves</link>
      <guid isPermaLink="true">https://aiordie.com/arguments/execution-layer-is-where-the-value-moves</guid>
      <pubDate>Tue, 10 Mar 2026 12:00:00 GMT</pubDate>
      <category>Argument</category>
      <category>Software</category>
      <dc:creator>Jason Wojahn</dc:creator>
      <description>Everyone is focused on better models. I think the bigger shift is happening one layer lower, where decisions become actions.</description>
      <content:encoded><![CDATA[For most of the last 30 years, enterprise software had a fairly clear job.

Capture information. Organize it. Analyze it. Show someone what happened.

Then a person took over.

A manager decided. An analyst interpreted. An engineer implemented. A services team executed.

The software stopped at the edge of the work.

AI starts to erase that boundary.

A system can increasingly understand context, determine what should happen next, initiate actions, evaluate the result, and continue.

That sounds like automation. It is bigger than automation.

Automation follows a path someone already designed. Execution systems increasingly determine the path.

That changes where value accrues.

The most valuable model may not win. The prettiest application may not win. The company with the most engineers may not win.

The advantage starts moving toward whoever can combine:

Context. Data. Policy. Decision authority. Workflow. Execution.

That is the layer I am watching, because the gap between knowing and doing has always been expensive. AI is starting to collapse it.

The next generation of enterprise technology will not be defined by what software can tell you. It will be defined by what software can reliably get done.

Find the boundary where your systems hand the work back to a person. That boundary is probably where the next redesign starts.]]></content:encoded>
    </item>
  </channel>
</rss>
