If you’ve ever been a part of a backlog planning session for business innovation, automation or AI envisioning (like I regularly run), you know the moment.

We spend hours with the business:

• Listing pain points
• Mapping processes
• Surfacing inefficiencies
• Identifying manual work, rework, double handling

By the end, some expert (like me) may inevitably ask:

“So… is this a low-code solution or an AI solution?”

And that’s where I think many automation experts are still framing the problem wrong.

It’s not Power Platform or AI.

It’s both.

The False Divide

Some problems are structured:

• “When a form is submitted, create a record.”
• “If the value is over X, start approval.”
• “Notify the team when status changes.”

That’s classic Power Platform. Clean logic. Clear triggers. Deterministic outcomes.

Other problems are messy:

• “Summarise this 4-page email thread.”
• “Work out what this customer is actually asking.”
• “Extract meaning from this blob of text.”
• “Classify this into the right category.”

That’s where AI adds value, interpretation, reasoning, contextual understanding.

But here’s the reality:

Most real-world business processes include both.

What We’re Actually Building Today

We’re not building “AI solutions.”

We’re not building “low-code solutions.”

We’re building agentic applications.

Applications where:

• AI interprets input
• Business logic applies rules
• Automations execute reliably
• Humans oversee where needed

For example:

• An AI model summarises a long customer submission.
• A workflow parses that AI-generated summary.
• Business rules check keywords or risk indicators.
• A Power Automate flow routes it to the correct team.
• A model-driven app provides oversight and audit.

That’s not AI replacing automation.

That’s AI enhancing structured systems.

The Workshop Shift We Need

In automation workshops, instead of asking:

“Is this low-code or AI?”

We should be asking:

1. Where is the process predictable?
2. Where does interpretation or judgment matter?
3. Where would AI improve speed or accuracy?
4. Where do we still need guardrails and governance?

Because here’s the truth:

Even your “low-code” apps are increasingly infused with AI

• AI tick/check validation
• AI text classification
• AI summaries feeding workflows
• AI-assisted data capture

And your “AI solutions” still rely on:

• Structured data models
• Defined triggers
• Approval workflows
• Governance frameworks

They’re not separate categories.

They’re layers of the same solution.

Agentic Apps: The Middle Ground

An agentic app isn’t a chatbot floating in isolation.

It’s an intelligent layer sitting on top of structured systems.

It can:

• Understand context
• Decide next steps
• Call structured processes
• Escalate to humans
• Log actions for review
• Improve over time

That’s Power Platform plus AI.

Not one replacing the other.

The Real Maturity Signal

If a client is still trying to categorise every idea into:

• “That’s automation”
• “That’s AI”

They’re thinking in tool silos.

Mature organisations think in outcomes.

They design:

• Structured where possible
• Adaptive where necessary
• Governed always

And they blend capabilities to achieve that.

So next time someone asks: “Is this Power Platform or AI?”

Smile. And say:

¿Por qué no los dos? Why not both?