Optimizing for the Agent Era

Welcome & Introduction

Roger Snyder | 00:00:00 – 00:03:00
Hello everyone, and welcome to Productside’s webinar, “Optimizing for the Agentic Era.” We’ll give people a moment to join, and in the meantime, feel free to drop your location in the chat.

We’ve got a global audience—Denmark, Portland, India, Chicago—great to see everyone here.

Let’s go ahead and get started. My name is Roger Snyder, Principal Consultant at Productside. Joining me today is Tom Evans, Senior Principal Consultant based in Austin, Texas.

We’ll keep this session interactive—please use the Q&A for questions. Yes, this session is recorded and will be shared afterward. Also, we invite you to join our LinkedIn community of over 40,000 product professionals.

Tom, I’ll hand it over to you.

Understanding AI Agents & the Evolution of Work

Tom Evans | 00:03:00 – 00:12:00
Thanks, Roger. Today we’ll cover three things:

1. Understanding AI agents and current trends
2. How product managers can use agents as teammates
3. Practical steps to get started

Let’s start with how work has evolved.

Before 2022, we were doing source-oriented work—manually gathering and synthesizing information. With generative AI, we moved to intent-oriented work, where we describe what we want and AI helps us generate insights.

Now we’re entering delegation, where AI agents execute workflows for us. And the next step is outcome-oriented work, where agentic AI pursues goals independently.

We’re moving from doing the work → to guiding → to delegating → to setting outcomes.

AI Maturity & Current Adoption

Tom Evans | 00:12:00 – 00:18:00
We asked how you’re using AI today. Many are still exploring, some use it daily, and a smaller group is already building agents.

Internally, AI agents are being adopted faster than external use cases. Why? Because internal environments give us:

• Better data control
• Lower risk
• Easier experimentation

It’s essentially a sandbox for learning before deploying AI into products.

AI Agents vs Agentic AI

Tom Evans | 00:18:00 – 00:25:00
AI Agents execute tasks and workflows under supervision. Agentic AI is goal-driven and more autonomous.

AI Agents:

• Execute tasks or workflows
• Operate under human supervision
• Follow defined instructions

Agentic AI:

• Goal-driven systems
• Autonomous decision-making
• Adapt based on context

The biggest differences are autonomy, decision-making, and goal orientation.

Regardless of type, accountability and guardrails are essential. You want autonomy—but not uncontrolled behavior.

AI Agents as Product Team Teammates

Tom Evans | 00:25:00 – 00:33:00
Think of AI agents as digital teammates.
Define their role, expectations, and escalation paths. Adoption is growing, but most organizations are still early in scaling.

That means:

• Define their role clearly
• Set expectations
• Establish escalation points
• Apply governance

Adoption data shows:

• 78% of organizations use AI in at least one function
• 62% are piloting AI agents
• Only ~23% are scaling

We’re still early—but momentum is strong.

Challenges in AI Adoption

Tom Evans & Roger Snyder | 00:33:00 – 00:40:00
Key challenges include:

• Accuracy and reliability
• Data privacy and security
• Ethical and legal concerns
• Integration and training

Roger shared an example where AI misclassified experts—highlighting the need for critical thinking and validation.

Tom emphasized:
Product managers must still understand first principles. AI augments thinking—it doesn’t replace it.

High-Value AI Agent Use Cases

Tom Evans | 00:40:00 – 00:48:00
Here are practical use cases for product teams:

• Auto-generating release notes
• Summarizing customer interviews
• Triaging feedback and feature requests
• Sprint readiness checks
• Stakeholder reporting
• Competitive and market analysis

Roger shared a real example of an AI agent delivering daily industry insights with product implications, saving significant time.

Key criteria for strong use cases:

• High volume
• Repetitive work
• Structured data (where possible)

Designing AI Agent Workflows

Tom Evans | 00:48:00 – 00:52:00
When building an agent, follow this journey:

1. Data – What inputs are needed?
2. Orchestration – What process should it follow?
3. Governance – What are the boundaries?
4. Output – What format is required?
5. Inspection – How do we validate results?

Governance applies across all steps—not just one.

Measuring ROI & Impact

Tom Evans & Roger Snyder | 00:52:00 – 00:56:00
Measuring AI ROI is still evolving, but key metrics include:

• Decision velocity
• Time saved
• Reduced manual work
• Increased throughput
• Improved outcomes

Quality matters too—not just output.

Roger shared an example where AI improved both speed and quality of a presentation, reinforcing that AI can enhance thinking, not just efficiency.

Getting Started with AI Agents

Tom Evans | 00:56:00 – 00:59:00
Start small and practical:

• Identify a clear pain point
• Choose a simple, high-impact use case
• Define governance and boundaries
• Benchmark before implementation
• Measure results

Then scale gradually:

• Individual use → Team → Cross-functional → Organization-wide

Don’t aim for perfection—aim for learning and iteration.

Q&A & Closing

Roger Snyder & Tom Evans | 00:59:00 – 01:01:00
In Q&A, key themes included:

• AI vs automation: Use AI when decision-making or synthesis is required
• Tool selection: The landscape evolves quickly—evaluate continuously
• Role of PMs: AI increases expectations for productivity and impact

Thank you all for joining. We appreciate the engagement and look forward to seeing you in future Productside sessions.

Go build smarter products—with smarter teammates.