How I/We/Collaboration.Ai Employees Prepare for a Webinar About AI, By Actually Using Agentic AI

Statistics and webinars

Intro

I was invited to join Scott Kirsner of InnoLead and our own Jan Fischer for a conversation about the AI and agentic future of innovation management software. A webinar. About innovation. About AI. So the question became obvious: do I prepare for this the old way, or do I practice what we preach? I chose to make the preparation itself the innovation.

What Most People Do

Most people preparing for a webinar Google a few stats, skim a few articles, jot down three talking points, and hope the moderator carries the conversation. It works. It is fine. It is also exactly what everyone else does. The problem with “fine” is that it is forgettable. More than half of webinar attendees leave before the session is complete. When preparation is generic, output is generic, and the audience can feel it.

What I Actually Did

I treated this webinar the way our platform treats any innovation challenge: I used our UWT (United We Transform) methodology. Short version: Scan, Focus, then Act decisively.

Scan

I ran deep research across platform-aggregated data from more than 2 million B2B webinars, including ON24, Goldcast, Banzai, BigMarker, and TwentyThree. Not a quick search. A structured investigation using AI research agents trained on specific contexts and pulling from multiple data sources simultaneously. Here is where it diverges from what most people mean by “I used AI.” I did not type a question into ChatGPT and copy the answer. I used a stack of tuned skills: specialized AI agents with layered context about our company, methodology, competitive landscape, and likely attendee psychographics. Each skill built on the others. Deep research fed content strategy. Content strategy fed messaging. A serendipity engine surfaced unexpected connections I would not have found manually. *easiest way to scan is to have everyone give an answer first

Focus

The research surfaced facts that reframed how I thought about the medium itself. 60% of webinar registrants never arrive. Half the live audience leaves before minute 30. Nearly half of all consumption now happens on-demand. The fastest-growing webinar output is AI-generated short-form video clips, up 2,903% in one year. Webinars are not events anymore. They are source material for content engines. That insight changed my approach. Instead of preparing to present at a webinar, I prepared to create a content asset designed for multiple formats, knowing the live audience is only part of the equation.

Act

From Scan and Focus, I produced three outputs from one intelligence cycle:

  • A one-minute “Day in the Life” video with verified statistics, using Sora, VEO3, and Kling to speed production.
  • This blog post.
  • A LinkedIn post.

Three assets from one intelligence cycle. That is what high-performing organizations do: treat each session as a learning engine, not a one-time broadcast. I also tested findings across different models, challenged framing in team Slack channels, and used psychographic profiling to map which points would resonate with specific audience segments.

Why This Matters Beyond One Webinar

“Could we use AI not just to assist with creative work, but to help shape, package, and even perform it?”

Scott Kirsner has been asking the right question. The answer is yes, but only if you move beyond prompting a chatbot and start building systems. The gap is not between people who use AI and people who do not. It is between people who use AI like Google and people who build AI into their operating rhythm with context, skills, feedback loops, and compounding intelligence. Jan has spent 15+ years building software that helps organizations bring diverse perspectives into innovation processes. At Collaboration.AI, we are extending that into the agentic era, where software does not just collect inputs, it actively participates in the work. This webinar is one small example. The principle scales to everything we do.

Closing

Brandon Klein is the founder of Collaboration.AI and co-creator of the United We Transform framework. He will be speaking with Scott Kirsner (InnoLead) and Jan Fischer (Collaboration.AI / innosabi founder) about the agentic future of innovation management. Register for the webinar → Interactive stats

More from the blog

Innovators are not one instruction.
AI enablement and adoption · Apr 30, 2026
How to Innovate Your Companies Town Hall / All Hands TRY 2
What's a focus group anyways?
The innovation function · Apr 23, 2026
Innovation Workshops/Focus Groups Don’t Work Anymore
AI and your job as a compression
AI enablement and adoption · Apr 20, 2026
From Farm to Factory to… Pontifications from a Dave Gray article

Connect your mission to the breakthrough it needs

Innovation work rarely fails for lack of ideas. It fails when what the program knows is spread across systems that don’t share a record. Bring the program, and we’ll show you where it could run on one.

Contracting vehicles

Direct award pathways, no new competition required.

SBIR Phase III

Under 15 U.S.C. § 638(r)(4), a Phase III award derived from Collaboration.Ai’s prior SBIR work may be made without further competition.

DAF DTO BOA

Order through the DAF Digital Transformation Office Basic Ordering Agreement.

Not sure?

We’ll walk through which pathway fits the program. Contact us →