Dave Killeen: [00:00:00] Hello, hello, Dave Killeen here, Field CPO at Pendo in EMEA, and a very warm welcome back to The Vibe PM episode two. What's the one task that crushes our souls? I think it's building presentations. We have all stared at that blank slide, wrestling with text boxes instead of what truly matters. The story.

Today, we're gonna fix this in a powerful way. I'm gonna show you my two-part AI workflow. Part one, we will leverage the foundational skill from episode one using our voice speech-to-text. If you haven't watched it, you must do. Very important, highly effective. It's the ultimate hack for getting ideas out of our heads at the speed of thought, which in turn frees up those precious hours for our customers and our strategy.

Speech-to-text is what I turned to in a moment of high-stakes panic before having to present to 12,000 product managers at ProductCon London in February of this year. And based on the success of that presentation it's gone bonkers ever since, it then forced me to think about the second part of [00:01:00] this episode, which is a game-changing technique using a tool called Gamma to bring the story to life more efficiently, more effectively.

And in a way that I just don't see anyone else doing with Gamma today. Let's do this. You're gonna love it.

Okie dokie. Let's kick things off with the first of two powerful pillars that I'll be showing you in this episode. When it comes to presentations, my core belief is always that you need to get the narrative down first and your design after.

So get your speaker notes all pinned down and then put the visuals on top. Despite that, I had a certain level of anxiety about presenting to so many people, regardless of the fact that I knew I had an awful lot of useful material to share. The question was, how best do I structure everything with a compelling narrative, a nice, strong arc?

So I don't bore people, I don't lose people, I don't lose the audience, so I then use what I call the magic word trick. I covered it all in episode one. If you haven't watched it, please do. It's incredibly important and it's very effective. [00:02:00] And basically what I do is I dump all of my thoughts into an LLM on mobile using my voice.

I use Claude for this, and that's explained why in episode one. And I'm like, hello, Claude, Dave here. I'm gonna give you a bunch of voice notes, Claude, in no particular order, and when I say the magic word process, Claude, you're gonna take all those voice notes. And you're going to turn them into a set of 20 brilliant, magical slides of speaker notes.

With a strong structure, add enough detail, Claude, to keep me on stage for 20-minutes. And Claude, I've liked this all written in the style of Ken Norton, who I think story tells brilliantly, I love him. So thank you, Claude. Love you. And then I go outside and I take the mobile out and I start ranting at the phone, dumping my voice notes in, and I'm literally like, boom, boom.

In, in, in, in, in it goes no particular order. And then I say process and job done. Bonkersly, bonkersly, effective, and then with that rock-solid narrative to hand, I then work in partnership with our amazing design team here at Pendo to bring the visuals to life [00:03:00] and then get the slides all nailed down and done.

So effective and the structure you get from it. I came up with a narrative, which I would never have thought about before. It's so, so good.

The presentation was a huge success, and this photo that you see here says it all, but that success created a new problem for me. I started getting more requests to present externally, and I simply couldn't scale my time, let alone that of the design team for every talk. I needed to get from that perfect narrative with the LM, that we didn't step one to a polished deck visually.

Far, far faster, and that's when the following workflow in Gamma struck me that I'm about to show you.

A few months back, a large private equity firm asked me to present to their portfolio CEOs and product leaders on aligning product KPIs with financial outcomes. The trouble was I only had a few hours, so I took the transcript from my initial call.

I put that into the LLM on desktop through web. And then I went for a walk outside for 20-minutes with Claude on my mobile, to dump my [00:04:00] own thoughts into the LLM using that magic word trick. Boom, boom process. Claude, thank you Claude. Love you, Claude. And the whole thing was done. I had then the draft of speaker notes that you see here in this Google doc.

Unreal. Just look at the detail, look at the structure. It's so, so good. So that's part one of this episode done.

Now next up, let's head on over to Gamma. Where the very first thing you do before any content is set up your theme. Here is Pendo's, with our own brand colors, our fonts, our logos. Setting up your theme is the big unlock to get your deck to pop.

Alright, branding, theme is all done. Now for the story, I click on Create New. I then paste that copied speaker notes from the Google doc you saw earlier and crucially, I tell Gamma to preserve the notes to keep them intact so the AI can use that full context. I then click Continue, and now we're in the prompt editor in Gamma.

Here are all the key moves that you need to do. You tell it to condense the text, but keep it concise. You don't go minimal because you want [00:05:00] to keep the meaning. I then select that Pendo theme, and for images, I choose AI Images with the Flux Pro model, which I find works really well. And this is really important, the custom style prompt.

This defines look and feel for the AI Images. I have a detailed one here that asks for a flat narrative illustration style with accents of Pendo's pink watermelon and no text because when text gets rendered, sometimes just not that good. It's a bit tacky. I'll put that prompt into the YouTube video description, by the way.

And now we hit generate and just look at that. Not too shabby, but the images are just okay-ish. You have this common AI bias that you see here. Very kind of male-centric, which just isn't great. So what we're gonna do now is another trick is we're gonna create a series of text-to-image prompts in the LLM on top of our speaker notes.

This is not always necessary, but I feel it's very effective as you're about to see. So what we'll do is we'll go into the Claude Console and here what you see, is I'm asking you to generate a text-to-image prompt essentially, for [00:06:00] each of my speaker notes slides, each prompt should align with corresponding slide content while maintaining thematic consistency across the entire presentation.

The images should exclude any text and feature diverse representation in terms of gender and ethnicity. That for most people would be a good prompt, but wait till you see the magic of the Claude Console, honestly, tick this box here because we're working with thinking models, and then click generate, and then when you click generate, here's what you get back.

It's much, much better. You'll see a fundamentally different response from the LLM when you use something like Claude Console. It's very, very good. It's worth its weight gold. It's about $5 for seven months worth of credits. So when you take all that through and you run that again, send your LLM where you have all those speaker notes, this is then what you get back.

Then I copy and paste all of this. I go into the presentation. And then I just swap out the prompt that Gamma created with the more detailed one. And then when I run that all through, then this is what you see. And this actually was a presentation that was given then to the [00:07:00] PE firm a few months back.

But I'm going to let you in on my favorite part of this whole system.

The presentation isn't the only asset that we've created here. The real golden asset is that rich narrative conversation that you've had with all that context in the LLM, that's your source of truth. So what I then do is I go back into that same chat and I say, hey, give me a three-page summary doc from our conversation to use as a follow-up.

And just like that, you have a perfect handout. It's a total game-changer. You're creating multiple assets from a single source of truth with almost zero extra effort. Honest to God, it's so effective. And so once you're happy with all that, then you can export that into PowerPoint or Google Slides and collaborate with your team over there to finesse and finish.

And that's the end-to-end process. We've used our voice to chat to the LLM to give it the context. We give it the magic word and we take that rich set of speaker notes and all the other context you might wanna throw at it. Deep research reports, whatever it might be. You know, call recordings, and then you take all of those speaker notes into Gamma [00:08:00] and then you'd be ready to present. In no time.

By the way, Gamma recently released Gamma 3.0 in September 25, which is still in beta, but let's you do even more with the visuals and with the data, so definitely do go check that out.

And so there you have it, the LLM as a story, copilot and Gamma as a rapid, execution designer. This is how you save hours and raise the quality, throughput, and impact of your work with your colleagues and more importantly for your customers.

And so that's it for me. All my episodes are short and sweet and under 10-minutes, so please share this with your colleagues so that they can save time with you, spend more time with customers. Thank you for joining me on another episode of The Vibe PM.

Please spread the vibes and please, I'll see you next time.

Thank you.