I question the idea that this generation of AI replaces (or even approximates) human creativity. But: I’ve always been fascinated by the vital role of technology in enabling human creativity – AI included. (In enterprise settings, AI clearly has a role to play in content generation and personalization – a topic for another time).
When I bear down on AI’s pros and cons, I want to identify a clear path for experiments, which lead to strong use cases. That includes my own creative workflows.
Can AI enhance creative projects? Time for the AI content dexterity test
Example: for years now, I’ve seen a potential role for AI in editing B2B video (see my 2023 post, AI for video editing – a viable use case for your B2B content strategy?) On the creative side, I look for use cases where AI can:
- amplify – and scale – human creativity
- summarize complex or long form content
- produce content in new formats, or in formats where a team isn’t well equipped
Of course, this needs to happen with sandboxes/guardrails, where we don’t lose our production instance when “vibes” go wrong.
Recently, I threw myself into a fresh round of experiments that honed in on #3: produce content in new formats. (This project builds on my post, Reaching the Informed B2B Buyer – time for the relevance versus entertainment debate!)
When it comes to content dexterity, I am polarized. I’ve always been a writer; I wrote my first spoken word poem on a Commodore 64, after I survived a hit-and-run accident in my ice cream truck. I’ve been a podcaster since I starting taping speaker phone recordings in 2008. Video is also in my historical wheelhouse, but: I lose patience for fussy editing quickly.
On the other hand, I absolutely stink at creating appealing slide decks. And: I couldn’t craft a decent infographic, even if diginomica’s fate depended on it.
Can AI prove itself in my creative workflows? New formats – and results
The most interesting story is my long/winding attempt to get a great infographic out of my B2B Informed Buyer d·book, but first: a tour of other formatting experiments.
Since I last reviewed “AI” services for producing highlight clips from longer B2B videos, those services are now everywhere. My personal go-to for this is OpusClip, but many services do a decent job. I like OpusClip because:
- It has multiple templates for different kinds of clips, including intro and outro options
- There is a good explanation and ratings system for why it selects the videos it does (sometimes the explanations are uncannily good)
- It is incredibly easy to use – paste in your YouTube link, pull up your template of choice, pick your style of clip and go.
OpusClips’ prompt-based video highlight generator is a winner. I have issues with generative AI’s buzzywordy techno-jargon, which doesn’t work for my personal style. Therefore, I push the prompts to give me “informal” language, or “outspoken” opinions. Example: I have prompts for OpusClip to surface more controversial/critical highlight clips. (Here’s one recent Opus clip from my Enterprise Month in Review show, with Brian Sommer going off on AI-generated news mishaps – in YouTube Shorts format).
AI for Powerpoint generation? I am nowhere on this one. I found one service that is better at creating the kind of informal language I am looking for, but the output of images and text are incomplete at best, makeshift at worse. If you have a good one, let me know.
AI for text-to-animation? Again, not much luck. I’ve had fun with Steve.ai, but only for frivolous birthday cartoons. I think Steve.ai would be good for satirical enterprise animations, but I have a hard time seeing it work for more serious treatments.
Article to video? I did another round of sandboxing AI output at Pictory.ai – their team is collaborative about testing new approaches, and they have plenty of formatting choices. Though I haven’t had much luck with Pictory’s text-to-Powerpoint option yet, I had an unexpected success with their URL-to-video option. This 2 minute video, which I generated from my article on rethinking B2B content, came out well enough for me to post on YouTube: Reaching B2B buyers – why we must redefine “entertainment.
One video image, however, was completely inappropriate. Which brings us to more user tips:
- As usual, I finessed the AI prompt to get more informal output with less buzzwords. Even so, I wouldn’t normally use all of the techno-jargon in this video.
- AI tools should allow for quick interventions to replace problematic scenes or images. (I’m still personally squeamish about using AI-generated images; more on that shortly).
AI-generated video summaries – a different challenge than highlight clips!
I also worked with Pictory.ai on a service that is harder to find: AI video summaries. The use case for AI video summaries is different than clips. For summaries, “AI” is making judgement calls on what represents an effective/compelling video, and splicing those clips together. That’s not easy to do, without a disjointed video. Using Pictory.ai’s tech, I produced a six minute version of my recent Enterprise Month in Review show with Brian Sommer. This was originally a 120 minute live show (highlight reel: Opinionated take on Mary Meeker’s AI report from BOND, and RTO mandates).
On my first summary video with Pictory, I ran into some overly-choppy switches, but I’ve now learned how to (mostly) avoid that. I’m now able to modify the clips before producing them; I can also quickly add more, via transcript highlights (you can do the bulk of the editing by searching and highlighting text).
For my recent State of Enterprise Video show with Brent Leary, I was able to customize (and produce) the six minute highlight video in about 30 minutes, featuring all three themes of the show:
- the TikTokization of LinkedIn video
- the power of live enterprise discussions, and:
- why watch parties are the key to rejuvenating enterprise keynotes
(also see Pictory’s highlight video below).
Could I edit this kind of highlight video myself? Yes, but I am not talented at this, nor do I enjoy it (I use Vimeo.com). Could I outsource this to a human? Perhaps, but there are issues of turnaround time, and capturing the topics I want. With Pictoria’s transcript interface, I am essentially “vibing” my way through video editing, with a push from AI on clip selection. But: I control the text content (and therefore the images).
These same considerations of labor and scale are relevant in an enterprise context. Too often, marketing teams have a high value content asset, but struggle to repurpose it across formats, platforms, languages, etc. Then: there is the question of which content to send/show to which person. AI is very capable – though far from infallable – in these areas. However, the excellence/relatability of the original content asset should not be underestimated.
In pursuit of a B2B buyer infographic – how AI fell short
That’s why I thought I might be able to use AI to turn my B2B Informed Buyer d·book into an AI infographic – but it was a total fail. AI was not even close to the challenge.
The first service I tried was Napkin.ai, but all I could get out of Napkin was a few bland looking, disconnected slides that did not frame a narrative. And, I was unable to export them – clunky interface.
In fairness, while this Napkin.ai slide is pretty bland (pictured right), and whiffs on “B2B management,” the illustration of the topics on this slide is quite accurate. It’s nowhere near an infographic, but it shows the tantalizing aspect of one of AI’s real strengths (summarizing concepts).
In Napkin’s case, we have a really effective conceptual flow, aside from the “B2B Management” weirdness on the right (I also tried a service called Piktochart, but the images were cheesy stock photo purgatory, and lacked any narrative flow)..
The clear winner of the three AI infographic services I tested was Venngraphic (pictured left). Though the images are nothing special, the layout had a logical flow, and uses mostly my own language – avoiding buzzword malfeasance.
Venngraphic made good use of my text content, before veering into a bizarre tribute to Pinterest, which I could have edited out. However, Venngraphic’s quality relies on the caliber and depth of the prompt material. And if that’s the case, how differentiated are these images?
While Napkin.ai didn’t come close to delivering an infographic, I thought how they framed my material in slides was the best value of the three.
Next, I attempted to outsource this infographic to an online marketplace. The service providers did a pretty good job of helping me boil down the concepts – a surprisingly good job, when you consider the language barriers of offshore work.
But: the images were more like consumer infographics, the kind you might use to generate Instragram excitement for a school trip to DisneyLand. By the end, the text looked good, but the visual personas were off the mark.
That’s when I turned to Two Line Studios. I know Two Line from my years of attending Constellation’s Connected Enterprise event, where Heather Willems and colleagues do live illustrations during the sessions. Willems has done illustrations of my sessions several times, such as this one from CRM Playaz’ first on the ground event, where we did a countdown of the biggest CRM mistakes of all time:
This led me to bring Willems on my podcast, to discuss the impact of visual narratives at enterprise events, and also the impact of AI on her work. Like most creative professionals, I’m sure Willems will find uses for AI. I’m much less confident, however, that AI could do anything that resembles the finished product we did here:
This is not the full infographic – you can see the full diginomica Informed Buyer infographic image here, or scroll down.
Why? Willems and I met once in person to discuss what worked – and what didn’t – with my prior infographic attempts. Yes, it was just one in-person meeting, but it made a difference. Willems took live notes using a cool designer tool – which became a project framework for us to follow. We didn’t require many subsequent meetings to complete this project, but there was plenty of email-based fine tuning, as we agreed upon the protagonist – and other personas who would appear in the infographic.
The personalities – and the vibrant colors – are thanks to the talents of the main infographic artist for this project, Two Line Studios’ Ona Rygelis. If you look closely at the infographic, you’ll see a number of small flourishes (AI imaging doesn’t do those). What I like most? Buyers typically involve teams in their decisions, but that feeling of uncertainty/overwhelm is still an individual predicament. Here, we get images of both, with a light touch of humor to offset the stress factor of making good decisions with disruptive change afoot.
My take – original content still matters, and so does the ethics of creative spend
Exceptional/original content that reinforces your personal (or brand) narrative still matters – and definitely still matters in B2B. We’re nowhere near the point where “AI” can do what humans can do, when it comes to humanizing concepts and creating ‘experiences” that change how audiences think. Eventually, that content – and the interactions it sparks – builds a trust you simply can’t summon with prompts alone – and neither can your competitors.
However, I’m seeing more evidence that the latest iteration of these tools can enable creators to do more – and for brands to “spin out” their best ideas in new formats, languages, etc. (Sometimes automation beats “AI” – I distribute my Busting the Omnichannel podcast series via Auphonic’s combination of auto-podcast optimization and automated distribution). What I don’t see enough of is the clear understanding that human creators remain invaluable.
Yes, budgets can crunch us on that front, but that doesn’t mean we should delude ourselves into thinking that small budgets replace brilliant work. The winning approach is still to invest in that core human exceptionalism, then expand and personalize that content across domains. Tight budgets are often less of an impediment than we think. Not when you build a culture of creativity, where experts inside your organization are freed up to share.
I strongly prefer using my own content and images, which is why I like AI video editing tools the best. In the ideal world, I would only use tools that (fairly) compensate creators. I remain uncomfortable using free AI services for that reason.
I don’t like seeing images in my AI output with no idea where they are from, how they were created, or how they were licensed or paid for. There are plenty of paid AI services that are not sourcing ethically either. These are lines in the sand we all need to draw, and I won’t tell you what yours should be.
But if we are going to sing the praises of the tools, we should be aware that many artists have either lost – or are in the process of losing – their livelihoods to these tools; tools which were trained on their work. This does affect which tools I’m willing to use, and how much. The good news? There is nothing to stop us from hiring those artists – and making the most of AI tooling also.
If I can save time on one aspect of creative formatting, I want to use that time to invest in other projects and creators. For my last book, I didn’t use a human copy editor for the first time, but I did pay a human audiobook narrator with the money instead. For all those reasons, AI video editing of my own content is the AI formatting use case I like the most.
By far, the most fun experiment was the collaborative work with Two Line Studios. It’s no accident that the result lived up to my expectations also.
In closing, here is the full size Informed B2B Buyer infographic:
For more context: Reaching the Informed B2B Buyer publication and intro post. Infographic by Two Line Studios; art by Ona Rygelis of Two Line.
Here is my B2B video strategy highlight reel from Pictory.ai:
