Announcing Cyber Content Revival (CCR) GPT

CMO: "Hey, we need to up our lead-gen game. We must find a way to produce at least one new campaign per quarter."

PMM: "Unfortunately, we've only budgeted two new thought leadership pieces this year."

CMO: "How about we get Avi [CTO] to write another piece like he did a few years back? That was the most successful lead-gen campaign we've ever run."

Avi: "Sorry guys, I'm slammed with the new release, so I couldn't even think about this till Q3."

Does this discussion sound familiar to you? Are you continually scrambling to develop new content for leads? This is the problem I want to solve with Cyber Content Revival (CCR). CCR is a GPT (from OpenAI ) I've designed that does two things:

1. Revive existing content. CCR extracts the main points of an existing cybersecurity white paper, highlighting areas that may be out of date (e.g., old stats, old practices, old concepts), and then gives you the option to determine which of these you want to update. CCR then pulls in web research and rewrites your white paper.
2. Refocus the buyer persona. CCR presents a set of eight cyber-buying personas. You choose which one you want for the white paper, and CCR will rewrite the white paper, targeting the new buyer persona.

The result is an updated white paper that maintains the original core messaging with more current statistics, best practices, and, when necessary, a different target audience. It may even be a more engaging read than the original!

TLDR: What's your call to action (CTA)?

If I've already hit your TLDR limit, I'm looking for test cases. Are you a senior marketing person (no bots, please) at a cyber company? Please DM me.

I'm looking for a handful of older 3-5 page white papers from established cyber companies to run through CCR. The process is simple and has no cost:

1. You send the content to me, and I run it through CCR. Let me know if you're looking for an update, refocusing the buyer persona, or both.
2. I'll be the anti-hallucinogen by curating the results and then ship the updated paper to you.
3. In return, I ask that you assess the output and give me feedback on the potential value of CCR to you and your company. Simple.

If you want to better understand why CCR is unique, please keep reading my short FAQ.

Why am I doing this?

I built CCR because I'd like to know if GPT and other large language models (LLMs) will eventually replace me. After all, most of my work these days is developing thought leadership content for cybersecurity products and services companies.

The good news for me - and content writers like me - is that after many hours on LLMs, I've learned that they suck when generating new cyber content (more on this below). They are, however, really good at updating/editing existing content. This aligns with a project I recently completed for a client (not using AI). The client wanted to repurpose a 3-year-old white paper written by the founder to generate new leads. While working on this project, it occurred to me that - like my client - most organizations have spent tens of thousands of dollars to develop thought leadership content (some written by me) for lead generation. Unfortunately, the half-life on these pieces is relatively short; most white papers die on the content shelf.

Putting it all together, I saw a potential role for LLMs to revive older cyber content. Hence, CCR!

Why is GPT like the love child of Jan Brady and Rain Man?

To develop CCR, I spent weeks tweaking the configuration file to get consistent, reliable output. Engaging with GPT (and other LLMs) is like working with a truculent, imaginative, self-absorbed teenager. As a frame of reference, imagine that GPT is the love child of an adult Jan Brady (from The Brady Bunch) and Rain Man. Yes, I know. This is a total Boomer analogy!

Like Jan, GPT is so eager to please. For example, no matter what I submit, it always spits back 2-3 times the word count. Unfortunately, also like Jan, it has a habit of making stuff up because it thinks it's what you want to hear. It loves to inject flowery language and cliches and consistently create new acronyms. It will even go so far as fibbing (now, Jan!). For example, GPT once told me not to be concerned because the 300-word output was just slightly more than the 150-word input. I could go deeper on the Jan Brady analogy, but Marcia has always been my favorite. Sorry, Jan!

Regarding Rain Man, GPT knows a lot and has access to nearly everything with its web search capabilities. Unfortunately, like Rain Man, 99% of what GPT knows is irrelevant to your needs. The challenge is that it has difficulty focusing on that 1%. Further, as Tom Cruise learned, you have to be extremely precise in what you ask. There is no room for inference, assumptions, or even common sense. To address this, I define seven clear rules of engagement for CCR. Still, I have to remind it continually to follow these rules!

Why is this different than feeding a white paper to generic GPT?

For me, the reason why CCR is different than a generic GPT session comes down to three requirements: voice, governance, and curation.

• Voice - CCR represents my voice. I've developed my voice after writing heaps of white papers and web content over the past 15 years. For example, I almost always use active voice, inject analogies often, limit cliches, motherhood and apple pie statements, and inject humor wherever possible. For CCR, I had to instruct GPT on how to use my voice.
• Governance - As mentioned above, getting an LLM to do precisely what you want is difficult. In my configuration of CCR, I provide clear steps it must take to accomplish the tasks at hand. For example, when changing the target persona, CCR had an annoying habit of writing about the persona rather than from the persona's viewpoint.
• Curation - Of these three requirements, curation touches on why I started looking at LLMs in the first place. Will it replace me? The answer is "no," at least for now ;-). From my perspective, everything CCR creates requires some subject matter expert (SME) curation. At a minimum, I must ensure that CCR follows my rules of engagement. More importantly, I must curate the output to confirm that it's engaging, thought provoking, technically accurate, and maintains (or enhances) the company's core messaging. For the foreseeable future, I do not see LLMs doing this level of curation.

Would you like to see if CCR, like Rain Man, is "an excellent driver?"

If you are interested, I'm looking for five white papers to test drive CCR. If you are a senior marketing person at a cyber company, please DM me. I'll go ahead and run your content through CCR; in return, I would like your feedback on the output.