What role should we give AI in content generation?
Image by DALL-E 3

What role should we give AI in content generation?

You want the partners in your ecosystem to succeed with their marketing and know they need great content. But you also know they lack the marketing horsepower to generate all of the content they need.

Generative Artificial Intelligence (GenAI) like ChatGPT has the potential to revolutionalise your partners' marketing efforts making content generation easier than ever. But should they? Should you encourage them to generate content using an AI?

I'm going to dissect the promise and pitfalls of leveraging GenAI, specifically Large Language Models (LLMs) like ChatGPT, in marketing and propose a 'safe' role for Generative AI in content generation.

Let me quickly declare my multiple (perhaps conflicting) biases.

  1. Firstly, I spend most of my day doing one of two things: advising large and small companies on their growth strategies and overseeing the execution of those strategies with our outsourced marketing work. So, in that first sense (my day job), I perhaps have a bias towards keeping the humans in the loop.
  2. My nights are largely filled with guiding the development of the multiple AIs we use in our own application called Funnel Plan which underpins much of that planning and execution work. So, in that second sense, I am a zealot for the role of GenAI in either doing things less expensively or doing the tasks we have always performed better.
  3. Recently, I have been advising growth companies on their own best-fit GenAI use cases - not narrowly in the realm of marketing but in the broad array of use cases for which GenAI presents new opportunities. So, this third sense has me perhaps less automatically in one or the other camp on the question of AI vs. humans.

GenAI: Analysis, not just words

The ability of GenAI to take a loose brief (a prompt) and create coherent content has many marketers salivating. It also has business leaders worried - some anyway.

The enthusiasm for AI's generative capabilities comes with a caveat. While these new AIs excel at generating content based on a prompt, they inherently lack the nuance and insight to capture a brand's unique intellectual property (IP) and tone of voice. The subtleties of creating content that resonates with a specific audience, reflects a brand's ethos, or articulates a nuanced argument about a pressing issue often elude the grasp of even the most advanced LLMs without highly engineered prompts and diligent iteration.

I'd argue that the ability to generate a blog article is not the greatest of the gifts that we have been given. Before ChatGPT changed the landscape a little over a year ago, marketers knew that AIs like machine learning (ML) could digest extensive data sets to find patterns. But they also believed that AI had little to offer the daily grind for most in the field of marketing.

The ability to make sense of imprecise prompts means that LLMs like GPT4 (the engine behind ChatGPT and Microsoft's Co-pilot), Anthropic's Claude2, Meta's Llama 2, and the many others that have since emerged, means that they can do more than generate new content. These AIs can also distill complex information into coherent, accessible insights.

For instance, consider the task of summarising key context to serve as background for that new content. AI can swiftly analyse thousands of sources, providing a succinct summary that outpaces and outperforms even the most diligent researcher (if you have one), let alone the time-poor copywriter who would otherwise have to satisfy themselves with doing just a little homework.

This capability extends to reviewing and refining existing copy and summarising dense transcripts to identify core points to inform subsequent content creation. The speed and efficiency with which AI accomplishes these tasks offer undeniable benefits, from time savings to potential enhancements in content quality.

The humans are irreplaceable

This brings us to the crux of integrating AI into our partner marketing strategies. The technology, for all its sophistication, operates within limitations. Without a highly engineered prompt and a disciplined iteration in the hands of a skilled AI prompt engineer, an LLM like ChatGPT generates vanilla content that could as equally have come from any of your competitors as from you. LLMs are unlikely to grasp your brand, your unique insights, or the specifics of the problems you solve. AI-generated content leans towards a generic average, lacking the distinctiveness and depth that come from human expertise. It could as equally have been written by you or any of your competitors.

Think now not of your own brand but the brands of your many partners operating in the complex ecosystem typical for partners for a major vendor. A key challenge for the partner marketing teams in major vendors since the dawn of time is that their partners are so keen to be 'available' to partner up on any deal offered to them by the vendor that their marketing messages become indistinguishable. "We do all things for all people". AI-generated content makes this worse, not better.

For the marketing professionals in your ecosystem and the copywriters supporting them, their role has not been replaced. We are tasked with guiding AI, shaping its output to align with our strategic goals, and imbuing the content with the unique perspectives and insights that only we, as domain experts, can provide. This involves a careful balance of leveraging AI for efficiency while ensuring the content remains impactful, relevant, and truly representative of our brand and the value we offer to our partners.

Strategies for AI integration in content generation

If you are in a partner marketing role and want your partners to harness generative AI effectively while maintaining differentiation in their marketing content, consider recommending the following strategies:

  1. Content roadmap: Before even touching an AI, build a roadmap of the content topics for which content should be generated. We usually suggest that the roadmap be populated with the many insights into the problem that each partner is focused on solving: what causes it, what are its consequences, how does it play out, and what do you need to solve it? Don't go asking an AI to do that for you - these are strategic questions able to be answered only by those 'in the tent'. Or if they do, use the AI to generate ideas, then build the roadmap by hand.
  2. Style guide: Whether as formal as a published style guide or not, it is helpful to be explicit about the dos and don'ts that are to be used in any content. Maybe the partner already has this. If not, it can be built by hand by a skilled copywriter. It's also feasible - again with the help of a skilled prompt engineer - to have an AI review the last 10 blogs and build the style guide retrospectively.
  3. Deep research: As they turn their minds to each topic, recommend that they craft prompts that elicit insights about what others are saying about that very topic. Do this not to copy those insights but to make their own insights sufficiently distinct.
  4. Unique views: Mindful of the unique problems that each partner solves and the market for whom they solve it, but also mindful of what competitors are saying, have them identify unique insights. What is it that the author(s) of this article believe that are consistent with the strategy and distinct from the competition?
  5. Ideation: Craft detailed, strategic prompts for AI to create 'angles' for the blog. This is a creative task to elicit alternative points or even narrative arcs that allow the partners' distinct insights to be shared while acknowledging that there may be alternative paths to doing so.
  6. Draft generation: Engineer a prompt to have a draft of the article generated that holds to the unique opinion, differs from those of the competition, honours the style guide, and leverages the points identified or narrative arc adopted from the prior step.
  7. Rewrite for nuance and personality: Subject all AI-generated content to thorough human review by a skilled writer working with the domain expert. This step is crucial for fine-tuning the content to reflect opinions held, facts known, and a brand's voice and ensuring it addresses the specific needs and challenges of our partners.

What have you seen?

Have you experimented with leveraging AI in your marketing efforts? Have your partners? Have you witnessed firsthand the impact of AI-generated content, either from the perspective of a creator or a consumer?

Let me know your thoughts. I'd love to know.

Hugh Macfarlane

Partner Growth Acceleration with Go-to-Market process optimisation for B2B businesses and their partners

9mo

And another email. This one made no attempt at AI-generated personalisation, but I bet this was written by an AI and not a copywriter. ChatGPT loves platitudes and cliches. Hi Hugh,  In the ever-evolving landscape of workplace safety, we must continually adapt to ensure the well-being of all employees. The introduction of new legislation across Australia has cast a vital spotlight on psychosocial hazards, recognizing their profound impact on workforce well-being.   This legislation not only places a legal obligation on employers but encourages proactive action to foster a safe and supportive workplace. Creating awareness is key. Employers must understand their responsibilities and the benefits of a mentally healthy workforce, while employees need to feel valued and informed. So, that's not horrible, but neither is it good. It is the sort of banal dross that blog content or outbound emails cannot afford to use. We are HUGE users of GenAI, and even run workshops for businesses to help them find legit use cases. But please don't outsource your content writing to a middle-of-the-road GenAI bot. Or if you do, get good at writing super-tight prompts to force personality and unique insight into your content and proof aggressively.

Like
Reply
Hugh Macfarlane

Partner Growth Acceleration with Go-to-Market process optimisation for B2B businesses and their partners

9mo

Here's a perfect example of GenAI being used without due care. Content in square brackets is mine: Hi Hugh, Firstly, we want to congratulate you on your remarkable journey with alignme [close], shaping a narrative of innovation and redefining the B2B marketing landscape [good]. As tristan, Co-founder at Buzzai, I've witnessed the transformative power of alignme's strategies [liar - you have likely never heard of us], reflecting in recent accolades like the groundbreaking'Show Me Story' with Invisalign-maker Align [not us]. Companies like SharpSpring and HubSpot have thrived by staying ahead of market trends, much like alignme [cheap and missed the mark]. We are impressed by alignme's commitment to growth and values alignment, evident in the Align Method's impact on the ballet community. complete miss. [this ttuff is dangerous in the hands of sloppy operators] Amidst this excellence, let's explore a unique strategy session tailored exclusively for alignme, leveraging Buzzai's AI-powered sales engagement platform. [not likely] Join us in a 20-30 minute chat to uncover new growth opportunities and stay ahead in the ever-evolving market landscape. Partly cloudy skies in Melbourne set the stage for a bright future ahead. [cheesy]

Like
Reply
Hugh Macfarlane

Partner Growth Acceleration with Go-to-Market process optimisation for B2B businesses and their partners

9mo

According to Trust Insights' Christopher Penn, there are three reasons we need to disclose if any of our content has been AI generated. Two of them legal. But what about having an AI proof your work? https://www.trustinsights.ai/blog/2024/02/disclosure-of-ai-and-protection-of-copyright/

Like
Reply

To view or add a comment, sign in

Insights from the community

Explore topics