I’ve been reluctant to write a post about generative AI because there’s so much already getting published on the topic every day, I fear it will get drowned out.
However, the time has come to take the plunge.
This post is about generative AI in the context of B2B content marketing. While some of my opinions apply more broadly, I’m trying my best to stay focused…
Specifically, I want to explore a handful of questions that content marketing leaders should be asking before putting generative AI to work.
Generative artificial intelligence (AI) models are deep learning models that have been trained on extremely large sets of unstructured data—usually wide swaths of the public internet.
Appropriately deployed, these models have the power to increase efficiency and productivity, while reducing human effort and cost.
Their power relative to earlier AI tools comes from their ability to perform more than one function. Previous models were narrowly trained to perform a single task, such as identifying objects from images. Today’s models are adept at classifying, editing, summarizing, answering questions, and—importantly—drafting new content with human-like skill.
However, they are not without their limitations.
For example, generative AI models have limited understanding of whether something found on the internet is real or fictitious. This can lead to the propagation of misinformation.
Their understanding is also limited to the most recent data on which they have been trained. ChatGPT 4, for example, was trained on data spanning until late 2021.
This makes generative AI rather useless for writing about current affairs or any subject that’s experiencing rapid change.
Generative AI lacks originality and emotion
Generative AI lacks originality and emotion. Pretty obviously, these models are designed to regurgitate relevant information without employing any feelings or subjective thinking.
This makes them useful for fact-finding but ineffective at writing compelling content that offers a unique perspective and stands out from the crowd.
Generative AI also has a propensity to hallucinate.
This happens when the model recites inaccurate information but expresses it in such a natural manner that it appears to be authoritative, making such inaccuracies perilously difficult to detect.
Content marketers can start by considering four broad questions:
> How will generative AI impact our work in the short and long term?
> How much risk are we taking if we experiment with generative AI?
> Where should we deploy generative AI first?
> Do we have the people, process, and technology to do this?
Forming a sensible AI strategy requires understanding how the technology might affect your work today and in the long run.
Research suggests that most of the early beneficiaries of generative AI are found in software engineering, marketing and sales, customer service, and product development.
Consequently, the biggest impacts are being felt by industries that rely heavily on those functions, such as media and entertainment, consumer goods, telecoms, medicine, and the tech sector.
Nevertheless, generative AI will eventually impact worker productivity across all industries as use cases and customized applications proliferate.
There are immediate opportunities to improve efficiency and effectiveness in practically any domain or department.
Failing to spot the quick wins in your own neck of the woods could leave you trailing behind competitors that can work faster and more accurately with generative AI helping them write code and content.
No one can predict the full implications of generative AI, but it is also important to think further out. How might AI change the competitive environment? Where are you most vulnerable and what, if anything, can you do to futureproof your strategy and business model?
Does this outlook lead you to pursue AI more aggressively or more cautiously?
Deciding how much to mess with generative AI requires finding a balance between potential value creation and risk taking.
The litany of spectacular accomplishments being attributed to generative AI inevitably makes leadership teams eager to begin adopting it and capturing its value.
But generative AI, if not well managed, also has the potential to destroy brand value and reputations.
The widespread use of AI is already raising privacy concerns and ethical questions, such as what the potential impact might be of hidden biases in training data.
There are sustainability concerns because running AI models requires jaw-dropping computing capacity and the associated energy supply.
Could it also raise the risk of a cybersecurity breach?
What about AI-generated deepfake videos that impersonate your company leaders, undermining your brand reputation?
What about AI-generated deepfake videos that impersonate your company leaders, undermining your brand reputation?
Existing risks are also amplified, such as the risk of infringing copyright, trademarks, and other protected information that was inadvertently collected by the generative AI model.
Try to evaluate both the value and the risks of each use case and how they align with your company’s business objectives and risk tolerance.
We’ve written many times about the importance of consistency and authenticity in effective content marketing.
Random acts of marketing don’t work.
So, it follows that sporadic applications of generative AI won’t be helpful, either. You will need a more coordinated approach.
There are several possible starting points to consider, such as:
> Efficient, comprehensive research – With the right prompts, generative AI can complete arduous research tasks in mere seconds. However, as we have discussed, beware inaccurate, outdated, and incomplete information.
> Content development – While you don’t want to put AI completely in control of creating your content—if nothing else, to retain a modicum of authenticity—it can be very helpful for ideation, outlining, and headline writing. Use it for jumping-off points not full-blown copywriting.
> Improving your writing – Generative AI can produce enhanced versions of an existing piece of content based on optimization areas supplied in a prompt—for example, readability, tone, style, keywords, or sectioning and headlines. At a grand scale, these could be A/B tested to discover which style your audience likes best.
Appoint someone to take the lead—most likely your Managing Editor or someone in a similarly significant position—and assign them responsibility for all generative AI activities.
A sensible second step is to tie your marketing AI efforts into a wider, cross-functional group. This will help the organization take a coordinated approach to multiple use cases, as well as raising awareness of where AI is being deployed and its potential impact.
You must also plan regular check-ins and reviews to assess the impact and efficacy of ongoing generative AI projects.
If any area of content marketing where generative AI is being used shows signs of stress or underperformance, don’t blindly assume things will get better.
To effectively take advantage of generative AI, you will need the usual trifecta: people, processes, and technology.
People
Many companies are already struggling to adjust their skills base amid the digital transition.
The shift from in-person to online purchasing at B2B businesses is no exception. Many vendors are scrambling to catch up, finding themselves ill equipped to deliver the content marketing that’s needed.
The introduction of generative AI has added a new wrinkle to reskilling the workforce for a data-driven world.
Generative AI puts a premium on analytical and creative skills, not to mention the newly popularized art of prompt writing
Generative AI puts a premium on analytical and creative skills, not to mention the newly popularized art of prompt writing.
Fortunately, off-the-shelf generative AI models will be sufficient to handle most marketing use cases. This means only generalist AI skills should be needed—supported by data and software engineers.
Those AI skills must be added to marketing job descriptions, included in hiring needs, and added to training plans for your existing team members.
You will also need appropriate support from the leadership team level.
Unless your CEO and other senior leaders understand generative AI and its implications, they won’t be able to gauge its likely impact or make appropriate decisions about time investment, risks, hiring, or technology.
Since generative AI is new to almost everyone, accessing sufficient expertise might mean engaging an AI consultant.
Processes
Effective content marketing teams depend heavily on a repeatable, robust process for delivering consistent, high-quality content.
On the one hand, therefore, generative AI represents a threat since it will likely disrupt the smooth-running process—at least initially.
Does your team have the sort of learning culture that will be needed for this to succeed?
Do you have processes in place to ensure the responsibility and accountability needed to manage the factual and ethical risks we described earlier?
Additional steps will be required to experiment with generative AI, review and edit its outputs, and to augment them with authentic, unique opinions.
You must ensure strict human oversight whenever you publish content written using generative AI—especially if the information will be used by the public in any non-trivial way.
Technology
At the broader business level, a company must have a modern data and technology stack to successfully deploy generative AI.
Within the marketing domain, however, basic models can support a wide range of use cases without having to be fed additional, proprietary data.
While it’s tempting to just start playing with generative AI—especially the miraculously easy to use, chat-based tools—it’s worth hitting pause and plotting a careful path.
Unleashing AI without first considering the possible consequences could destroy brand equity that you’ve spent a lot of time building, undermine your company’s reputation, and run it afoul of even more serious ethical and legal concerns.
Make sure you have the right people and processes in place to deploy generative AI in a coordinated manner and review the outcomes regularly to help nip any unintended consequences in the bud.
Stay in the know. Our weekly newsletter delivers the latest blog content, a selection of B2B content marketing insights gathered from across the web, and quick, actionable tips for taking your content marketing to the next level.
Sign up here!
Image credits: Adobe Stock
