Harnessing the Value, Volume, and Velocity of AI-Generated Marketing
How to move fast in an AI-powered world without sacrificing quality or brand integrity
A recent Gartner poll shows that 55% of companies are already piloting or using generative artificial intelligence (GenAI), and that by 2025, 30% of outbound marketing will be produced by GenAI.
And why not? The use of lightning-fast GenAI for content creation allows your marketing team to reach a wider audience more often, while freeing up bandwidth to focus on more creative endeavors. It can also allow marketers to experiment with messaging more quickly and cost-effectively.
GenAI is a powerful tool that can unlock immeasurable value.
But this exciting innovation also brings new risks: As GenAI rapidly creates a larger volume of content, it opens the door for potential errors, like factual and grammatical mistakes and messages misaligned with your audience. If marketers aren’t careful, these errors may slip through the cracks of the vetting process, gumming up an otherwise smooth copyedit cycle and resulting in embarrassing missteps.
Effective mitigation of these risks requires a holistic approach that incorporates several strategies. Reviewing the data sources that train GenAI models and applying safety guardrails can help prevent undesired outputs.
Marketing teams must also optimize downstream processes to ensure quality content and position their company to take full advantage of the benefits of GenAI.
As you deploy GenAI, you’d be wise to also adopt a process for reviewing that content.
The continued importance of the internal review
GenAI’s impact at the top of your content supply chain is undeniable: it accelerates content creation, eliminating bottlenecks at the top of the workstream. However, companies often overlook its impact on downstream processes, where new traffic jams can occur.
GenAI models find words and phrases that flow well and present a seemingly rational response, but they’re not necessarily trained to return responses that are accurate, recent, or precise. They can sometimes present “hallucinations” — nonsensical outputs based on false information. Unfortunately, we have no easy way to identify which pieces of data the models based their responses on or to vet the information the models give.
Because of this limitation, creative teams, subject matter experts, the legal team, and marketers must review outputs before publication. This is similar to the process you’d follow with human-generated content; we all know our internal teams must review material before it’s set free into the world. You wouldn’t assign a content creation job to an intern without reviewing their work; we follow the same principle here.
However, as GenAI multiplies the volume and speed of content, reviewing every piece becomes more difficult. It can be easy to allow a few messages to slip through the cracks as we chip away at the sheer volume. So, how do you eliminate that bottleneck while ensuring quality and consistency?
Assessing risk and planning accordingly
A tiered review process, in which you assess the risk level of each piece of content you generate, will allow you to eliminate that review bottleneck and take GenAI-created content to the next level where it truly resonates with the audience, giving you the greatest benefits from its use. For this to work, you need to divide your content into three categories:
1. Content that hasn’t been reviewed or reached an audience yet
This content is at the highest risk for misalignment or legal or usage errors. Category 1 content requires a team of reviewers with deep knowledge of the subject. Your legal team will need to review the entire document for any applicable regulatory requirements and to approve any claims the company makes. Meanwhile, marketers will need to bring a creative eye to the content created by the AI model: How can they optimize it for the best customer experience? This requires an understanding of customer needs and the best way to appeal to different segments, something AI can’t do. It means incorporating visuals that will resonate with the audience and correcting factual or grammatical errors.
2. Content that you’re creating a different variation of
For instance, content that is already testing well with specific audiences for which you’re looking to adapt for new segments. This partially modified content has likely been reviewed and approved before, so it won’t need as intensive a review as a net-new piece. The legal team may not need to inspect this content, or they may only need to review a part of it. The burden on this tier falls on the creative team, but the revisions might be minor, requiring fewer resources.
3. Content that just needs cosmetic revisions, like a styling variation
For example, content that doesn’t change, but you’re distributing it to a different audience. This likely doesn’t require a subject specialist or legal review, but the marketing reviewer must ensure the content is customized for the new audience. This is likely a quick review.
This tiered system allows you to allocate the proper resources for review. It’ll take longer to review new content than to check content variation. By creating an accurate view of the new content and building a longer review cycle into the content supply chain, you can better plan your content distribution cycles and resources, avoiding downstream bottlenecks.
Once you’ve established your different tiers and are comfortable with how that process works, you can introduce automation into the review and approval process. For regulatory and quality reasons, you’ll still need a human to approve content. It’s similar to how a doctor still needs to examine you when you’re sick. Even though AI assistance can provide input on an initial diagnosis, a doctor still needs to conduct an exam, review the diagnosis, make sure nothing is missing from the record, and write the prescription.
More tips for creating efficiencies
By grouping your content into three categories, you ensure each piece gets the appropriate review. You can streamline the review process further by incorporating the following practices:
- Become familiar with what GenAI does and how it works, so you and your team know what to expect from its generated content. For example, maybe it slips up when trying to use modern metaphors. When you learn its patterns, you can look for them and make those revisions faster.
- Integrate workflow tools with creative applications so people working on a photo, video, ad, or other content will automatically be notified when it’s their turn to take over. This allows them to efficiently collaborate on the project from a single program instead of having to switch programs to complete each part of the assignment. This not only limits the likelihood of human error but also eliminates version control issues.
- Deploy a capacity-planning solution to allow you to see who has the availability to take on a new project or complete the next step of a collaborative project. This will prevent bottlenecks when a member of the staff can’t complete a task, which, as a result, joins the pile of stalled projects.
- Integrate workflow tools, such as Workfront, with content distribution platforms like Adobe Experience Manager and marketing automation tools like Marketo, so that when content is approved to go out, it’s automatically sent for final assembly. Here’s where automation can be a strong advantage.
It’s time to streamline your content supply chain
GenAI is an exciting technology that can get your message out to a broader audience more frequently. Who doesn’t want that? You can meet an increased demand for marketing and digital content. Because you can so quickly and easily revamp content, you’ll be able to change your messages on the fly to better engage with your audience.
But you must ensure that your company is operationally ready for this new volume and velocity. As your company adopts AI, it’s an excellent opportunity to refine your content supply chain for this new environment, which is iterative rather than linear. By investing the time to ensure that every step of your content supply chain is as efficient and automated as it can be, you’ll be able to keep the quality content flowing.
Slalom is a next-generation professional services company creating value at the intersection of business, technology, and humanity. Learn more about Slalom’s partnership with Adobe.
