What our actual clients are saying about GenAI tools

What’s the role of a content agency like Ascribe in the era of generative AI? It’s a valid question, given the growing prevalence of (and possibly preference for) AI-generated content. And it got us wondering about how our own clients are or are not using AI in their own work. So we interviewed several of them to ask about the platforms they’re using, the benefits they’re seeing and where the limitations are — with the goal of better understanding how we can continue to support them going forward. Here’s what we learned.

How are our clients using AI?

Given the vast number of AI tools that are out there, it wasn’t surprising to hear that our clients are using many different ones as part of their workflows. Several mentioned ChatGPT, of course, with one using the Pro version intended for researchers and engineers. Some are using alternatives such as Claude, IBM Edge and Grammarly, while others are using specialized AI tools like Canva (graphic design), Beautiful.ai (presentations), Perplexity (deep research with quality citations) and DeepL (translation). Others are relying on the AI assistants built into the platforms they already use: specifically, Adobe Firefly and Microsoft Copilot.

What are they doing with these tools? Our clients vary in size and sector, so there’s no single universal application of GenAI. But we learned they’re using AI to help with:

• Marketing content, including web pages, articles and social media posts
• Quotes and proposals
• Internal and client-facing emails (e.g., polishing for voice and style)
• Research
• Translation
• Image generation and editing
• Speaking notes (e.g., to help set the tone for a speech or presentation)
• Employee learning plans
• Strategy development
• Note-taking and meeting summaries

Over the next year or so, some of our clients said they’ll continue to explore and evaluate AI’s potential as the technology as it evolves, especially if functions like image generation get significantly better in the short term. Others said they might specifically branch out into areas of AI assistance that are new to them, like translation.

Why are our clients using AI?

For most of our clients, it’s all about efficiency and speed. For example, one mentioned how AI helps them run their organization’s social media account, something that would have been a full-time job not too long ago. Another talked about how AI empowers their staff to generate content or assets without having to go through the communications team every single time.

Others said AI is useful for sparking ideas or just getting a first draft out quickly, no matter how rough it is, so they can start iterating and adjusting. As one person said, “I’m better at editing something than drafting it from scratch.” It was also mentioned that AI can add new perspectives and insights to broaden their way of thinking.

But is it possible to quantify these benefits? Not yet. None of our clients indicated that they are tracking performance and engagement metrics of human- versus AI-generated content. One even said, “I don’t know if we’ve seen any cost savings yet.”

How is AI affecting our clients’ day-to-day workflows?

Our clients consistently mentioned two main challenges or limitations of GenAI:

• Accuracy and trustworthiness: Many of our clients operate in highly regulated industries or handle communications for organizations where public trust in the content they release is paramount. As such, they aren’t super confident that the content being generated by their AI tools is accurate or relevant to their needs. As one said, “It creates good first drafts, but I’m not confident enough beyond that.” Another mentioned feeling like they need to double-check everything created by the AI. That is, they can’t take anything AI produces at face value due to concerns it will have made something up or gotten crucial details wrong — and that’s simply unacceptable for government agencies, national associations or high-tech companies.

• Tone and style: Similarly, the confidence isn’t there when it comes to AI’s ability to reflect the brand voice of their organizations. “The tone is often not quite right,” said one client, while another expressed concern about the originality of the AI’s output. One client put things this way: “There’s rarely a first draft from an AI tool where I think, ‘Yeah, we can go with that.’” Another client noted that AI has difficulties consistently following their organization’s style guide.

These challenges are having a noticeable impact on their internal workflows. In general, while the initial act of content creation is accelerated and made more efficient by AI, the vast majority of our clients said more steps have had to be added into the publishing process to account for additional editing and reviewing of the AI-generated content, and then getting the necessary approvals.

Several clients said that, unlike the content created by our writers, AI-generated content must always be reviewed by their communications team or a content strategist before it can be published, with some having established a formal review process to ensure an AI’s output is seen by at least one additional person. This is adding complexity and slowing down the process of publishing content. Others are still in the process of planning and implementing policies around AI — like the one organization that has decided to use AI only for internal communications and never for client-facing material.

In addition to adding more time in editing and reviews, there’s also the considerable amount of time spent at the start of the “writing” process. A person needs to enter a highly detailed prompt telling the AI tool what to do. Prompts need to be tested, tweaked and then tested again — often multiple times — before the AI will generate a first draft that is somewhat usable. Prompt design was mentioned by a few clients as something that needs more attention within their organizations, as it can easily be overlooked when there’s a rush to get content out the door.

When is human expertise still better than AI?

Given the challenges they’re facing with AI, it wasn’t a surprise to hear that our clients still prefer the human touch for most types of content and campaigns. When asked about the areas they would continue to need support from a strategic content partner, they mentioned creative work (e.g., slogans, taglines), media relations, highly technical or long-form content (e.g., thought-leadership articles), executive briefings, research-based educational materials (where accuracy and credibility are paramount), translation, or anything requiring on-the-ground knowledge of what resonates with specific customers or members of the organization.

Similarly, our clients said human oversight is still needed across several parts of their operations, such as client interactions, crisis communications, interviews and in-person events. And they simply can’t use AI at all when it comes to financial data or government confidential information (i.e., anything higher than “Protected B”).

So what does this all mean? Based on everything we heard, there’s still a clear role for Ascribe’s writers and editors, as there’s still a lot of work that remains best suited for real people.

But we also can’t ignore the evolution of the needs of our clients. We know that, in the not-too-distant future, more and more of our clients will be leveraging generative AI in their day-to-day work. Which is why we’re evolving our service offering, too. We want them to be able to turn to us for guidance in areas like prompt engineering or training AI platforms on their organization’s tone of voice. And we can also take on the time-consuming editing and reviewing steps of the content publishing process — to make sure any AI output reads well, is accurate, and fully aligns with their brand voice and style.

Contact us today to see how we can bring your AI-generated content to the next level.