While opinions about the broader impact of generative AI on publishing vary, there’s no denying the technology is poised to significantly impact many traditional publishing workflows. And as one who closely follows the development of AI technologies, I’ve seen firsthand how the integration of AI is not some distant future for publishing but a present reality.
My own experiments with AI tools so far have demonstrated AI’s significant potential across a range of publishing operations. For example, when it comes to contract analysis, AI algorithms are adept at identifying and dissecting things like key royalty clauses and financial breakpoints crucial for both negotiation and compliance. AI has also shown promise in terms of digital conversion, such as turning OCR’d text into cleanly tagged HTML. AI can also help analyze and optimize key aspects of the supply chain, such as warehouse throughput. And AI tools are also proving to be versatile assistants in the creative process.
The emergence of agent-based workflows hold significant promise for the publishing industry. Inspired by thought leaders like Andrew Ng and propelled by recent discussions at industry conferences, agent-based workflows are iterative or procedural models in which a large language model refines its outputs through multiple steps that mimic human processes—for example, writing a draft, editing it, and then shepherding it through production.
Publishing workflows are especially ripe for AI integration that enables the automation of rote tasks and AI-driven support for more complex decisions and judgments. And in some areas, notably the STM sector, this integration is already happening. In a recent email conversation, Pradeep Jain, from automation-based content services company Intect, told me that it is becoming increasingly common to see workflows in which a person is assisted by multiple AI “agents,” with the humans focusing on the more intellectual tasks and AI performing the routine ones. An editor, for instance, might review a manuscript for writing quality and subject matter accuracy, and use AI tools for reference formatting and checking image accessibility.
The synergy between AI tools and workflow management platforms (such as Jira and Asana) represents another opportunity for publishers. Integrating AI agents and workflow assistants into project management systems can provide a cohesive and trusted human-AI ecosystem that can enhance productivity and foster greater innovation at scale.
“
It’s not just about adopting new technologies to carry out our work—it’s about rethinking how we approach our work.
”
In my work with AI so far, it has also become clear to me that the true potential of generative AI in publishing lies beyond its current applications. But to tap this potential, we must engage those who know these tasks and workflows best—our employees. Only by leveraging their insights can we create AI tools that can refine and expedite the less engaging, more repetitive aspects of their roles. And successfully engaging employees requires a plan of action.
At my current employer, we have initiated a program to catalyze the adoption of AI tools. It includes providing a corporate subscription to OpenAI, and training to help employees understand key concepts, such as how to build effective prompts and to identify the red lines they shouldn’t cross (for example, creating content that will go into our publications). We are also building a shared library of effective AI prompts (Ethan Mollick, entrepreneurship professor at The Wharton School and a well-known AI commentator calls this a grimoire—a shared productivity spell book). Furthermore, we are implementing an incentive program to recognize and reward those who develop prompts that provide value to multiple individuals.
Of course, challenges remain—including the lack of a robust feedback mechanism to gauge the impact of the shared prompts, and a means of integrating them effectively into everyday workflows. Addressing these issues and others will be crucial as we refine our approach.
As we enter a new era in publishing, integrating generative AI into our workflows presents an exciting opportunity. It’s not just about adopting new technologies to carry out our work—it’s about rethinking how we approach our work. It’s about cultivating a culture of innovation and collaboration. The journey is just beginning. But we can already see that the possibilities are vast.
Columnist Ken Brooks is the founder of the consulting firm Treadwell Media Group and is a founding partner of Publishing Technology Partners.
A version of this article appeared in the 05/27/2024 issue of Publishers Weekly under the headline: