admin
The rapid ascent of generative artificial intelligence has presented a dual-edged sword for global organizations: the promise of unprecedented productivity and the peril of cultural disruption. For RTI International, a leading independent, nonprofit research institute, the challenge was not merely to adopt the technology but to integrate it into their corporate communications framework in a manner that was ethical, effective, and human-centric. Under the leadership of Kami Spangenberg, Vice President of Corporate Communications, the organization has pioneered a methodical, four-stage approach to AI adoption that prioritizes employee psychological safety and practical utility over rapid, top-down implementation. This strategy reflects a broader trend in the communications industry where "human-in-the-loop" systems are becoming the gold standard for navigating the complexities of automated content generation and data analysis.
The Genesis of the Gen AI Impact Team
RTI International’s journey began not with a mandate from the executive suite to use specific software, but with an open call for curiosity. Before any widespread rollout of AI tools, the Corporate Communications department recruited a diverse test group of volunteers. This group, which eventually branded itself the Gen AI Impact Team, was designed to be a microcosm of the larger workforce. It included employees across various career stages, from entry-level coordinators to senior directors, and spanned different tenures within the organization.
Crucially, the team was not composed solely of early adopters or "tech evangelists." Spangenberg intentionally sought out a spectrum of technological comfort levels. By including those who were naturally skeptical or "trepidatious" alongside those who were eager to experiment, the organization ensured that the feedback gathered would address the very real fears of job displacement and the learning curves associated with new interfaces. This inclusivity was a strategic choice; a group of power users might overlook the friction points that a less tech-savvy employee would encounter, leading to a rollout that could alienate a significant portion of the staff.
The Impact Team’s composition also reflected the diverse functions of RTI itself. Participants came from internal communications, public relations, and teams responsible for scientific and technical content. This cross-functional approach allowed the group to assess how AI could assist in varied tasks, from drafting employee newsletters to summarizing complex scientific reports tied to RTI’s mission of improving the human condition through multidisciplinary research.
Phase One: Controlled Experimentation and Freedom to Fail
Once the Gen AI Impact Team was established, Spangenberg initiated a two-month period of intensive experimentation. During this phase, the group was given limited funding to explore paid versions of prominent AI tools, such as ChatGPT Plus, Claude, and Midjourney, as well as specialized platforms designed for research and drafting. The leadership provided a high degree of autonomy, allowing the team to be self-directed in their explorations.
The primary objective was not to produce immediate deliverables but to evaluate the landscape. Spangenberg tasked the group with documenting their findings in a structured report that focused on four key areas:
- Identifying which tools provided the highest quality output for specific communication tasks.
- Assessing the ease of use and the time required for "prompt engineering."
- Noting the limitations, such as "hallucinations" (the tendency of AI to generate false information) or tone inconsistencies.
- Evaluating how these tools could be integrated into existing workflows without creating additional administrative burdens.
By framing this period as a "learning journey" rather than a software trial, Spangenberg lowered the stakes. The goal was information gathering, not the immediate replacement of human effort. This phase allowed the team to demystify the technology, moving from a place of intimidation to one of informed assessment.
Phase Two: Identifying Strategic Wins and Workflow Alignment
The conclusion of the two-month experiment provided RTI with a wealth of qualitative data. The Impact Team was able to categorize AI tools based on their strengths. For instance, they discovered that certain Large Language Models (LLMs) were exceptionally skilled at drafting internal memos and brainstormed subject lines, while others were more reliable for synthesizing long-form research papers into executive summaries.
These "early wins" were critical for building momentum. By identifying specific, low-risk tasks where AI could immediately save time, the team could demonstrate the technology’s value to the wider organization. For example, using AI to generate initial outlines for blog posts or to repurpose technical reports into social media snippets allowed communicators to focus more on high-level strategy and less on the "blank page" stage of writing.
During this phase, the team also began to identify the specific workflows that matched the tools’ functions. This prevented the "tool-first" trap, where an organization buys software and then tries to find a problem for it to solve. Instead, RTI adopted a "problem-first" approach, matching established communication needs with the most effective AI solutions.
Phase Three: The Guided Pilot and Departmental Expansion
Building on the success of the initial experiment, RTI moved into a broader pilot program. This stage involved expanding the use of AI tools to different departments within the communications umbrella. However, this was not a "free-for-all" release. The expansion was "guided," meaning the original Impact Team acted as mentors to their colleagues.
Employees in specific roles were paired with tools that the Impact Team had already vetted. For example, staff members focused on internal employee engagement were given access to tools optimized for conversational drafting and sentiment analysis. These employees were asked to use the tools during their regular work cycles and provide feedback via standardized forms.
This feedback loop was essential for avoiding what Spangenberg describes as the "giant wave" effect—the feeling of being overwhelmed by too many disconnected tools. By providing a structured environment where employees knew exactly what was expected of them, RTI maintained a sense of order. The feedback forms asked targeted questions: Did the tool save you time? Did the output require significant editing? Did the tool understand the organization’s specific brand voice? This data allowed the leadership to refine their strategy in real-time, ensuring that the AI adoption remained relevant to the actual work being performed.
Phase Four: Knowledge Sharing and Institutional Governance
The final phase of the current cycle involves the synthesis of all feedback into a cohesive institutional strategy. As employees across the department began using the tools, they naturally discovered "tips and tricks"—better ways to prompt the AI or creative ways to use it for data visualization. Spangenberg emphasized the importance of swapping these insights across the team to foster a culture of collective learning.
This phase is also where governance becomes paramount. The insights gathered from the pilot programs have helped RTI form a broader AI structure that includes ethical guidelines and quality control measures. In a research-heavy organization like RTI, the accuracy of information is non-negotiable. Therefore, the governance model emphasizes that AI is an assistant, not an author. Every piece of content generated or assisted by AI must undergo a rigorous "human-in-the-loop" review process to ensure scientific integrity and adherence to organizational values.
Industry Context: The Evolving Role of AI in Communications
The approach taken by RTI International aligns with emerging industry data regarding AI adoption. According to the 2024 Muck Rack "State of AI in PR" report, approximately 64% of public relations professionals are now using generative AI in some capacity, up significantly from the previous year. However, the same data suggests that "lack of training" and "ethical concerns" remain the primary barriers to deeper integration.
RTI’s method addresses these barriers directly. By providing a safe space for experimentation and a structured pilot program, they have effectively provided the necessary training while simultaneously building the ethical guardrails required for responsible use. Furthermore, Spangenberg’s focus on "relevance"—asking if a tool actually helps an employee do their job better—addresses a common critique in the tech industry: that AI is often a solution in search of a problem.
Analysis of Implications: Human-Centric Innovation
The RTI model suggests several key implications for the future of corporate communications:
- Psychological Safety as a Metric of Success: By including skeptics in the initial test group, RTI prioritized the emotional and professional well-being of its staff. This reduces the "AI anxiety" that often leads to quiet resistance or decreased morale during technological transitions.
- From Efficiency to Effectiveness: While AI is often touted as a way to do things faster, RTI’s approach focuses on doing things better. By automating the rote tasks of drafting and summarizing, communicators can dedicate more time to strategic storytelling and stakeholder relationship management.
- The Death of the "One-Time Rollout": Spangenberg’s assertion that AI adoption is a "journey" with no "final destination" reflects the reality of the rapidly changing tech landscape. Organizations must remain agile, constantly re-evaluating their tools as the technology evolves.
- Governance as a Creative Enabler: Rather than viewing rules and guidelines as a hindrance, RTI uses governance to provide the clarity employees need to experiment safely. When people know where the "off-limits" zones are, they feel more confident exploring the "safe" zones.
Conclusion: A Responsible Path Forward
As Kami Spangenberg prepares to share these insights at the upcoming PR Daily Conference, the RTI International case study serves as a blueprint for other organizations. The core lesson is that successful AI integration is 20% technology and 80% culture and process. By putting humans at the center of the transition, RTI has transformed a potentially disruptive force into a collaborative tool.
In an era where "automated" often becomes a synonym for "impersonal," RTI International is proving that generative AI, when handled with intentionality and inclusivity, can actually enhance the human element of communications. The journey continues, but the foundation—one of curiosity, rigorous testing, and collective learning—ensures that the organization is well-equipped to navigate whatever technological shifts the future may hold. By focusing on responsible adaptation rather than reactive adoption, RTI is not just keeping pace with the industry; it is setting the standard for the future of the profession.
