Five Takeaways from AMEC North American AI Day for PR Leaders
The conversations at AMEC Measurement and Evaluation’s North American AI Day reflected a shift in how information is discovered, how communications is governed, and how value is understood and measured. For PR leaders, the implication is that AI is fundamentally expanding the scope of the discipline itself. Across the sessions, and reinforced by emerging research, five ideas stand out as particularly important for communications leaders navigating what comes next.
1. We are now marketing to machines
Jonny Bentwood, Global President Data & Analytics, Golin & Louis DeCosmo, Sr Director, Corporate Brand Reputation Strategy and Insights, PepsiCo spoke at AMEC AI Day about how they made generative AI visibility a CCO-tracked KPI and implemented targeted optimizations tactics across channels.
For years, communicators have organized distribution across paid, earned, shared, and owned media. But from an audience perspective, discovery has typically happened through a smaller set of environments: media coverage, search, social feeds, and direct research. Increasingly, AI-generated answers are becoming another layer in that discovery process.
Large language models (LLM) and AI assistants now act as intermediaries between brands and audiences. They summarize, recommend, compare, and shape perception in ways that are not always visible or easily measured. As a result, PR teams need to expand how they think about visibility.
This shift is already underway. Yext’s Rise of AI Search Archetypes report describes a move away from traditional keyword-based behavior toward more conversational, intent-driven queries mediated by AI systems, which is changing how information is surfaced and trusted.
But visibility alone is not the full challenge. What matters increasingly is how AI systems interpret the information environment around a brand.
This is where the concept of context engineering becomes a strategic discipline for communications. At AMEC AI Day, Rob Key, CEO, Converseon, and Geoffrey Sidari, Founder & President, Airadis, described context engineering as the process of structuring and contextualizing media, social, and brand information so AI systems can retrieve and use it more accurately and with fewer hallucinations.
That framing aligns with how the concept is being defined more broadly in AI development. LangChain describes context engineering as providing “the right information and tools, in the right format” so an AI system can perform effectively. OpenAI has similarly emphasized that high-quality outputs depend on assembling the right evidence before generation begins.
For PR leaders, this reframes the challenge. The goal is not simply to appear in AI-generated answers, but to shape the underlying context those systems rely on through authoritative coverage, structured brand information, credible third-party validation, and content aligned to how real people ask questions.
During the panel, “The New Answer Engine Era,” Devon Bottomley, Head of Research & Analytics, Prosek Partners & Siqi Jiang, Senior Lead, Insights & Analytics, Codeword spoke to how agencies can audit AI systems for visibility, citation, and accuracy while using purposeful prompting.
“Share of voice” remains a useful construct. But in an AI-mediated environment, a complementary question is emerging of, “what is our share of answers and what evidence is shaping those answers?” Seen this way, marketing to machines is a new form of information stewardship that sits squarely within the evolving remit of PR.
| Food for Thought As you think about your own organization, where is your brand or client’s brand showing up in AI-generated answers today and just as importantly, where is it missing? What sources are shaping those answers, and to what extent do you have influence over them? And if AI systems are increasingly interpreting the information environment on your behalf, how might you begin to more intentionally shape that context? |
2. “Human in the loop” is not a strategy
As AI adoption accelerates, “human in the loop” has become a common reassurance, but the concept is often too loosely defined to serve as a meaningful governance model. Jennifer Sanchis, Insights & Consulting Director, CARMA West, shared practical guidance at AMEC AI Day on how to build an “AI moat” by investing in skills AI cannot replace such as critical thinking, ethics, and strategic judgment.
Human oversight alone is not a “cure-all” for AI risk and can create a false sense of security if not paired with clear accountability structures and system design (IAPP).
Effective governance requires that humans:
- define the purpose and constraints of AI use,
- design meaningful checkpoints and escalation paths, and
- retain accountability for outcomes.
In a communications context, this matters because outputs shape narratives, reputations, and public understanding. A superficial review at the end of a process is not the same as intentional oversight throughout it.
| Food for Thought As you evaluate your current approach, where does true accountability sit within your AI-enabled workflows? Are you actively designing governance into the system, or relying on end-stage review? And what risks might be flying under the radar simply because no one has taken ownership yet? |
3. Synthetic audiences are a tool with limits (not crystal balls)
One of the most intriguing areas of discussion at AMEC AI Day was the use of synthetic audiences: AI-generated personas used to simulate how different groups might respond to messaging. Kaitlin Hileman, VP Data Innovation, Ketchum, walked through a few demos where a synthetic Gen-Z persona reacted to different messages around a new premium egg product.
Used well, these tools can accelerate testing, surface blind spots, and enable faster iteration. They offer a form of directional insight that can complement traditional research.
But emerging research suggests clear limitations. A study from the Nuremberg Institute for Market Decisions (NIM) found that while AI-generated respondents can improve efficiency, they tend to reflect mainstream patterns more strongly than real populations, often missing niche perspectives and emerging viewpoints. The researchers conclude that AI should be used to “complement, not replace, human insights.”
Similarly, academic research published in Scientific Reports found that while LLMs can approximate certain patterns in human responses, they still produce systematic errors and fail to capture the full variability of real-world opinion and behavior.
These findings reinforce that synthetic audiences are best understood as a testing environment. They can help refine messaging under controlled conditions, but they cannot fully replicate how real stakeholders will interpret, react, or behave, particularly in complex or high-stakes contexts. While the value of synthetic audiences increases when supported with a client’s in-depth target research, these tools are most effective when used within a broader research approach that continues to prioritize direct engagement, qualitative insight, and real-world validation.
| Food for Thought As you experiment with these tools, where can synthetic audiences genuinely accelerate your work? At what point do you need to bring in real human insight to validate decisions? And how do you guard against overconfidence in outputs that feel precise but may not reflect real-world complexity? |
4. Reputation should be understood as enterprise value
A recurring theme throughout AMEC AI Day and across recent research is the need to move beyond describing reputation as a “soft” outcome.
Burson’s Global Reputation Economy report frames reputation as a measurable driver of enterprise value, linking it to trust, stakeholder behavior, and long-term performance. Similarly, Echo Research’s $13.8 Trillion in Plain Sight report estimates the scale of intangible value tied to reputation across global markets, cementing the idea that reputation is foundational.
For PR agencies and in-house teams, this creates an opportunity to translate communications into the language of business:
- How does trust influence customer acquisition and retention?
- How does credibility affect partnerships and market access?
- How does reputation shape resilience in times of crisis?
The more PR can connect its work to enterprise value, the more central it becomes to strategic decision-making.
| Food for Thought When you look at your current measurement approach, how are you actually quantifying reputation today? Which metrics truly matter to your leadership teams? And where is there an opportunity to more clearly connect your communications work to enterprise value? |
5. We need to consider the ROI of proactive issues management
If reputation is an asset, then issues management is one of the primary ways it is protected. The LexisNexis keynote session at AMEC AI Day with Kate LaVail, PhD, Segment General Manager for PR, Communications, Media, and Government & Nicola Johns, Principal Product Manager, reinforced how early intervention matters.
Identifying emerging risks—whether reputational, operational, or societal—and addressing them proactively can prevent significantly larger consequences later. The challenge has been articulating that value in terms that resonate with business leaders.
Framing issues management as risk mitigation helps close that gap. It shifts the conversation from reactive crisis response to proactive investment:
- monitoring signals,
- preparing response scenarios, and
- engaging stakeholders early.
The business case becomes clearer when viewed through the lens of avoided loss. Research suggests the cost of a major issue often extends well beyond the immediate impact. McKinsey & Company’s risk report, for example, found that across nearly 500 operational-risk events, companies experienced an average $1.9 billion decline in shareholder value over 120 days, which is approximately 3.7 times the average direct loss.
At the same time, those costs are rarely captured in a single line item. As PwC notes in their 2026 Global Digital Trust survey, reactive expenses tend to be distributed across legal, communications, operations, IT, marketing, and lost opportunities—making them easy to underestimate and difficult to attribute to any one function.
For PR leaders, issues management is part of a broader enterprise risk strategy, where early detection, preparedness, and response can help preserve both reputation and long-term value.
| Food for Thought As you think about your current approach, what risks are you already seeing but not yet acting on? How are you quantifying the cost of inaction or are those costs still invisible across the organization? And how might you more effectively communicate the value of prevention to leadership? |
A broader shift for PR leaders
These five ideas point to a broader evolution in the role of communications. PR sits at the intersection of:
- how information is discovered (including by AI systems),
- how organizations govern emerging technologies,
- how audiences are understood and modeled,
- how value is defined and measured, and
- how risk is anticipated and managed.
The opportunity and responsibility is to help clients and stakeholders navigate this expanded landscape, which means asking more precise questions such as:
- How is our organization represented in AI-mediated environments?
- What does responsible AI use look like in communications?
- Where can new tools surface reliable “signals” but rely on human judgment for the narrative and recommended action?
- How do we connect reputation to outcomes that matter to leadership?
- And how do we make the case for proactive investment, not just reactive response?
