June New York Roundtable Recap

Catch Up on What You Missed!

1. Setting the Context

On June 3, 2026, DCA and CardLinx convened a cross-functional group of leaders from across financial

services, loyalty and rewards, card-linked offers, commerce media, influencer marketing, travel rewards, data

strategy, and emerging AI shopping platforms. The session began with a deliberately broader question than

“Should financial services use influencers?”

The question was:

How is influence changing across the full commerce journey, and what does that mean for financial

institutions, loyalty platforms, fintechs, retailers, networks, and technology providers?

Participants entered the room with different relationships to the topic. Some were practitioners in influencer

and employee advocacy. Some were financial-services leaders wrestling with governance, risk, and

reputational constraints. Others were B2B platforms trying to understand whether influence mattered for

enterprise sales, partner credibility, and customer education. By the end of the conversation, the group had

reframed influence as a layered system that now includes creators, employees, customers, partners,

communities, platforms, and AI agents.

2. How We Framed Influence

The roundtable quickly moved beyond the familiar image of a social media celebrity endorsing a product.

Influence was described as the set of human, institutional, social, and technological signals that shape what

people notice, trust, consider, and ultimately choose.

That led to a wider frame: influence is not one channel. It is an operating layer across the customer journey.

• At the awareness stage, influence can come from celebrities, platform-native creators, podcasts, short-

form video, and cultural moments.

• At the consideration stage, influence can come from credible niche experts, customer testimonials,

employee-generated content, local community voices, or B2B thought leadership.

• At the conversion stage, influence can come from affiliates, offers, rewards optimization, card-linked

value, and increasingly AI shopping assistants.

• After the transaction, influence can come from customer advocacy, partner credibility, community

impact, and the visible behavior of employees.

This broader framing mattered because it allowed participants to see that financial services may already be

using forms of influence without naming them that way: branch-level community service, mobile financial

centers after disasters, partner testimonials, rewards platforms, and client success stories all shape trust and

brand preference.

3. Influence Has Become Layered, Not Linear

A useful model emerged around different influence layers. Traditional influencer marketing still matters, but it

is only one part of the picture. The group discussed mega influencers, mid-tier creators, micro and nano

creators, employee advocates, B2B partners, customers, platform-native creators, and AI agents.

Each layer carries a different mix of reach, credibility, cost, risk, and measurability.

• Mega influencers can create visibility and board-level recognition, but they are expensive and carry

significant reputational risk.

• Mid-tier and micro creators can bring credibility, niche relevance, and content volume without tying the

brand to one highly visible public figure.

• Nano and hyper-local voices can shape community trust in ways national campaigns often cannot.

• Employees can offer authenticity and local relevance, but only if participation is governed, trained, and

supported.

• B2B partners and customers may be the most important influencers for enterprise platforms because

they create trust through proof, networks, and shared context.

• AI agents are becoming a new influence layer because they will increasingly filter, rank, recommend, and

perhaps execute choices on behalf of consumers.

The practical implication is that organizations should not ask whether they need “an influencer strategy” in the

narrow sense. They should ask which influence layer is best suited to the outcome they are trying to create.

4. “Influence for What?” Became the Central Strategic Filter

One of the clearest strategic insights came late in the conversation: influence has to be tied to a purpose. The

group repeatedly returned to the question of whether the goal was awareness, credibility, acquisition,

reputation repair, community connection, employee engagement, sales conversion, rewards optimization, or

AI/LLM visibility.

This distinction prevents mismatched strategies and misleading measurement.

• If the goal is credibility, then citations, expert references, and thought leadership may matter more than

follower count.

• If the goal is knownness for a B2B solution, credible voices inside the industry may matter more than

mass-market reach.

• If the goal is emotional connection, human stories and creator authenticity matter more than structured

product data.

• If the goal is agentic discovery, machine-readable, specific, data-rich content may matter more than

polished creative.

• If the goal is conversion, affiliate content, offers, rewards, and last-mile transaction paths become

central.

This was especially important for participants operating in B2B-to-C environments. A provider may not market

directly to the end consumer, but it may still influence the financial institution, the issuer, the offer ecosystem,

or the agent-mediated decision layer that determines which product or offer surfaces.

5. Employee Advocacy Was Seen as High-Potential but Operationally Demanding

Employee advocacy was one of the most substantive areas of discussion. Several participants described

programs in which employees create or share content that reflects local stories, product availability,

community moments, hiring narratives, or service experiences. The strongest examples came from retail, tax

services, QSR, telecom, and other sectors with distributed workforces and local customer touchpoints.

The promise was clear: employees can bring a level of authenticity, specificity, and local relevance that

national campaigns struggle to match. In some examples, employee-generated content substantially

outperformed paid influencer content in engagement and sentiment. Participants also noted that strong

creators are not always the youngest employees; in several large retail contexts, some of the most effective

employee creators were older employees with distinctive voices and strong community resonance.

But the group also surfaced real complexity:

• Manual programs often stall once they reach roughly 100 advocates because permissioning, review,

asset collection, and measurement become too difficult.

• Scaling requires coordination across marketing, brand, legal, compliance, HR, field leadership, and

sometimes labor relations.

• Employees need guardrails not only to protect the brand but also to feel safe participating.

• The more regulated the industry, the more likely participation rules will become so complex that

employees disengage.

• The authenticity question is different for employees than customers: an employee can be genuine, but

audiences may still perceive bias because the person works for the company.

A pilot-first approach emerged as the most credible path: start with a trusted cohort, study what works, build

the right brand-safety and compliance tools, develop training, then expand gradually.

6. Financial Services Faces a Distinct Governance Burden

The financial-services participants made clear that influencer and advocacy programs cannot be evaluated

only through a marketing lens. For banks and credit-related businesses, governance is not a late-stage review

step; it shapes what is possible from the beginning.

Several constraints stood out:

• Fair lending and disparate impact concerns affect targeting, even when the product being marketed is a

loyalty or rewards experience rather than a lending product in the narrow sense.

• FCRA-governed data requires fairness testing and careful usage controls.

• GLB and privacy obligations add another layer to how data can be used and combined.

• Reputational risk functions as a separate filter, often framed around the headline that would appear if

something went wrong.

• Compliance timing can make time-sensitive content stale before it is approved.

The group also identified an internal operating problem: many people inside large organizations can stop an

initiative, but far fewer can approve one. This creates a structural bias toward delay. Participants did not argue

for abandoning compliance; rather, they argued that the pace of market change requires faster ways to

identify, mitigate, and govern risk instead of repeatedly deferring action.

7. Trust Is Both the Barrier and the Product

Trust connected nearly every part of the conversation. It appeared in different forms: trust in influencers, trust

in employees, trust in brand claims, trust in AI-generated content, trust in agentic transactions, trust in

financial institutions, and trust in compliance systems.

The group distinguished between formal control and perceived authenticity. A campaign can be legally

reviewed and still feel inauthentic. A creator can be authentic and still create reputational risk. An employee

can be close to the customer and still raise questions about bias. An AI agent can optimize value and still

provoke discomfort if it crosses from recommendation into financial authority.

The practical conclusion was that trust must be designed, not assumed. That means:

• clear rules for what advocates can and cannot say;

• transparent disclosure of relationships;

• contractual protections, including takedown rights and morality clauses where appropriate;

• background checks and audience analysis for creator partnerships;

• systems that protect employees from accidentally violating guidelines;

• human review for judgment-sensitive content; and

• careful calibration of how much autonomy consumers are asked to give AI agents.

8. Technology Can Scale Influence, but It Cannot Eliminate Judgment

The group discussed technology as both an enabler and a risk multiplier. Employee advocacy platforms can

scan text, images, audio, captions, overlays, and metadata. They can apply brand guidelines, social policies,

role-specific rules, and industry regulations. They can flag sensitive visuals, restricted terms, product claims,

and content that fails to match a creative brief. They can also support gamification, rights management, asset

collection, and paid amplification of high-performing content.

At the same time, participants were cautious about assuming that AI can replace human judgment. One

example captured the issue: a content post can be technically compliant but still feel odd, off-brand, or

contextually wrong. AI may catch prohibited terms, but it may not know that a sandwich photographed on a

shag rug is simply not the right representation of a food brand.

The emerging consensus was that AI can accelerate review, identify risk, and make scaled programs feasible.

But in regulated or reputation-sensitive contexts, human accountability remains necessary. The opportunity is

not to remove people from the loop; it is to use technology to make the human review process faster, more

consistent, and more focused on true judgment calls.

9. Agentic Commerce Changes the Meaning of Influence

The conversation eventually moved from human influence to agentic influence. Participants discussed how AI

agents and shopping assistants are beginning to shape product discovery, offer evaluation, rewards

optimization, and potentially transaction execution. This extended the Q1 roundtable discussion on agentic

commerce into a new question: if agents are influencing or executing purchases, who or what is the

influencer?

Several themes emerged:

• AI agents may become both the audience and the intermediary, requiring brands to persuade systems as

well as people.

• Consumers may be more comfortable allowing agents to optimize rewards, points, offers, or purchase

paths than allowing agents to spend money directly.

• Agentic commerce introduces new fraud and liability questions when an agent acts on bad information,

manipulated content, or an “agentic influencer.”

• The near-term opportunity may be rewards and value optimization rather than fully autonomous

spending.

• The infrastructure needs both deterministic transaction controls and probabilistic AI reasoning; one

participant described this as “bones” and “soft tissue.”

The group treated recent market announcements as evidence that agentic commerce is moving from

abstraction toward implementation. The significance is not that autonomous purchasing is already

mainstream. The significance is that infrastructure, guardrails, and commercial incentives are forming now.

10. Brands Now Need Dual Content Strategies

A major insight was that human-facing content and agent-facing content should not be treated as the same

asset with different distribution. They serve different audiences and should be measured differently.

Human-facing influence depends on emotional connection, simplicity, visual clarity, storytelling, identity,

trust, and resonance. It is measured through attention, engagement, sentiment, brand lift, recall, and eventual

downstream behavior.

Agent-facing influence depends on structure, specificity, credibility, breadth, freshness, machine readability,

and the ability to answer concrete queries. It is measured through citations, search visibility, LLM references,

recommendation frequency, and conversion pathways that may not look like traditional media attribution.

This distinction matters because a creator campaign designed for emotional loyalty may not produce the

structured signals that an AI system needs. Conversely, dense, data-rich content designed for agentic

discovery may not move a human audience emotionally. The strategic task is to coordinate both, not collapse

them into one.

11. Measurement Remains Underdeveloped

Participants repeatedly returned to the problem of measurement. Influencer marketing, employee advocacy,

affiliate commerce, LLM visibility, and agentic transactions each have different success metrics. Yet

organizations often try to evaluate them through familiar performance-marketing or last-click attribution

models.

The group identified several measurement gaps:

• How should financial institutions measure emotional influence when the desired outcome is reduced

price competition or increased brand preference?

• How should employee advocacy be evaluated when its value may include engagement, trust, recruiting,

community connection, sentiment, and sales?

• How should B2B platforms measure influence when the most important outcome may be credibility

among decision-makers rather than direct conversion?

• How should LLM or agent-facing content be measured when the outcome may be a citation,

recommendation, or ranked inclusion rather than a visible click?

• How should attribution work when a human influencer creates awareness but an AI agent completes the

transaction?

The shared implication was clear: influence strategy will not mature in financial services without better

measurement frameworks. The field needs metrics that match the actual objective, not just the channel that

carried the message.

12. Adoption Will Be Driven by Persona, Context, and Use Case

The conversation challenged simple generational assumptions. Younger audiences may follow YouTubers,

TikTok creators, Twitch streamers, and platform-native personalities more than traditional celebrities. But

older consumers are also influenced by online recommendations, podcasts, communities, and trusted digital

voices. The divide is not simply old versus young.

Adoption will vary by persona, trust posture, financial stakes, product category, and task complexity. A

consumer may trust an agent to find the best rewards path across 20 items, but not to initiate a large

transaction. A financial institution may tolerate employee posts about community service before allowing

employees to discuss product features. A B2B buyer may be more influenced by a peer case study than by a

public creator. A consumer may respond to a creator story in one context and an affiliate price alert in another.

This reinforces the need for segmented influence strategies. One voice will not reach every customer. One

format will not serve every purpose. One risk model will not fit every use case.

13. Strategic Postures in the Room

The roundtable surfaced several strategic postures:

• Practitioners were focused on proving that influence can be structured, governed, scaled, and

measured.

• Financial institutions were interested in the upside but constrained by fair lending, privacy, reputational

risk, and internal approval models.

• B2B and platform companies were thinking about influence as credibility, customer advocacy, and

partner-driven trust rather than mass-market awareness.

• Startups and agentic commerce platforms were focused on speed, infrastructure, rewards optimization,

and new forms of AI-mediated discovery.

• Data and technology providers were focused on identity, targeting, governance, citations, and the

tension between innovation and regulated use.

Despite those differences, the group aligned on one point: doing nothing is not risk-free. The market is moving

quickly, and the influence layer is becoming more complex. Organizations that wait for perfect certainty may

find that standards, customer expectations, and agentic discovery pathways have been shaped by others.

14. Questions That Remain Open

The conversation did not resolve every issue. It clarified the next set of questions DCA and CardLinx members

may need to explore:

• What are consumers actually comfortable delegating to AI agents on a crawl-walk-run basis?

• What governance model allows financial institutions to test influencer and employee advocacy

programs without unacceptable risk?

• How should brands structure content for both human emotional resonance and agentic discoverability?

• What does authenticity verification look like in a world of AI-generated content, deepfakes, and synthetic

endorsements?

• How should attribution work when influence is distributed across creators, employees, affiliates,

rewards platforms, agents, and transaction systems?

• Can agentic compliance accelerate review without removing human accountability?

• What metrics will convince skeptical executives that influence is worth the operational complexity?

• Which financial institution or fintech will move first at meaningful scale, and what advantage will that

create?

15. What We Left With

The group did not leave with a single answer. That was appropriate. Influence is no longer a discrete marketing

tactic, and agentic commerce is no longer a distant abstraction. Both are becoming operating conditions for

financial services, loyalty, and digital commerce.

We left with a clearer set of realities:

• Influence now extends across people, platforms, employees, partners, communities, content systems,

and AI agents.

• The most useful strategic question is not “Should we use influencers?” but “Influence for what?”

• Employee advocacy has high potential, but it requires tooling, training, governance, and careful scaling.

• Financial services cannot copy less-regulated playbooks without adapting for fair lending, privacy, and

reputational risk.

• Trust is the core constraint and the core product.

• AI can help scale compliance and content review, but human judgment still matters.

• Agentic commerce will make content, data, offers, and rewards increasingly machine-interpretable.

• Human-facing and agent-facing influence strategies must be designed separately and coordinated

intentionally.

• Measurement is the next frontier; without better metrics, adoption will remain uneven.

Perhaps most importantly, we left with a broadened definition of influence. It is not only who speaks for a

brand. It is what earns attention, what carries credibility, what reaches the customer at the right moment, and

what becomes visible to the systems that increasingly mediate choice.