Early Warning, Big Spaceship Anchor AI Ad Scale in Human Oversight

Where AI Belongs in Marketing: Human Judgment, Real Oversight With Big Spaceship and Early Warning — Ai4

Ai4

While generative tools eliminate physical shoot limits to produce infinite asset permutations, marketing teams face a mounting review burden that demands human gatekeeping to safeguard brand tone and emotional resonance.

AI-assisted summary of Where AI Belongs in Marketing: Human Judgment, Real Oversight With Big Spaceship and Early Warning — original source. Check the original for full context.

Source published:

Detailed notes

## Background Published by Ai4 on YouTube on August 20, 2026, this session features Big Spaceship CEO Taryn Crouthers and Early Warning CMO Andrea Gilman. The presentation explores where generative AI fits into modern marketing workflows and how client-agency partnerships can leverage new tools while maintaining rigorous human oversight, brand integrity, and creative judgment.

## TL;DR Drawing from their collaborative work launching the Paze digital checkout campaign, Andrea Gilman and Taryn Crouthers outline a structured approach to integrating AI into creative advertising. They break down the creative pipeline into human-only brainstorming, AI-augmented asset systems and photorealistic pre-visualization, and scalable personalization that still requires human sign-off. The discussion highlights that 30-second generative outputs rely on 30 years of accumulated human craft, while cautioning against the mounting review burden created by infinite asset variations. Ultimately, they argue that AI-driven time savings should be intentionally redirected into deep strategic thinking, relationship building, and creative joy.

## Atmosphere & style The presentation carries an energetic, pragmatic, and collegiate stage presence characterized by easy banter and mutual respect between client and agency leaders. Grounded in actual production war stories rather than speculative tech hype, the dialogue addresses employee anxiety with candid empathy and clear-eyed operational insight. It feels thoughtful and empowering, pivoting the industry conversation from mere cost-cutting efficiency toward elevating human craft, emotional nuance, and strategic clarity.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=4s) Marketers must deliberately identify and map where human judgment and oversight belong across the creative pipeline to prevent brand inconsistency and context errors. - [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) Viewing AI solely as a tool for speed and cost reduction overlooks its true strategic value in freeing teams to tackle complex business problems from fresh angles. - [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) Executing an AI prompt in 30 seconds requires 30 years of accumulated human craft, taste, and domain experience to guide the tool effectively. - [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) Brainstorming and conceptual development must remain human-led because generative AI cannot read a room or evaluate authentic emotional nuance and humor. - [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) Campaign development is transitioning away from rigid, fixed asset counts toward dynamic, scalable asset systems tailored across multiple formats and audiences. - [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) Combining high-end base physical photography with modular generative styling allows marketing teams to react to cultural trends without funding full commercial reshoots. - [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) High-fidelity photorealistic pre-visualization eliminates guesswork and accelerates executive buy-in across complex stakeholder networks before production spending begins. - [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) Moving from linear production to system-based generative workflows exponentially expands the review burden, requiring automated brand checkers paired with mandatory human approval. - [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) Time saved through generative automation should be intentionally reinvested in deep thinking, cross-functional collaboration, and creative joy rather than added administrative volume. - [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) Sustainable adoption requires marketing leaders to establish explicit governance principles regarding human accountability and ethical oversight at every stage of production.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=4s) Introduction and the Core Questions of AI in Marketing

The session kicks off with pre-lunch energy as Andrea Gilman, Chief Marketing Officer at Early Warning Services, and Taryn Crouthers, CEO of Big Spaceship, take the stage. Gilman opens by prompting the audience to consider two foundational questions regarding the role of AI in the creative process. First, while generative AI can produce impressive outputs—such as clever taglines, striking visual creative, or polished copywriting—marketers must critically ask where human judgment and oversight actually entered the process. Second, organizational leaders must evaluate whether their own creative and operational teams clearly understand where human oversight must be inserted.

Gilman emphasizes that these questions are urgent because AI systems are prone to lapses in judgment, knowledge omissions, context gaps, and brand inconsistencies. As powerful as the technology is, it cannot replace human discernment. Therefore, mapping out where human judgment belongs across the creative pipeline is essential for maintaining brand integrity and accountability.

- AI can produce taglines, creative imagery, and copy rapidly, but without guardrails, it risks brand inconsistency and judgment errors. - Creative teams must clearly define at what operational stages human intervention and oversight take place.

---

[01:29](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=89s) The Partnership: Explorers, Digital Innovation, and Paze

After a playful microphone shuffle, Taryn Crouthers introduces herself as CEO of Big Spaceship, an independent creative agency with a 25-year heritage in digital innovation and integrated campaigns. Big Spaceship serves as the agency of record for Paze, a digital wallet and online checkout solution developed by Early Warning Services. Crouthers notes that having a name like Big Spaceship inevitably leads to endless spaceship puns, but the studio leans into the identity of being "explorers, not settlers." In an era dominated by rapid technological change, Crouthers argues that an obsession with curiosity and human-centered storytelling is vital to navigating the landscape while keeping humans in the loop at every key juncture.

```mermaid flowchart LR accTitle: Partnership Ecosystem Between Big Spaceship and Early Warning Services subgraph Agency["Creative Agency"] BS["Big Spaceship<br/>(Taryn Crouthers, CEO)"] end subgraph Client["Financial Network Operator"] EWS["Early Warning Services<br/>(Andrea Gilman, CMO)"] Zelle["Zelle Network"] Paze["Paze Checkout Solution"] end BS -- "Agency of Record / Integrated Campaigns" --> Paze EWS --- Zelle EWS --- Paze ```

Andrea Gilman introduces Early Warning Services, widely recognized as the network operator behind Zelle. The company delivers financial innovation to over 2,500 banks, community banks, and credit unions across the United States. Early Warning recently launched Paze, designed to offer streamlined online checkout directly backed by consumers' existing banks and credit unions. Gilman highlights their latest high-profile national campaign developed alongside Big Spaceship, starring Elizabeth Banks and Gabrielle Union—a deliberate creative "memory hook" tying the talent's names to "banks and credit unions."

- Big Spaceship has 25 years of experience in digital innovation and brand storytelling. - Early Warning Services operates the Zelle network and powers financial technology for more than 2,500 U.S. financial institutions. - Paze provides online bank-backed checkout; its national campaign leverages Elizabeth Banks and Gabrielle Union to reinforce the banking connection.

---

[03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) Employee Anxiety vs. Creative Leverage: 30 Seconds and 30 Years

Gilman reflects on the emotional reality facing employees in the generative AI era. Staff often experience excitement mixed with intense overwhelm, walking into offices feeling like an unresolved "squiggly line" and worrying about job security, tool selection, skill development, and adaptation. Crouthers notes that many clients initially approach AI with a narrow focus on efficiency—asking how to produce campaigns faster or cheaper. Crouthers dismisses that view as "last year's questions," insisting that the real opportunity lies in using the saved time to think strategically, solve business challenges from new angles, and expand creative ambition.

To illustrate the relationship between generative speed and human expertise, Crouthers shares an anecdote about an agency creative director who generated an extraordinary prompt output. When asked how long it took to create the piece, the director replied: "Well, 30 seconds and 30 years." Crouthers underscores that while the tool executed the generation in 30 seconds, it required 30 years of accumulated craft—in photography, art direction, framing, and visual taste—to know exactly how to guide and prompt the system. True innovation sits at this exact juxtaposition between deep human capital and AI capability.

- Employees face uncertainty around career paths and tool adoption amid rapid AI integration. - Shifting the conversation away from pure cost/time-cutting allows teams to focus on deeper strategic problem-solving. - AI execution time (30 seconds) relies entirely on decades of human artistic and domain experience (30 years).

---

[06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) Deconstructing the Creative Process: Concept, Socialization, and Personalization

Gilman outlines a stylized, three-phase model of creative development, detailing how human leadership and AI capabilities interact across each stage:

```mermaid flowchart TD accTitle: Three Phases of AI-Augmented Creative Development subgraph Phase1["1. Brainstorming & Concepting"] P1["Human-Led Ideation<br/>Emotional resonance, room-reading & humor"] end subgraph Phase2["2. Development & Socialization"] P2A["Asset Systems<br/>Replacing fixed asset counts"] P2B["Stakeholder Buy-In<br/>Photorealistic visual mocks for leadership"] end subgraph Phase3["3. Customization & Personalization"] P3A["Automated Permutations<br/>Channel formats, asset sizing, localized variants"] P3B["Final Human Approval<br/>Sign-off on brand alignment and tone"] end Phase1 --> Phase2 P2A --> Phase3 P2B --> Phase3 ```

1. **Brainstorming and Concept Creation:** This phase remains strictly human-centered. Gilman insists on keeping people together in a room to debate concepts, test humor, and gauge genuine emotional resonance. Because "AI cannot read the room," humans are irreplaceable in determining whether a cultural joke lands or an emotional hook feels authentic. 2. **Development and Socialization:** Here, AI integration accelerates dramatically. Rather than planning production around fixed deliverables (e.g., three TV spots and ten digital banners), teams now develop dynamic "asset systems." Concurrently, AI imagery aids internal socialization, allowing marketing leaders to show executive suites and partner banks precisely what a campaign will look and feel like before committing major production budgets. 3. **Customization and Personalization:** Generative tools enable massive variation across channels, aspect ratios, and targeted demographics—such as automatically swapping out burger bags for pizza boxes to match localized contexts. However, human oversight serves as the essential gatekeeper: people make the final approval call on every output.

- Brainstorming relies on human intuition because AI lacks empathy and situational room-reading ability. - Production planning has transitioned from counting individual assets to designing cohesive, scalable asset systems. - Personalization tools allow rapid asset swapping (e.g., localized packaging), but humans retain unilateral sign-off authority.

---

[08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) Practical Applications: Dynamic Asset Libraries and High-Fidelity Pre-Visualization

Crouthers shares concrete examples of how Big Spaceship incorporates AI into creative execution, emphasizing visual workflows:

| Creative Workflow | Traditional Approach | AI-Augmented Approach | Strategic Benefit | |:--- |:--- |:--- |:--- | | **Asset Library Generation** | Full tabletop shoots requiring every scenario and prop to be planned months in advance. | High-end base shoot combined with low-fi DSLR capture of modular props (e.g., glasses, calculators) styled via generative tools. | Allows real-time cultural adjustments without incremental studio shoot costs. | | **Storyboarding & Pre-Visualization** | Hand-drawn black-and-white sketches with mood board fabric/wardrobe swatches pasted in. | Detailed, photorealistic generative imagery matching the intended lighting, lens ratio, and wardrobe. | Eliminates client guesswork; secures executive and partner bank alignment before filming. |

In tabletop production, Big Spaceship previously had to storyboard and shoot every possible asset configuration during expensive physical production days. For Paze, they conducted a core tabletop shoot with a director, stylist, and photographer, capturing five baseline setups. When the creative team later realized they needed additional elements—like spectacles or a calculator—they captured low-fi photography with a DSLR and fed those items into the established style guide using generative tools. This shift empowers agencies to react to real-time cultural moments without funding incremental physical shoots.

Similarly, traditional advertising relied on black-and-white storyboard sketches with separate swatch inserts for wardrobe and props, forcing clients to imagine the final result. Using generative imagery, Big Spaceship now creates high-fidelity concept previews. Gilman notes that for Early Warning, this capability is invaluable: client-side CMOs must build consensus among executive teams and convince external stakeholder banks and credit unions to back large investments. Presenting rich visual representations upfront dramatically reduces organizational friction.

- Practical base photography can be combined with generative pipelines to produce endless on-brand variants. - Modular asset creation reduces the need for expensive, repetitive commercial reshoots. - Photorealistic pre-visualization improves stakeholder alignment across distributed partner networks like banks and credit unions.

---

[13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) Linear Production vs. System Thinking, and Managing the Review Burden

Crouthers explains the structural transformation from traditional, linear production pipelines to non-linear systems. The legacy model followed a rigid sequence: brief, creative concept, pre-production meeting, physical production, and final editing into a finite pool of perhaps 50 to 100 assets. Today, creative teams design open-ended narrative systems capable of generating an infinite variety of tailored creative assets. Crouthers dispels the myth that this is an effortless "push-button" trick, noting that human creatives actively guide every iteration to preserve storytelling nuance.

```mermaid flowchart TD accTitle: Linear Advertising Workflow vs System-Based Generative Workflow subgraph Legacy["Linear Traditional Workflow (Finite Output)"] direction LR L1["Client Brief"] --> L2["Creative Concept"] --> L3["Pre-Pro"] --> L4["Physical Shoot"] --> L5["Max ~100 Assets"] end subgraph Modern["Modern System-Based Workflow (Infinite Scale)"] direction LR M1["Strategic Concept"] --> M2["Core Design System"] --> M3["Generative Permutations"] --> M4["Human Verification Gate"] end ```

This explosion of creative scale creates new operational bottlenecks. Gilman points out that "no good deed goes unpunished"—when an agency produces hundreds or thousands of tailored variations, the client's marketing team still bears the legal and brand responsibility to inspect and approve every single asset. While automated brand-checking tools help flag layout, color, and guideline violations, both speakers agree that automated filters cannot replace human review. Campaign schedules must deliberately allocate time for real people to verify tone, consumer resonance, and messaging accuracy.

- Legacy ad production was linear, costly, and capped at roughly 100 assets per campaign. - Systemic generative design enables near-limitless variations across audiences and platforms. - Higher asset volumes exponentially increase the review burden, requiring automated brand checkers paired with mandatory human sign-off.

---

[15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) Redesigning Marketing Around Human Value, Deep Work, and Joy

Gilman turns the conversation toward work redesign, arguing that offloading repetitive, mundane tasks to AI should prompt leaders to re-evaluate how they allocate human time. Instead of simply absorbing more rote work, marketing organizations should intentionally direct saved hours toward "deep work, deep thinking, creativity, and more relationship building and socializing." Freeing teams from low-leverage administrative tasks creates space for the high-impact work that delivers both business outcomes and genuine professional joy. Crouthers agrees enthusiastically, observing that creative professionals "have lots of feelings and they do their best work when they're filled with joy."

In their closing remarks, Gilman and Crouthers challenge marketing leaders to clearly define and articulate their core principles regarding human oversight. Rather than becoming passive consumers of AI capabilities, organizations must intentionally decide where human judgment, creativity, and accountability belong at every step of their creative roadmap.

- AI should eliminate mundane administrative tasks to unlock space for deep, focused creative thinking. - Time savings must be reinvested in stakeholder relationship-building, cross-functional collaboration, and strategic craft. - Sustainable AI adoption requires teams to explicitly define their ethical and creative oversight principles.

## Notable quotes > [01:29](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=89s) explorers, not settlers — Taryn Crouthers

> [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) squiggly line — Andrea Gilman

> [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) last year's questions — Taryn Crouthers

> [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) Well, 30 seconds and 30 years. — Taryn Crouthers

> [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) AI cannot read the room — Andrea Gilman

> [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) no good deed goes unpunished — Andrea Gilman

> [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) deep work, deep thinking, creativity, and more relationship building and socializing — Andrea Gilman

> [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) have lots of feelings and they do their best work when they're filled with joy. — Taryn Crouthers

> the tools will keep advancing, but the most important work in marketing still depends on human judgment, creativity, accountability, and trust.

## Mentioned Ai4, Andrea Gilman, Big Spaceship, Early Warning Services, Elizabeth Banks, Gabrielle Union, Paze, Taryn Crouthers, Zelle

## Video description Taryn Crouthers - CEO, Big Spaceship Andrea Gilman - Chief Marketing Officer, Early Warning

Most conversations about AI in marketing start with capability: what it can do, how it can scale the work, and how it can expand capacity. This conversation starts with judgment: where AI sharpens the work, where it introduces risk, and where people have to lead.

Join Andrea Gilman, CMO of Early Warning, and Taryn Crouthers, CEO of Big Spaceship, for a conversation grounded in real decisions from their ongoing partnership. They’ll discuss how AI helped shape an ambitious national campaign from concept to production — giving teams a clearer view of creative ideas before they were filmed, organizing and flagging brand materials for human review and supporting brand consistency from strategy to execution.

The discussion will also look ahead at a broader shift in consumer discovery: consumers are increasingly moving from traditional search engines to AI-powered platforms to research, compare and make sense of information. For marketers, that raises new challenges. Throughout, one theme holds: the tools will keep advancing, but the most important work in marketing still depends on human judgment, creativity, accountability, and trust.

[Source](https://www.youtube.com/watch?v=HZFsWLG_2-E)

中文简报

Early Warning与Big Spaceship：以人工把关筑牢AI广告规模化基石 尽管生成式工具打破了实地拍摄限制，能产出无限的素材排列组合，但营销团队也面临着日益加重的审核负担，需要人工把关以维护品牌基调与情感共鸣。 Andrea Gilman 与 Taryn Crouthers 结合双方在推出 Paze 数字结账推广活动中的合作实践，概述了将 AI 融入创意广告的系统化方法。她们将创意流程细分为：纯人工主导的头脑风暴、AI 增强的素材系统与逼真的前期预览，以及仍需人工签署确认的可扩展个性化生成。讨论强调，30秒生成的成果依托于30年积累的人类工艺，同时提醒业界防范无限素材变体带来的巨大审核负担。最终她们提出，AI 节省下的时间应当有意识地投入到深度战略思考、人际关系构建以及创意乐趣之中。 营销人员必须明确界定并梳理整个创意流程中需要人工判断与监督的环节，以避免品牌调性脱节和语境偏差。 (0:04) 仅仅将 AI 视为提速降本的工具，会忽视其解放团队从全新视角化解复杂商业问题的真正战略价值。 (3:48) 在30秒内执行一条 AI 提示词，需要30年积累的人类技艺、审美与行业经验来对其进行有效引导。 (3:48) 头脑风暴与概念发想必须坚持由人工主导，因为生成式 AI 无法洞察现场氛围，也无法衡量真实的情感细微差别与幽默感。 (6:02) 营销活动的开发正在告别僵化固定的素材数量模式，转向针对多种规格和受众定制的动态、可扩展素材体系。 (6:02)

Reading summary

Andrea Gilman of Early Warning Services and Taryn Crouthers of Big Spaceship argue that while generative AI accelerates content production, maintaining brand integrity requires explicit, structured human oversight across every stage of the creative pipeline. Grounding Generative Speed in Human Experience (3:48) Crouthers notes that many organizations initially treat AI primarily as a mechanism to produce campaigns faster and cheaper, framing mere efficiency as 'last year's questions.' True leverage occurs when rapid technical execution is directed by deep professional background. Crouthers illustrates this dynamic through an agency creative director whose generative output took '30 seconds and 30 years'—thirty seconds for the tool to render the asset, enabled by thirty years of accumulated craft in art direction, lighting, and visual taste. Without that foundation, automated outputs risk brand inconsistencies, judgment lapses, and context omissions. Gilman observes that marketing personnel frequently experience anxiety regarding tool selection, shifting roles, and job security amid rapid technological adoption. Rather than using automation simply to accelerate output volume, leadership must guide teams toward strategic problem-solving, acknowledging that generative tools lack the empathy and room-reading capacity essential to nuanced creative work. Human Gates in the Creative Lifecycle (6:02) Gilman outlines a three-phase model mapping where human intervention and automation belong. Brainstorming and concept generation remain strictly human-led. Because AI systems cannot read a room or evaluate authentic humor, Gilman insists that creative teams must debate concepts face-to-face to ensure emotional and cultural resonance. By contrast, the middle phase of asset development and socialization heavily integrates AI. Teams replace static deliverable quotas with flexible asset systems and generate photorealistic visual mockups to win internal consensus. The final phase, customization, leverages generative tools to adapt formats, aspect ratios, and regional elements, such as swapping packaging items for local markets. However, both leaders stress that final sign-off authority on all variations must remain exclusively with human reviewers. Shifting from Linear Shoots to Scalable Systems (8:45) In traditional commercial production, agencies planned and captured finite configurations through rigid tabletop shoots. For Early Warning's Paze checkout solution—developed alongside a national campaign featuring Elizabeth Banks and Gabrielle Union to reinforce bank backing—Big Spaceship employed a hybrid approach. The agency captured baseline setups during a core physical shoot, then photographed supplementary props with a handheld DSLR and styled them via generative models against established brand guidelines. This workflow enables real-time adjustments for cultural moments without funding incremental studio days. Generative tools also alter pre-visualization workflows. Replacing traditional black-and-white sketch storyboards with photorealistic concept imagery eliminates guesswork for corporate stakeholders. Gilman highlights that for network operators serving more than 2,500 financial institutions, clear visual proof of concept significantly reduces friction when securing alignment across executive teams and partner bank leadership. Navigating the Operational Burden of Infinite Output (13:09) The transition from legacy linear workflows—which historically yielded 50 to 100 finished assets—to open-ended generative systems unlocks near-infinite permutations across target demographics and platforms. However, Crouthers refutes the idea that this process is an effortless push-button shortcut, emphasizing that human guidance remains necessary to maintain narrative tone. Gilman points out that this surge in creative assets creates an operational bottleneck for brand teams: every additional variant carries legal and brand liability. While automated compliance checkers can flag basic layout and color guide violations, organizations must allocate sufficient human review time before publication, as automated filters cannot evaluate messaging tone or consumer resonance. Reinvesting Productivity Gains into Deep Thinking (15:56) Offloading repetitive operational tasks should prompt marketing organizations to intentionally redesign work patterns. Gilman argues that time saved by AI must not be absorbed by more rote assignments, but reinvested into deep strategic thinking, cross-functional collaboration, and relationship building. Crouthers reinforces that creative teams perform at their best when unburdened by administrative churn, urging marketing leaders to establish clear operating principles for where human intuition and judgment reside. Both speakers conclude that sustainable AI adoption requires marketing executives to establish deliberate operational guardrails rather than adopting generative platforms purely for speed. While automation resolves production bottlenecks, the responsibility for brand trust, emotional nuance, and legal accountability remains strictly human.

精炼阅读摘要

Early Warning Services首席营销官Andrea Gilman与Big Spaceship首席执行官Taryn Crouthers指出，生成式人工智能虽能极大提高营销内容的产出速度与规模，但企业必须在全流程中明确设立人类判断与监管机制，以防范品牌偏差与语境失误。 超越效率幻觉与人类经验积累 (3:48) 许多企业初期接触人工智能时，往往狭隘地聚焦于如何更快、更廉价地完成营销活动。Crouthers将这种单纯追求降本提效的视角形容为过时的思维，主张真正的机遇在于将节省下来的时间转化为更深层的战略思考与创意野心。面对员工在AI转型期对职业前景与技能迭代产生的普遍焦虑，管理层需要明确AI产出与人类专业知识之间的共生关系。 针对生成式工具的极速响应，Crouthers分享了一位创意总监在30秒内生成高水准提示词成果的案例。当被问及耗时多久时，该创意总监回答道“30秒加30年”。这表明AI在30秒内执行的自动化产出，完全依托于人类在摄影、艺术指导与构图审美上历经30年积累的专业眼光。没有深厚的人类经验引导，技术工具无法独立完成高质量的创意突破。 创意三阶段中的人机协作边界 (6:02) Gilman将创意开发划分为三个阶段，明确界定了人类判断与AI工具的边界。在第一阶段“脑暴与概念孵化”中，团队完全由人类主导，因为AI无法理解复杂的社交语境与氛围，唯有人类面对面的推敲才能捕捉真实的情感共鸣与幽默感。第二阶段“开发与内部沟通”则高度依托AI加速，团队由规划固定交付物转向构建动态资产系统，并借助高保真视觉模拟图，使管理层与合作银行在正式拍摄前直观预览成片效果。 进入第三阶段“定制与个性化分发”，AI工具能够快速生成适应不同渠道规格及本地化场景的衍生内容，例如根据地域差异自动将汉堡包装替换为披萨盒。但Gilman强调，人类在这个阶段扮演着不可逾越的最终把关人角色，任何资产上线前都必须由人工审核其是否符合品牌调性与沟通语境，防止自动化带来的逻辑漏洞与知识偏差。 动态资产库与高保真预演流程 (8:45) 在传统制作流程中，平面桌面拍摄需要提前数月规划每个道具与场景，并在昂贵的实景影棚中逐一拍摄；而故事板则长期依赖黑白线稿和织物样本拼贴，迫使客户凭空想象最终效果。针对为早期预警服务旗下数字钱包Paze打造的推广项目，Big Spaceship展示了AI增强的工作流。团队在完成核心机位的拍摄后，仅需使用单反相机捕捉眼镜或计算器等零散道具，即可借助生成工具将其融入既有的品牌风格指南。 这种模块化的资产构建方式使代理商无需追加实体影棚预算，即可实时跟进文化热点并进行微调。同时，高保真的概念预览彻底改变了创意的内部沟通流程。Early Warning作为连接全美2500多家银行与信用社的网络运营商，其首席营销官必须在重大投资前说服庞大的利益相关方。高度还原的预演图像消除了决策层的猜测成本，显著降低了跨组织推行新创意的摩擦力。 系统化生产与审批瓶颈管理 (13:09) 广告制作正在从线性生产模式转变为开放式的系统化生产。传统的线性流水线自客户简报出发，历经概念确立、前期筹备、实地拍摄与后期剪辑，最终产出的物料规模通常被限制在50到100个资产以内。现代模式则围绕核心设计系统展开，能够衍生出海量针对不同受众与平台的定制物料。Crouthers澄清，这种规模化产出并非一键式的自动化把戏，整个生成过程始终需要创意人员的积极介入以维系叙事细节。 然而，资产数量的几何级数增长也带来了严峻的审批瓶颈。Gilman指出，当代理商交付成百上千个衍生版本时，客户团队依然承担着全部的合规与品牌责任。尽管自动化品牌核查工具能快速筛查版面、色彩和规范性违规，但机器无法代替人类审查语气偏差与情感共鸣。营销团队必须在项目排期中专门预留人工审查窗口，确保每一份对外物料都经过真实人员的严格复核。 Gilman与Crouthers总结强调，企业引入AI不仅是技术升级，更是工作方式的重塑。通过将重复性的机械劳动移交技术系统，营销团队应将释放出的精力重新配置给深度思考、艺术创作与跨部门协作。企业领导者若要长期维持品牌完整性，必须主动厘清并确立组织在各业务节点上保留人类判断的核心原则。

Detailed notes — zh

## Background 该内容由 Ai4 于 2026 年 8 月 20 日发布在 YouTube 平台。页面简介显示，Early Warning 首席营销官 Andrea Gilman 与创意机构 Big Spaceship 首席执行官 Taryn Crouthers 展开对话，探讨在 AI 浪潮下营销工作如何从关注技术能力转向聚焦人类判断与监督，并结合双方合作的全国性营销案例分享实战经验与前瞻思考。 ## TL;DR 本场对谈深入剖析了生成式 AI 在现代营销全流程中的合理定位与边界，强调人类判断力、监督力和审美经验始终是不可替代的核心支柱。两位嘉宾结合数字钱包 Paze 的实际操盘经验，详细阐述了 AI 如何赋能动态设计系统、高保真分镜宣讲以及大规模资产定制，同时指出“30秒 Prompt 背后凝聚着30年行业经验”。对谈强调 AI 绝非一键生成的万灵药，面对海量内容变体，团队必须保留严格的人工审查机制，并将技术解放出的时间重新投入到深度战略思考、高阶创意与人际连接等高价值工作中。 ## Atmosphere & style 整场对话务实、从容且充满行业前沿的敏锐洞察，品牌方与代理商掌舵人之间的互动默契自如、张弛有度。现场交流没有浮于表面对 AI 工具的盲目追捧，而是直面从业者的技术焦虑，以真实业务决策为锚点抽丝剥茧。整体节奏沉稳且富有启发性，传递出一种在算法时代坚守人类创造力本质与工匠精神的从容与底气。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=4s) 在 AI 能够规模化生成文案与视觉素材的时代，营销团队最关键的任务是明确界定人类判断与审查介入的具体节点，以防范逻辑漏洞与品牌调性偏差。 - [01:29](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=89s) 创意代理商与品牌在数字钱包 Paze 等实际战役中践行“人机协同（Human-in-the-loop）”准则，将 AI 定位为赋能工具而非完全自治的替代品。 - [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) 行业对 AI 的关注焦点已从早期的“如何做得更快更便宜”，升级为“如何利用释放的时间进行高维战略思考与解决复杂商业难题”。 - [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) “30秒与30年”形象地阐释了人机协作的本质：AI 生成结果只需 30 秒，但决定 Prompt 深度与产出质量的是背后 30 年的行业积淀与审美把控。 - [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) 头脑风暴与概念构思阶段必须以人为核心，因为 AI 无法感知现场氛围（Read the room），也难以精准拿捏复杂的人类情绪与幽默感。 - [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) 在方案开发环节，AI 推动营销团队从单一物料生产转变为构建灵活的“资产系统”，大幅拓展了物料延展的可能性。 - [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) 在定制化与个性化阶段，尽管 AI 能规模化产出针对不同受众与场景的多尺寸变体，但最终的发布与审批权仍必须由人类把关。 - [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) 通过基础实物采集结合 AI 品牌风格规范系统，品牌能够摆脱昂贵的全年预设实拍，实现对突发文化热点的低成本、即时敏捷响应。 - [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) 利用 AI 替代传统黑白线稿分镜渲染高保真效果图，消除了高层管理层与合作方的脑补门槛，极大缩短了方案沟通与投资获批周期。 - [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) 营销界存在“一键生成完美广告”的误区，AI 虽能打破生产上限，但全流程中没有任何一个环节可以脱离人类的深度参与。 - [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) AI 带来的物料指数级爆发加重了审查责任，营销流程中必须为人类团队预留充足的时间，逐一核验所有变体是否符合品牌标准与受众共鸣。 - [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) 应用 AI 的深层价值在于重塑工作意义，将人类从机械事务中解放出来，专注于深度战略工作、高阶创意探索与人际信任关系的构建。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=4s) 开场核心反思：AI 时代的营销创作与人类判断 Early Warning 首席营销官 Andrea Gilman 与 Big Spaceship 首席执行官 Taryn Crouthers 共同登台，在午餐前的活跃氛围中向与会者抛出了两个直击本质的核心问题：虽然 AI 能够产出令人惊叹的标语、精美的创意素材和高质量的文案，但营销人是否曾停下来思考，**人类的判断与监督究竟发生在流程的哪个环节？** 团队内部是否清楚界定人类判断与审查应当切入的具体节点？ AI 在生成内容时难免会出现判断失误、逻辑漏洞、信息遗漏，甚至与品牌调性相违背。因此，在整个创意生成过程中建立清晰的人类监督机制至关重要。 ```mermaid flowchart TD accTitle: AI 与人类判断在创意流程中的关系 subgraph AI_Gen["AI 自动化生成能力"] A1["生成文案与标语"] A2["生成视觉与创意素材"] A3["规模化变体输出"] end subgraph Human_Oversight["人类核心把控与判断"] H1["识别逻辑漏洞与遗漏"] H2["校准品牌调性 (Brand Alignment)"] H3["最终决策与责任承担 (Approval)"] end AI_Gen -->|"输出草稿与变体"| Human_Oversight Human_Oversight -->|"终审定稿"| Output["最终商业营销产出"] ``` --- ### [01:29](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=89s) 嘉宾背景：Big Spaceship 与 Early Warning 的合作历程 Taryn Crouthers 介绍了创意代理商 Big Spaceship。这家拥有约 25 年历史的机构深植于数字创新，长期为合作伙伴打造整合营销战役，并担任 Early Warning 旗下数字钱包产品 Paze 的记录代理商（Agency of Record）。Taryn 分享道，团队自视为“探索者而非定居者（explorers, not settlers）”，始终保持对事物的好奇心，并在每个环节坚守“人机协同（Human-in-the-loop）”的准则。 Andrea Gilman 随后介绍了 Early Warning Services 的业务背景。作为 Zelle 的网络运营商，Early Warning 为美国超过 2,500 家银行、社区银行和信用社提供金融创新服务。近期，双方联合推出了数字钱包 Paze 的最新全国营销战役，由 Elizabeth Banks 和 Gabrielle Union 联袂主演——巧妙利用两位演员的名字构成了“Banks and Union（银行与信用社）”的记忆记忆点（Memory Hook）。 --- ### [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) 职场焦虑与提问的演进：“30 秒与 30 年” 在 AI 浪潮下，许多员工陷入了兴奋与迷茫交织的状态，心中充满疑虑：我的工作会被取代吗？我该如何学习并适应新工具？ 针对客户对 AI 的诉求变化，Taryn Crouthers 敏锐地指出： * **去年的过时问题**：“如何更快、更便宜地制作营销活动？” * **当下的核心问题**：“如何利用省下来的时间进行更高维度的战略思考？如何用 AI 解决更复杂的商业挑战并释放更大的创意？” Taryn 分享了一个生动的行业轶事：一位创意总监最近写出了极为出色的 Prompt，生成的画面质量极其惊人。有人问他：“你写这个 Prompt 花了多长时间？”那位创意总监回答道：**“30 秒，加上 30 年（30 seconds and 30 years）。”** > “30 秒是生成那个结果的时间，但背后是 30 年在摄影、创意指导、艺术指导领域的深厚积累——知道该 Prompt 什么、如何精准 Prompt。” 这一回答深刻揭示了当下创意行业的人机关系：AI 的高效产出并非凭空而来，而是由深厚的人类经验与智慧在底层进行精准引导。 --- ### [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) 创意工作流的三阶段重塑 Andrea Gilman 梳理了现代营销创意开发的三个核心阶段，剖析了 AI 与人类判断在各环节的分工： ```mermaid flowchart LR accTitle: 创意流程三阶段演进 subgraph Stage1["阶段一：头脑风暴与概念构思"] S1["以人为核心<br/>面对面探讨情绪与幽默感<br/>(AI 无法读取现场氛围)"] end subgraph Stage2["阶段二：方案开发与组织宣讲"] S2["从单一资产转向资产系统<br/>AI 生成高保真分镜/效果图<br/>向高层与合作方直观演示"] end subgraph Stage3["阶段三：定制化与个性化"] S3["多渠道/多尺寸变体生成<br/>受众场景替换 (如汉堡换披萨)<br/>人类把关做最终定夺"] end Stage1 --> Stage2 Stage2 --> Stage3 ``` 1. **头脑风暴与概念构思（Brainstorming and Concepting）**：高度以人为中心。团队需要围坐在一起碰撞火花，感受幽默是否到位、情感共鸣是否准确。正如 Andrea 所强调的，“AI 无法感知现场氛围（AI cannot read the room）”，在这个最原始的创发起点，人类的情感洞察不可替代。 2. **方案开发与组织宣讲（Development and Socialization）**：AI 大幅介入的阶段。团队不再只是预先规划固定数量的 TVC 或数字物料，而是构建灵活的“资产系统（Asset Systems）”。同时，利用 AI 快速可视化概念，能够帮助营销人员在拍摄前将构想生动地展示给企业高管及合作银行，降低沟通成本。 3. **定制化与个性化（Customization and Personalization）**：AI 规模化生成多渠道、多尺寸的物料变体，并根据不同受众进行场景微调（例如将画面中的汉堡包装袋替换为披萨盒）。但最终的审批权和决策权依然牢牢掌握在人类手中。 --- ### [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) 落地实践：资产库构建、高保真分镜与宣讲革命 Taryn 结合 Paze 营销战役的具体案例，展示了 AI 工具在 agency 实际工作中的三种落地方式： #### 1. 动态资产库（Asset Libraries）与设计系统 * **传统模式**：为应对全年的传播需求，必须预先搭设桌面静物台（Tabletop Shoot），配备导演、摄影师、造型师，在前期将一年内可能用到的所有场景与组合全部拍完。 * **AI 增强模式**：团队只需用单反相机拍摄低保真、基础的实物素材（如团队成员实拍的眼镜、计算器等道具），随后输入 AI 系统并结合品牌规范，即可按需生成匹配风格的全新素材。这种方式将思维模式从“预设固定物料”转变为“构建设计系统”，使品牌能够在文化热点发生时即时响应，无需追加实体拍摄预算。 ```mermaid flowchart TD accTitle: 传统拍摄模式与 AI 增强资产系统对比 subgraph Traditional["传统线性拍摄模式"] T1["高成本预设布景与拍摄"] --> T2["产出有限的固定资产 (~100个)"] T2 --> T3["后续调整需追加预算重新拍摄"] end subgraph AISystem["AI 增强资产系统"] A1["基础低保真摄影采集 (DSLR)"] --> A2["输入 AI 风格规范系统"] A2 --> A3["无限量实时生成与微调"] A3 --> A4["根据文化热点即时敏捷响应"] end ``` #### 2. 分镜头脚本（Storyboards）与概念宣讲 * **传统痛点**：传统广告拍摄前使用黑白线稿分镜，造型细节（如帽子款式、服装质感）通常只能用零散的参考图附在旁边，极度依赖客户和利益相关者在脑海中脑补最终成片效果。 * **AI 变革**：利用生成式 AI 工具，代理商在拍摄前就能直接渲染出高度逼真、包含精确比例和光影质感的概念图。 Andrea 对此深有共鸣：在客户侧，营销负责人的核心职责是“带着领导层与合作伙伴一起走（bring leadership along）”。利用 AI 生成的逼真视觉方案，能让 C-Suite 高管以及合作银行的产品与市场团队在开拍前就清晰看到成片效果与品牌契合度，极大提升了方案推进和投资获批的效率。 --- ### [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) 规模化叙事与审查责任：没有一键生成的奇迹 Taryn 指出，传统广告制作是严格线性的：从 Brief 到创意开发、前置沟通（Pre-pro）、实地拍摄、后期剪辑，受限于预算和时间，最多只能精细化产出约 100 套物料。 引入 AI 系统后，叙事视界被彻底打开，团队理论上可以输出无限量的内容变体。但 Taryn 强调，这绝非“按一个按钮就能搞定一切”，市场上充斥着“一键生成完美物料”的误解。在实际操作中，团队的每个环节都必须有人类紧密参与。 随着资产产出量指数级爆发，“没有一桩善举不受惩罚（No good deed goes unpunished）”： * 尽管团队开始引入 AI 品牌校验工具（Brand Checkers）自动排查基础合规问题，但面对成百上千的变体，**人类团队必须在流程中留出充足的审查与决策时间**。 * 每一个面向消费者的资产，在正式发布前都必须经过人类的双眼进行最终把关，确保其符合品牌标准并能引发真实的受众共鸣。 --- ### [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) 重塑工作价值：将时间投入到带来深度与快乐的事情上 在对谈尾声，Andrea Gilman 提出了一个更深层次的组织议题：**当 AI 帮我们消除了琐碎、机械的日常事务后，我们该如何围绕“人类价值”重新设计工作？** 省下来的时间应当投向以下领域： * **深度思考与深度工作（Deep Work & Strategic Thinking）**：专注探索深层的商业逻辑与前瞻战略。 * **高阶创意探索（Advanced Creativity）**：打磨真正具有穿透力的创意概念。 * **人际连接与关系构建（Relationship Building & Socializing）**：在团队、跨部门及合作伙伴之间建立更深厚的信任。 Andrea 总结道，拥抱 AI 的终极哲学在于解放人类双手，去做那些能带来更大业务影响力、同时在生活中带来更多快乐的工作。Taryn 对此表示强烈赞同：创意人员往往情感充沛，当他们内心充满愉悦时，才能创作出最出色的作品。 两人最后呼吁所有营销从业者：在为 AI 的能力感到兴奋的同时，必须时刻保持清醒，清晰界定并向团队和合作伙伴阐明——**人类的判断力、监督力和创造力究竟应该在何处发挥不可替代的核心作用。** ## Notable quotes > [00:04](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=4s) 虽然 AI 能够产出令人惊叹的标语、精美的创意素材和高质量的文案，但营销人是否曾停下来思考，人类的判断与监督究竟发生在流程的哪个环节？ — Andrea Gilman & Taryn Crouthers > [01:29](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=89s) 探索者而非定居者（explorers, not settlers） — Taryn Crouthers > [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) 如何利用省下来的时间进行更高维度的战略思考？如何用 AI 解决更复杂的商业挑战并释放更大的创意？ — Taryn Crouthers > [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) 30 秒，加上 30 年（30 seconds and 30 years）。 — 创意总监（引述） > [03:48](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=228s) 30 秒是生成那个结果的时间，但背后是 30 年在摄影、创意指导、艺术指导领域的深厚积累——知道该 Prompt 什么、如何精准 Prompt。 — Taryn Crouthers > [06:02](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=362s) AI 无法感知现场氛围（AI cannot read the room） — Andrea Gilman > [08:45](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=525s) 带着领导层与合作伙伴一起走（bring leadership along） — Andrea Gilman > [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) 按一个按钮就能搞定一切 — Taryn Crouthers > [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) 没有一桩善举不受惩罚（No good deed goes unpunished） — Taryn Crouthers > [13:09](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=789s) 人类团队必须在流程中留出充足的审查与决策时间 — Taryn Crouthers > [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) 当 AI 帮我们消除了琐碎、机械的日常事务后，我们该如何围绕“人类价值”重新设计工作？ — Andrea Gilman > [15:56](https://www.youtube.com/watch?v=HZFsWLG_2-E&t=956s) 人类的判断力、监督力和创造力究竟应该在何处发挥不可替代的核心作用。 — Andrea Gilman & Taryn Crouthers ## Mentioned Big Spaceship, Early Warning Services, Taryn Crouthers, Andrea Gilman, Paze, Zelle, Elizabeth Banks, Gabrielle Union, Ai4 ## Video description Taryn Crouthers - CEO, Big Spaceship Andrea Gilman - Chief Marketing Officer, Early Warning Most conversations about AI in marketing start with capability: what it can do, how it can scale the work, and how it can expand capacity. This conversation starts with judgment: where AI sharpens the work, where it introduces risk, and where people have to lead. Join Andrea Gilman, CMO of Early Warning, and Taryn Crouthers, CEO of Big Spaceship, for a conversation grounded in real decisions from their ongoing partnership. They’ll discuss how AI helped shape an ambitious national campaign from concept to production — giving teams a clearer view of creative ideas before they were filmed, organizing and flagging brand materials for human review and supporting brand consistency from strategy to execution. The discussion will also look ahead at a broader shift in consumer discovery: consumers are increasingly moving from traditional search engines to AI-powered platforms to research, compare and make sense of information. For marketers, that raises new challenges. Throughout, one theme holds: the tools will keep advancing, but the most important work in marketing still depends on human judgment, creativity, accountability, and trust. [Source](https://www.youtube.com/watch?v=HZFsWLG_2-E)

Source quotes

[0:04] Andrea Gilman & Taryn Crouthers: 虽然 AI 能够产出令人惊叹的标语、精美的创意素材和高质量的文案，但营销人是否曾停下来思考，人类的判断与监督究竟发生在流程的哪个环节？ [1:29] Taryn Crouthers: 探索者而非定居者（explorers, not settlers） [3:48] Taryn Crouthers: 如何利用省下来的时间进行更高维度的战略思考？如何用 AI 解决更复杂的商业挑战并释放更大的创意？ [3:48] 创意总监（引述）: 30 秒，加上 30 年（30 seconds and 30 years）。 [3:48] Taryn Crouthers: 30 秒是生成那个结果的时间，但背后是 30 年在摄影、创意指导、艺术指导领域的深厚积累——知道该 Prompt 什么、如何精准 Prompt。 [6:02] Andrea Gilman: AI 无法感知现场氛围（AI cannot read the room） [8:45] Andrea Gilman: 带着领导层与合作伙伴一起走（bring leadership along） [13:09] Taryn Crouthers: 按一个按钮就能搞定一切 [13:09] Taryn Crouthers: 没有一桩善举不受惩罚（No good deed goes unpunished） [13:09] Taryn Crouthers: 人类团队必须在流程中留出充足的审查与决策时间 [15:56] Andrea Gilman: 当 AI 帮我们消除了琐碎、机械的日常事务后，我们该如何围绕“人类价值”重新设计工作？ [15:56] Andrea Gilman & Taryn Crouthers: 人类的判断力、监督力和创造力究竟应该在何处发挥不可替代的核心作用。

Source quotes

[1:29] Taryn Crouthers: explorers, not settlers [3:48] Andrea Gilman: squiggly line [3:48] Taryn Crouthers: last year's questions [3:48] Taryn Crouthers: Well, 30 seconds and 30 years. [6:02] Andrea Gilman: AI cannot read the room [13:09] Andrea Gilman: no good deed goes unpunished [15:56] Andrea Gilman: deep work, deep thinking, creativity, and more relationship building and socializing [15:56] Taryn Crouthers: have lots of feelings and they do their best work when they're filled with joy. the tools will keep advancing, but the most important work in marketing still depends on human judgment, creativity, accountability, and trust.

Original transcript

[0:04] Hello. Hello. How is everyone? [0:10] This is just before lunch energy. But we're going  to have some fun. We're going to have some fun. [0:16] Okay, we'll just switch mics. Oh, there we go. Hey  everybody, let's start by asking you to consider [0:22] two questions. The first one is this. We all know  that AI can turn in some really amazing marketing. [0:30] It can be a wonderful tagline, a beautiful piece  of creative, or some well-written content. But [0:38] have you stopped to ask yourself where does human  judgment and oversight, where did that happen in [0:44] the process? And the second question we'd like  to have you all think about is what about your [0:52] teams? Do they know where the human judgment and  the human oversight in the creative process needs [0:58] to come in? And the reason why these questions  are so important is that we know that AI can make [1:05] lapses in judgment. It can have gaps. It can omit  things. It can be inconsistent with your brand. [1:14] And so it's really important to think about as  we're going through the creative process, where [1:19] does human oversight, human judgment come in.  And that's what we're going to talk about today. [1:29] Okay, we're just going to shift. It's a good thing  we like each other. We're just going to shift. Uh, [1:34] so hello everyone. I am Taran Kthers. I am the  CEO of a creative agency called Big Spaceship. [1:42] Oh, we have a second. Thank you. Um, if you're not  familiar with us, we've been around for about 25 [1:48] years, always rooted in digital innovation, and  we create integrated campaigns for our partners, [1:54] and we are very fortunate to be the agency of  record for pays, part of early warning. Um, [2:00] and you know, it's interesting when you work  at a agency that's called Big Spaceship, [2:05] you can imagine we have a lot of spaceship puns,  right? Uh, because that's what creatives do. We [2:11] have fun. We like to think of ourselves as uh  a group of creative folks who are explorers, [2:18] not settlers. And I think there is no better  framing in what's happening today than having [2:24] that obsession with curiosity. So when we think  about how we work with partners, how we storytell, [2:32] it really is at the heart of everything that  we do that curiosity and um you know even some [2:39] of the questions Andrea that you just mentioned  around how you're framing all of this um it it's [2:45] incredibly important for us to make sure that we  keep humans in the loop in every single moment. [2:50] That's right. Um and by way of introduction,  hi everybody. I'm Andrea Gilman. I am the chief [2:55] marketing officer at Early Warning Services. Many  of you may know us as the network operator of Zel [3:02] uh and we provide financial innovation to over  2,500 banks, community banks, and credit unions [3:09] in the United States. We also have launched pays,  which is online checkout from your bank and credit [3:17] union. And together our teams with our agency Big  Spaceship have just launched our latest campaign [3:24] starring Elizabeth Banks and Gabriel Union. And  there is that fun little memory hook there of [3:31] banks and union being the product being from  your banks and credit unions um that we worked [3:36] on together. So, we're going to talk about some  examples of creative processes um that we worked [3:42] on together through both that campaign and the  first one we did for pays uh a couple years back. [3:48] But let's start with a little bit of a moment of  reflection about what is going on with employees [3:55] in the day and age of AI. First and foremost, um  it can be overwhelming and exciting all at once, [4:04] all at the same time. Both things can be true. And  a lot of employees sometimes come into the office [4:09] and feel a little bit like this squiggly line  and they're asking a lot of questions such as [4:15] what is going to happen to my job? How should I be  using AI? What tools should I take on? How should [4:22] I learn? Um and how do I need to adapt? And I  think Taran too, a lot of your clients are asking [4:27] similar types of questions. Yeah, absolutely.  There are so many partners and typically they come [4:33] and start with things like efficiency, right? So,  how can we make a campaign for a shorter period [4:38] of time or potentially cheaper? I personally  think those are last year's questions. Really, [4:45] where we're at now is how are we thinking about  using that incremental time to be more strategic, [4:50] to solve business problems differently, to use  AI, to be more creative. Um, and I'll use a really [4:57] great example that I personally love. So, one of  um uh our creative directors did this incredible [5:05] prompt recently and the output was really really  good and someone asked them how long did it take [5:12] you to do that prompt and their response was  well 30 seconds and 30 years. I love that. um [5:21] right and I'm I obviously it wasn't me but I am  obsessed with it because I think that is the heart [5:27] of what's happening right now in our industry  certainly creatively right so 30 seconds was [5:32] how long it took for that you know uh to create  that output but 30 years of you know photography [5:44] creative direction art direction knowing what to  prompt knowing how to prompt and we are at that [5:50] oxiposition truly between um how you have human  capital and intelligence that are informing these [5:56] systems to make the best outputs. Yeah, that's  right. So, let's really talk a little bit more [6:02] about the process. And so, what we have up here  on the slide is what we'll say is a very stylized [6:08] version. If only it was this simple. Yes. If  only it was this simple of the creative process, [6:13] right? Um and so we're talking about three  phases of creative development today. [6:18] Um the first one is brainstorming and creating  the concept. Uh this is one where it's very human [6:26] centered. I am such a fan of getting a bunch of  humans people in the room brainstorming throwing [6:32] ideas around and really trying to understand  what works, what's for us relatable and [6:38] humorous. Um is are people getting the joke?  Is the emotion landing? And we always say, [6:45] you know, AI cannot read the room. So really  having people in the room at this concepting [6:50] brainstorming process and really leaning heavily  into people is very very important. Um the second [6:57] phase that we want to talk about here a little bit  today is development and socialization. So this is [7:03] one of the areas where AI is starting to lean in  very heavily. So when you think about developing [7:09] uh a campaign uh just even a year or two ago we  used to think about creating x number of TV spots [7:17] y number of digital assets etc etc but now we're  starting to think much more in terms of asset [7:25] systems and Taran's going to walk you through some  examples of what that looks like. The other thing [7:31] that has started to become really interesting is  in the creative process is how you can use AI to [7:38] help socialize a concept. Um, and so for many  of us who are on the client side, bringing our [7:45] leadership along and really trying to get them to  understand the emotion in a campaign or what it's [7:52] going to look like that has also been aided very  much by AI. And then the third step of the process [8:00] that we're going to talk a little bit more about  today is customization and personalization. And we [8:05] all know how AI can really help us create several  permutations. Um, and I like to think about things [8:12] like different channels, different sizes, but also  make it more relevant for different audiences. [8:18] Swapping out burger bags for pizza boxes. So these  are the things that AI can help us run through the [8:25] different versions um throughout the process. But  the thing that remains and while there's humans [8:30] really leaning in in different ways in each of  the steps is that people are making the final [8:36] call and people are making the decision about  what gets approved and what does not. But Taran, [8:42] maybe it'd be great for you to walk through  some of these examples for the audience. [8:45] Absolutely. As soon as I saw you say pizza,  everyone was like pizza. Is it lunchtime? Um, [8:53] so there are literally hundreds of ways that  we have integrated AI into our workflows at the [8:59] creative agency. Here are three. I am uh nothing  if not a daughter of two artists. And so visual [9:07] examples are my love language. So the first is  storyboarding which we'll talk a little bit about. [9:12] Of course asset libraries and then how we think  about creative and strategic testing all along [9:17] the way. So let's start with asset library. So  this is an example from our current pays campaign [9:24] that I love. What you see on your left hand side  is a stylized actual practical photography shoot, [9:32] a tabletop shoot, right? So we had a director, a  photographer, a stylist. We set up this tabletop [9:40] shoot. We did five different executions of  it. And this was all part of a campaign. And [9:45] we think about assets and versioning. In the old  day, we would really have to pre-think of all of [9:52] the different scenarios that we would need for  an entire year-long campaign in advance. Now, [9:58] with the new AI tools and generative tools that  we're using, we can take a DSLR camera and go out [10:05] and have a relatively lowfi shoot. What you see  in the middle is actual photography of someone who [10:13] was on our creative team who realized we wanted  spectacles and we wanted um you know uh what is [10:20] the other thing in here? Um a calculator. And  so we then were able to input that in the style [10:25] guide in the exact way. And so what does this  mean for clients? It means that we don't have [10:30] to think about everything in advance. We can  think in systems and design systems instead of [10:35] in specific assets. It also means in real time as  things are happening culturally, we can add really [10:42] interesting nuance to the work without having to  do physical incremental production which of course [10:48] not only saves money but also saves time and gives  us better creative flexibility and output. So to [10:54] me this is a really great example of a win-winwin.  Uh another really fun example. So, what you see in [11:02] this black and white drawing is, if you're not in  advertising, what we call a story board. And so, [11:08] before you typically um execute a complicated,  you know, commercial shoot, which of course is [11:16] time and it can be expensive. Um, on the agency  side, we will create these story boards. And so, [11:21] you can think of them as stepping out different  images of how we're going to tell the story for [11:27] this 30-second, 60-second commercial, whatever it  is. And in the pre-pro or the production meeting [11:33] that we have with our partners, um there  historically used to be little inserts that [11:39] our production team would put in. So here's what  the hat will look like for wardrobe. Here's some [11:44] nuance of what people are going to wear. And we  would really need to lean on partners to have the [11:49] ability to put that all together in their mind to  know what they were about to pay for. Now instead [11:55] with generative imagery, we can show what you see  here on the right. So people really can understand [12:02] and I know this has fundamentally changed how  we work even internally as an agency. We're [12:07] able to mock stuff up in advance, think about the  different ratios, what's working, what isn't. But [12:13] I'm curious, Andrew, from your perspective,  even socializing inside of an organization, [12:18] how this has changed. Yeah. Right. I know it  definitely helped us understand working with you [12:23] in terms of what we think the campaign is going  to look like. Uh, and that is always very fun. But [12:28] when you think about the marketing on the client  side, our job is to bring leadership along. We've [12:34] got to convince seuite and other folks to invest  in the campaign. And then because we work across a [12:40] number of banks and community banks, often we're  also working to bring along the banks with us, [12:45] either their product people or their marketing  people. So being able to really bring the [12:50] campaign to life before we've even shot it  and show them how it's going to connect, [12:55] how it's going to be on brand and how it's going  to resonate um through the power of AI imagery [13:00] is super super powerful. Yeah, absolutely. I think  when we talk about workflows and how our workflows [13:09] have changed again if you look at the left hand  side everything historically for creative and [13:15] advertising from a production standpoint was quite  linear. So initially you would have a brief which [13:20] would come from you know a partner saying we want  to execute something solve a problem then we will [13:26] put together a creative development concept.  We will then do a pre-pro which we just showed [13:30] you. We will go out and do physical production.  That physical production then gets edited into a [13:36] certain number of assets and it's all very much a  finite and linear process and now we are able to [13:44] think in systems and as opposed to in assets and  it has just been incredibly exciting I think on [13:52] the creative side and the agency side because  it opens up the aperture of how we're able to [13:57] storytell. So if historically you had a finite  budget and a finite amount of time and you could [14:03] create let's say maybe at maximum a 100 different  assets and that was if you were being incredibly [14:09] thoughtful and strategic and make sure you had all  of your versioning. Now you can make an unlimited [14:15] number of assets because you can go in and change  workflows. you can add and build that narrative. [14:22] And so it really opens up the aperture of how  we're able to tell stories creatively and for [14:28] partners. Now, of course, this sounds incredible.  This is not a click of a button, right? I think [14:35] there's probably quite a bit of misinformation  about you can plug something in and all of a [14:41] sudden get an incredible asset and output. We  have our team in the loop every step of the way [14:47] throughout this process. But what the technology  has allowed for is for us to really create an [14:54] infinite number of assets and storytelling at  scale. And I think you know also um almost like no [15:02] no good deed goes unpunished. When we start to be  able to create all of these assets, it's up to my [15:07] team to approve every single one. And so we also  had to leave enough time into the whole process [15:13] and roadmap to make sure that people are checking  and making sure that all of the creative and all [15:19] of the assets are correct, that they're good,  that they're going to resonate with the consumer, [15:23] and that they are on brand. Yeah, absolutely. So  when we think about onbrand um I think one one [15:32] extra piece um that just you know to mention  there are so many ways historically you would [15:38] have to have a human who was checking everything  and now we have the ability um to run through and [15:45] look at brand checkers um but again at the end of  the day regardless of what those brand checkers [15:51] say again we we always have our team doing one  last one last look. Yeah. And so this starts to [15:56] actually open up a third area of question that I  think is very interesting for folks here in the [16:01] creative process to think about. Um because AI  does allow us to remove some of the mundane bits [16:09] of our work. It does allow for organization of  things that we have to check and and processes [16:15] we have to go through. And so there is another set  of questions that we could be asking ourselves in [16:20] our teams which is how do we think about really  redesigning the work that we do around human [16:26] value. So as you free up time from some of these  tasks that we did even just a year ago, what are [16:32] the things that you're switching to? And I think  it's really important to be asking yourselves [16:38] and your teams those questions. What are actually  the things that you're going to take on time for? [16:43] uh and what I do believe that we will be taking  more time for things are are things that require [16:49] deep work, deep thinking, um creativity, uh and  more relationship building and and socializing. [16:56] And so that is very powerful and I do think  that when we go through this type of exercise, [17:02] um we also end up finding that we're able to take  on work that brings more joy. So my philosophy is [17:08] really about really thinking about how can we  leverage AI to free ourselves up to do that [17:13] work that brings more impact and more joy uh in  our lives at the end of the day. Well, you know, [17:19] as someone who runs a team of creatives, they have  lots of feelings and they do their best work when [17:25] they're filled with joy. So, I'll double down on  that all day. Absolutely. Amazing. All right. So, [17:31] with that, maybe we'll just ask for final  thoughts. My final um couple words I want [17:36] to share with you is just really as we are very  excited about the world of AI and all that it can [17:42] bring to us. It's really important to pause and  to step back and say, where does human judgment, [17:49] where does human oversight and where does human  creativity really fit into this process and making [17:54] sure that you know how that works every step  of the way with your teams and that you can [18:00] articulate those principles to your teams and  to the partners that you work with. Yeah. And [18:05] uh I will I think now just have the pleasure  of saying I hope everyone has an incredible [18:09] rest of the conference and a good lunch. Right.  Wonderful. Thanks everybody for joining us today.