AI is having a moment. But for most marketers, that moment feels out of reach. Too often, AI tools feel disconnected from strategy, demanding more setup than they save. But pressure is mounting. Marketers are being asked to do more with less. And while AI can be a powerful force multiplier, it only delivers when built on the right foundation: clean data, clear definitions, governed systems, and most importantly, human-led oversight.
That’s exactly what we explore in Step 4 of Iterable’s Moments-Based Marketing Playbook. In this session, leaders from Iterable, Merkle, and Nextdoor unpack how to operationalize AI without overcomplicating it. Below, we break down what that looks like in practice—from foundational requirements to real-world applications.
| We’ll cover the top takeaways here, but be sure to check out our full conversation, Step 4: Put AI to Work with Human-Led Expertise. |
The AI Opportunity Is Real — But So Is the Execution Gap
Nearly every marketing leader has AI on their radar, and most already have it on their roadmap. But awareness isn’t the problem. The real challenge is activation: translating promise into performance, and integrating AI meaningfully into real workflows.
That gap between awareness and activation becomes clear when you look at how marketers describe the impact AI is actually having today.
- 65% of AI adopters say it gives them a competitive advantage…but that advantage is only realized when AI is embedded into core systems and decision-making, not when it lives off to the side as an experiment or a pilot project.
- 63% of digital marketing leaders admit they’re still struggling with personalization at scale…not because they lack ideas, but because they’re blocked by operational and technical limitations.
- 47% of marketers say improving efficiency is their top reason for using AI…yet many teams end up doing more manual work because their AI tools require extra configuration, oversight, or workaround processes.
What’s Slowing AI Adoption, According to Marketers
To ground the conversation in real-world experience, we polled attendees during the Step 4 webinar about their biggest barriers to making data AI-ready. The responses made one thing clear: the challenge isn’t belief in AI’s value—it’s the foundational friction that prevents teams from operationalizing it.
The top barriers marketers cited to being AI-ready were:
- Data quality (35%) — difficulty keeping data clean, consistent, and up to date
- Data silos (32%) — critical information trapped across disconnected systems
- Data access (29%) — reliance on engineering or workarounds to get usable data
Similarly, when asked about challenges with adopting agentic AI, the responses pointed to a related but more operational set of challenges:
- Lack of roadmap clarity (23%) — uncertainty about how to apply AI in a practical, scalable way
- System integration (19%) — AI tools that feel bolt-on rather than embedded into daily workflows
Together, these results reinforce a consistent theme: AI adoption doesn’t stall because marketers don’t see the value. It stalls when teams lack the foundational data, clear pathways, and integrated systems needed to turn AI from an idea into something they can actually use day to day.
Nextdoor’s Foundation-First Approach: Three Keys to AI That Actually Works
Many of the challenges surfaced in the attendee polls were familiar to Shweta Puri, Senior Product Manager of Marketing Technology and AI Operations at Nextdoor. During the session, Shweta shared how Nextdoor approached AI not by rushing into advanced automation, but by first doing the foundational work required to make AI usable at scale.
For Shweta and her team, it came down to three foundational pillars:
1. Clean Identity Resolution
It sounds simple, but achieving consistent customer identity across systems is one of the most overlooked (and underappreciated) steps in building an AI-ready infrastructure. By standardizing how customer records were matched across platforms, Nextdoor avoided one of the most common pitfalls in AI: duplicative or conflicting user profiles that made it difficult to distinguish one user from another. Clean identity resolution gave them confidence that AI outputs were rooted in reality, not guesswork.
2. Shared Definitions Across Teams
Nextdoor didn’t stop at clean data. They went further to create shared understanding across lifecycle marketing, product marketing, and revenue operations. Terms like “active user” and “high-value customer” were codified and aligned across functions. As a result, AI models were trained against definitions that actually reflected business goals. Marketers and machine learning teams were optimizing toward the same outcomes rather than working at cross-purposes.
3. Governed, Streamlined Data Ingestion
To reduce inconsistency and operational drag, Nextdoor moved away from numerous one-off lists and custom feeds. Using a single Databricks pipeline and Iterable’s Smart Ingest, they eliminated redundant pipelines and “backdoor” data paths, making AI inputs more consistent and reliable. With a strong data foundation, Nextdoor built a system that enabled AI to scale personalization and optimization with confidence.
Where AI Adds Value Today — And Where It’s Headed
In the age of generative buzz and predictive hype, it’s easy to assume that the most exciting AI innovations are the flashiest. But for marketers on the ground, the most impactful applications of AI quietly remove friction, eliminate guesswork, and accelerate decisions at scale.
What’s Working Now: AI That Streamlines the Everyday
Today’s highest-value use cases are designed to operate behind the scenes, embedded directly into lifecycle platforms like Iterable. These tools work with marketers, optimizing tasks that previously ate up hours of testing or manual tweaking, such as send-time optimization, channel selection, and content personalization.Â
What’s Next: AI That Powers Full-Funnel Orchestration
Looking ahead, AI is evolving beyond isolated optimizations toward journey-level orchestration. In this emerging model, marketers define the goal (purchase, conversion, retention, engagement), and AI analyzes real-time signals to determine the next best step for each individual. The system dynamically adapts the journey logic, pacing, and channel mix based on what will move that person closer to the outcome.
“Instead of asking what emails are we sending next, we’ll ask what outcome do we want for this person right now, and what’s the best path to get there?”
~ Shweta Puri, Sr. Product Manager of Marketing Technology and AI Ops at Nextdoor
How Iterable Embeds AI Into Everyday Work
AI’s transformational power comes from how and where it’s applied. That’s why Iterable has focused on integrating AI directly into marketers’ day-to-day tools and decision flows. These examples highlight four practical ways AI is being applied today, aligned to a human-led design philosophy.
1. AI Embedded Into Everyday Work
Ideas are easy. For most marketers, the biggest bottleneck is execution. Iterable’s MCP Server reduces that gap by letting marketers interact with Iterable using natural language through tools like Claude. Instead of jumping between systems, marketers can create or update templates, translate content, or ask performance questions using simple prompts, keeping momentum without giving up control.
2. AI Providing Intelligent Guidance
As programs scale, it gets harder to stay focused on what’s actually happening and what needs attention. Nova, Iterable’s in-platform intelligence engine, is designed to help marketers stay oriented without pulling them out of their workflow. By surfacing recent work, messaging health signals, and progress toward goals in one place, it reduces the need to hunt through dashboards or reports. Explainable guidance gives marketers clarity on why something matters while leaving decisions firmly in their hands.
3. AI Optimizing Campaign Execution
Some decisions matter, but don’t need to be remade by hand every time. Choosing when to send, which channel to use, or how often to message is critical—but hard to optimize manually at scale. Iterable’s AI-powered optimization features handle this tactical tuning within marketer-defined boundaries, improving timing, channel selection, and frequency while leaving strategy firmly in human hands.
4. AI Improving Audience Targeting
Not every user needs the same level of attention. Iterable applies AI to help marketers decide where to focus—and where not to over-invest—without turning targeting into a black box.
Brand Affinity enables strategic cohorting, like nudging neutral users toward loyalty or protecting brand equity by suppressing disengaged audiences. Marketers can see exactly which behaviors, campaigns, or moments caused affinity to shift, turning sentiment into a practical signal they can act on. Predictive Goals extend that thinking to outcomes, helping teams prioritize users who are most likely to convert and adjust effort accordingly.
Guardrails Make AI Scalable and Safe
As AI becomes more capable, the risk isn’t that it does too little—it’s that it does too much without direction. That was a clear warning throughout the session, emphasized by Shweta and Erin Kelsh, VP of Messaging Solutions and Innovation at Merkle.
They cautioned against treating AI as a black box or giving it too much autonomy. When teams mistake automation for strategy, AI can quickly introduce new problems instead of solving existing ones:
- Audiences become too small to matter, shaped by overly narrow or inconsistent signals
- Recommendations don’t fit cleanly into journeys, creating fragmented experiences
- Decisions drift from brand and business intent, optimizing for the wrong outcomes
AI doesn’t know what “good” looks like unless humans define it. That starts with understanding three critical layers of structure:
- Strategic Outcomes: Marketers should own the “why” — the goals and success metrics. AI should work toward those ends, not define them.
- Creative and Brand Guardrails: Messaging cadence, tone, voice, and relevance
- Data Governance: As Nextdoor proved, consistent inputs with shared definitions str non-negotiable.
When AI is built into the systems marketers already use, guardrails stop being abstract rules and start becoming part of how work actually gets done. That’s what makes AI scalable and safe.
Dos and Don’ts for AI in Lifecycle Marketing
| Do | Don’t |
| Keep humans in control of strategy and standards | Don’t give AI full autonomy or treat it like a black box |
| Invest in clean data, shared definitions, and governed ingestion | Don’t expect AI to fix poor data quality or siloed systems |
| Embed AI directly into tools and workflows | Don’t bolt on disconnected AI features |
| Use AI to prioritize audiences, timing, and channels | Don’t over-target low-likelihood segments |
| Leverage explainable AI to build trust | Don’t rely on outcomes you can’t interpret |
| Apply AI consistently to learn and optimize over time | Don’t expect overnight results without feedback loops |
From Complexity to Clarity: AI That Works With You
When applied intentionally, AI makes personalization possible at a scale no human team could manage alone.
But the real differentiator is whether your systems are designed to let AI amplify your best thinking — your strategy, your creativity, your judgment — rather than adding another layer of complexity to manage.
If you want to see what that looks like in practice, watch Step 4 on demand.
