Why We’re Building Nexie

B2C marketing is entering a new phase.

Customer acquisition is getting harder. Retention matters more. Customers are flooded with messages, and inboxes are crowded. Getting attention now requires more relevance, better timing, and more restraint.

It is clear what needs to change: brands must reach customers with deeper personalization on their core owned channels - email, SMS, and in-app.

The current operating model is simply not designed for these demands.

Most retention marketing still relies on humans making manual campaign decisions for every key element: audience, timing, channel, offer, messaging, testing, and learning.

Generative and agent-based AI enable a fundamentally new operating model.

Our view: the future of B2C retention marketing should be operated by systems, with marketers guiding strategy and setting boundaries.

That needs a new way of working, not a better version of the current one.

That is why we are building Nexie.

The Problem: A marketing model that has hit its ceiling

Retention marketing today has more data, infrastructure, channels, and AI than ever. Still, many brands struggle to improve repeat purchase, engagement, lifetime value, and churn.

This is not a gap in data or tools. The issue is structural.

The current model expects marketers to consistently make campaign decisions at the right level of detail, every single time.

Who gets the message? What is the right thing to say? Which angle does it use? Which product, offer, content, or experience does it highlight? When is the right moment to send it? Which channel is fit for purpose? What needs to be tested?

These decisions are not occasional. They happen constantly. No team can maintain this level of precision and consistency at all times.

As a result, teams simplify.

Most teams default to calendar-driven campaigns, broad segmentation, and templated journeys. Testing is infrequent and often reactive. Insights from reports are applied manually, well after the fact.

Most retention marketing is generic. Not because marketers want it that way, but because operational complexity forces them to keep it that way.

AI copy writing, smarter segment builders, and more dashboards do not solve this. They still assume a human can keep deciding what should happen next.

The model itself has to change.

Our Take

B2C retention marketing should be operated by agentic systems and governed by humans. This is the foundational shift Nexie is built to enable.

The bottleneck is not human creativity. It is human operation.

Humans should define direction, taste, brand, boundaries, and key judgment. Systems should run the cycle: identifying opportunities, generating content, executing, capturing feedback, and improving.

The technology is now ready. Generative AI can deliver consistent, on-brand communication at scale. Agentic systems are capable of making decisions, executing actions, and drawing meaningful lessons from real outcomes. With first-party data now sufficiently rich, these systems have the grounding required to operate with higher intelligence and relevance.

All of these capabilities are now available at once. Brands that integrate them early will move beyond the curve.

This is a new incarnation of marketing with no campaigns and no fixed customer journeys.

Campaigns should not be the default planning unit. Campaigns exist today because they make operational sense for marketers. Customers do not experience campaigns. They experience moments. A message has to be relevant to the moment, otherwise it will be ignored.

Fixed journeys should also go away because no customer fits neatly into pre-set categories with predetermined paths. Each customer’s context is unique and has to be understood and acted on in real time for communication to be relevant.

Learning cannot wait in a dashboard for someone to notice and act. In an agentic model, the system uses every data point to inform the next step.

The marketer’s role does not vanish in this model. Their focus shifts from hands-on execution to defining and guiding the system’s direction.

This shift is already underway, to some extent, with existing vendors. But it deserves to be built correctly, with fresh thinking from the ground up, rather than by bolting AI on top of legacy frameworks.

The Agentic Marketing Model

The shift begins with a fundamentally different question.

The traditional approach frames the question as: Which campaign should we launch this week?

The agentic model reframes the question: Where is the highest potential right now, and what is the right action to take?

That change alone restructures the marketing process at its core.

Retention moves from a human’s best-guess campaign to a probabilistic opportunity identified by systems. It moves from manual production to generated actions. It moves from sparse testing to continuous learning. It moves from the marketer as operator to the marketer as guide.

A serious agentic marketing system needs to excel at four things.

Agentic decisioning

In an agentic model, the system analyzes customer, behavioral, product, and business data. It identifies opportunities, forms micro-segments dynamically, and decides if an action is needed.

The system might identify customers likely to respond to free shipping instead of a discount, or suppress high-intent buyers from an offer they do not need. It could also flag a high-value cohort showing early signs of disengagement.

The decision of who, what, how, and when moves from a human’s best guess to a system’s calculated probability.

Agentic execution

Decisioning is only valuable if the system can act on it.

Today, even when marketers know the right action to take, execution is slow. They still need to write copy, create variations, design creative, define audiences, apply suppression rules, and QA every detail before launching.

Generative AI streamlines this process. Once the system identifies an opportunity, it can generate the necessary assets and configuration required to act on it. This includes messaging, product recommendations, audience, channels, timing, and suppression rules.

AI writing content is useful, but the real shift is AI turning an opportunity into a complete, ready-to-review action.

Self-optimization

The strongest advantage of an agentic system is the speed and granularity of learning.

An agentic system can run experiments continuously and in parallel across message tone, offers, products shown, channels used, timing, and frequency.

The results are analyzed at a micro-segment level and fed back into the next decision the system makes. That flywheel is what separates an agentic system from an AI-assisted tool.

Human oversight

Agentic marketing should not be a black box.

Marketers define the goals, brand voice, offer rules, frequency caps, approval flows, and constraints. The system operates within that framework.

They also choose the level of autonomy. Some actions can be suggested. Some can be drafted for approval. Some can be automated once trust is earned.

Every important decision should be explainable: Why this audience? Why this action? Why now? What past learning did it use?

Control does not disappear in an agentic model. It moves to the right altitude.

The Foundational Bets Behind Nexie

Decisioning, execution, learning, and human oversight are not optional. They are foundational for any genuine agentic marketing system.

Yet this alone does not define what Nexie is.

What sets Nexie apart is a set of focused, domain-specific convictions about how agentic marketing best serves B2C brands.

These are the three core bets we have made.

Built for no-campaign marketing

We are certain campaigns will not remain the foundation of retention marketing over time. Nexie is being built for that shift.

Campaigns are a byproduct of a marketing era where humans had to organize outreach, group customers, coordinate workflows, and manage communication schedules manually.

Customers do not experience campaigns. They experience moments. A message either helps in that moment or becomes noise.

If a system can track customer state, understand business context, and act dynamically, there is no reason for campaigns to exist.

This creates space for adaptive, intelligent marketing without relinquishing control.

Most AI-native solutions focus on making campaign production easier for marketers. Few challenge whether campaigns should remain at the center of retention marketing at all.

Nexie will support some version of batch-and-send where the market still needs it. But we see it as a stop-gap toward becoming a no-campaign marketing system.

Aware of the business, not just the customer

Customer and business context must be evaluated together. Nexie is designed with this approach at its core.

Most marketing systems build their foundation on customer insights, tracking behaviors and demographics. Yet they rarely account for the underlying drivers of the business itself: why it exists, what it offers, what makes it unique, and what the goals are for this quarter.

Historically, that was fine. Marketers carried that knowledge in their heads. For a system to operate marketing, it needs to internalize the business context.

This business knowledge should go deep. Which products carry margin, where inventory is sitting, how competitive a category is, and which metrics matter this month.

The system is able to make informed decisions such as:

This new category carries higher margin, but awareness is low. Promote it to customers who buy related products.

Or:

Winter inventory from December is still sitting. Clear it by biasing recommendations toward this.

This happens by going beyond system integrations with inventory, catalog, and supply chain. The business user becomes an active stakeholder in the process, reviewing and approving strategic goals and adding context where data alone may be inconclusive.

Adaptability and continuous learning at the core

Our position is clear: adaptability is not optional for agentic systems. It must be foundational to how they are designed and how they operate.

Every business brings a unique operating model and perspective. Systems must adapt to these realities instead of forcing businesses to give in to how systems operate.

This is even more critical in marketing systems. For B2C businesses, brand identity shapes the product experience itself, not just its appearance. An agentic marketing system must align with and reflect how the brand communicates, sells, makes offers, sets discounts, and builds relationship and trust with customers.

Learning and system evolution are fundamental and must be part of the foundation. They should be central to the system’s core architecture and not a layer on top.

Nexie is built with adaptability and learning as a corner stone. This adaption and learning happens in two ways. Policies and guardrails that marketer sets, and things it learns and concludes.

Every learning changes the behavior of the system.

For example, consider the strategy for driving a first purchase in a specific customer segment. Nexie can test different approaches and determine, through real results, that building brand trust outperforms simply offering 20% off. Once this learning is approved by the marketer, Nexie shifts its approach for that segment accordingly.

In agentic B2C marketing, every learning should become a state change that compounds over time, creating a marketing machine that adapts and optimizes to how a brand operates.

The Path Forward

If you extend this model to its logical conclusion, marketing starts to look very different.

Picture a marketer whose sole focus is one customer at a time. The system listens, responds only when it adds value, and delivers information that is genuinely beneficial to the individual.

That does not really sound like marketing anymore. It sounds like an advisor.

That is the destination.

Brands that make this shift early will not just see improved performance metrics. They will reshape their relationship with customers.

The ability to earn genuine attention is now the durable advantage. Building that relationship is what sets brands apart.

Customers are no longer overwhelmed by generic messaging. Instead, they receive communication that actually matters to them.

This is where the category has to go.