The End of 'Do Not Reply': Direct CRM Is Becoming a Conversation

For decades, direct CRM channels were built to broadcast. Marketing email came from an address no one could answer. SMS arrived in 160 characters with a link and little else. Brands talked. Customers clicked, or didn’t.

That model is ending. The two channels marketers rely on most, email and messaging, are becoming conversational. Customers can now reply, ask questions, and expect answers in real time. The inbox now talks back, and the implications run far deeper than a friendlier tone.

What is actually changing

Two things are happening at once.

The first is two-way conversational email. Agentic AI now lets a brand respond automatically to customer replies. Questions about sizing, availability, shipping, or returns no longer disappear into a do-not-reply inbox. AI agents can understand intent, access customer context, and continue the conversation within the brand’s guardrails.

Salesforce made this shift explicit when it shipped two-way conversational email in Marketing Cloud Next, describing it as the end of the do-not-reply sender.

The second is the rise of RCS. Rich Communication Services is the successor to SMS, and now that Apple has supported it since iOS 18, brands can deliver rich, interactive messaging across both Android and iPhone. By April 2026, RCS Business Messaging reached roughly 41% of U.S. Android handsets and about 62% of iPhones. Global RCS business traffic hit about 50 billion messages in 2025, up from 33 billion the year before. The performance gap is the reason for the rush: Read rates consistently outpace email, click-through rates are substantially higher, and consumers increasingly prefer rich messaging to traditional SMS.

Taken together, these shifts signal something much larger than a channel update. Direct channels built to broadcast are becoming channels built for conversation.

The consequence most teams miss

The interesting part is not the conversation itself. It is what the conversation triggers.

When a channel becomes interactive, the customer journey stops being a straight line. A reply in email or a tap in an RCS carousel launches the subscriber into a destination: a product page, a landing experience, a shopping cart, a support flow, a loyalty hub. The conversation can fan out into several of these in minutes. Subscribers now bounce across destination experiences, and they carry context with them every time they move.

That context is fragile. If the destination does not honor that context, the thread breaks. The customer who just told you they were shopping for a gift lands on a generic homepage. The customer who asked about a return is dropped onto a promotional banner. The conversation that felt personal one screen ago feels anonymous the next. Momentum dies on arrival.

This is where the expectation gap bites. McKinsey's research is now the standard reference: 71% of consumers expect personalized interactions, and 76% are frustrated when they do not get them. Companies that get it right are not running a marginal play. Fast-growing firms derive about 40% more of their revenue from personalization than slower-growing peers. In a conversational channel, those expectations are not set at send time. They are set the instant the customer acts, and they have to be met at the next destination, not the next morning.

Why adaptability becomes mandatory

In a one-way world, personalization, decisioning, and web curation were competitive advantages. You could win without them and still send a respectable campaign. In a conversational world, they become the cost of entry. Three capabilities move from optional to mandatory.

Decisioning must happen continuously. The old question was "What should I send." The new question is "What should I do next based on what this customer just said or did." Real-time decisioning has to operate across the conversation and the destination as a single loop, choosing the next best experience the moment a customer acts.

Personalization must span the message and the destination. The same profile shaping an email or signal stream that shaped the email or the RCS card must be the same profile and signal stream that shapes the page they land on. When the message and the destination run on different data, the experience splits in two and the customer feels it immediately.

Web curation has to be live. The destination can no longer be a static page waiting to be found. It has to assemble itself in the moment, adjusting recommendations, imagery, layout, and even tone to match the conversation that brought someone there. Miss any one of the three and the conversational channel works against you. You will have invited a dialogue and then failed to follow through on it, which is worse than never opening the door.

How the platforms are responding

The major engagement and data platforms have all reorganized around this shift in the last year. The capabilities are real and worth knowing by name. Razorfish works across these platforms, and the right answer is almost always a fit-to-stack decision rather than a single vendor.

Salesforce, Adobe, Braze, Zeta, and others are combining conversational messaging with real-time decisioning, personalization, and adaptive web experiences. Meanwhile, data platforms including Treasure AI, Amperity, Databricks, Snowflake, Hightouch, Segment, and mParticle are investing heavily in identity resolution, real-time customer profiles, and agentic orchestration.

The technology stacks differ. The direction doesn't. Every major platform is moving toward continuous, connected customer experiences rather than isolated campaigns.

The precondition no one can skip

Beneath all of these investments is the same prerequisite: unified customer identity. Conversation only works when brands know who they’re talking to and can carry that understanding across every handoff. An agent making decisions on stale or fragmented data is not an upgrade. It is a faster way to be wrong in front of a customer who is now paying closer attention than ever, because they are in a conversation.

The data backs the warning. Gartner has found that at least half of the generative AI projects are abandoned after proof of concept, often because of poor data quality and unclear value. The agentic capabilities arriving in every platform above are only as good as the identity and signal layer underneath them. That layer is what determines which brands the conversational shift rewards and which it exposes.

The Razorfish point of view

We see conversational direct CRM as one connected system, not a set of channel upgrades to roll out one at a time. The conversation, the decision, and the destination are all part of the same customer experience seen from three angles, and they have to be built that way: on shared identity, shared decisioning, and shared personalization.

That’s what we mean by the Connected Advantage. Conversational email and RCS are the front door. What determines success is everything behind them: connected decisioning, adaptive destinations, and the data foundation that powers both. 
 
Soon every CRM channel will be able to hold a conversation. The question is whether everything downstream of that conversation, and the data foundation beneath it, is ready to keep it going.