Mid-2026 Review: What Does AI-Powered B2B Marketing Actually Look Like?

Half-Time Whistle: The Split Between AI Penetration and Budget

Last week I visited a well-known business school in China. When the conversation turned to enrollment, the anxiety was unmistakable—applications used to come easily, but recruiting has become a major pressure point.

Why start with a business school? Because enrollment logic maps remarkably well to B2B marketing. Both require understanding the prospect's company, industry, role, and decision-making position. Average deal sizes are high (an MBA program can run from tens of thousands to hundreds of thousands of RMB). And the buying journey is long—no single person makes the call alone. The challenges business schools face are, at their core, the same ones every B2B marketer in the room is dealing with.

That business school's anxiety is not an isolated case. Halfway through 2026, B2B marketers collectively feel a split: on one side, AI has moved from slide-deck talking point to everyday tool, with penetration at an all-time high; on the other, budgets are tightening, and the prevailing feeling is that methodology is scarcer than tools.

"Budgets did not rise with AI penetration—they fell. You can no longer solve problems by throwing money at them. You have to be smarter, not louder."

This article captures trends and insights from a recent stretch of intensive customer visits. No fear-mongering—just what works in the field.

1. AI-Native Websites — Your 24/7 Sales Engine

Corporate websites are going through a quiet transformation. Updating a site used to be painful. Publishing a blog meant planning, drafting, formatting, and uploading; changing a case study page meant pulling in engineering—and a one- or two-week delay was normal. Today, much of that work has been turned into repeatable Skills.

What does that mean? You can use AI agents to automate these tasks. Blog updates, case study updates, title/description/keyword (TDK) optimization, internal linking, and layout all become much simpler—updated directly through Skills. Layout quality improves too, with SEO keywords and internal link structure handled consistently every time.

But AI-native websites are not just about update efficiency. The deeper shift is that the site is moving from a static information page to an AI-powered engine for data capture + personalized outreach + lead conversion.

"A website is no longer just an information page. It is becoming an AI-driven agent—your 24/7 sales engine that never sleeps."

2. AI Voice Outreach — MDR, Not Spam

Many people hear "AI outbound calling" and think spam. That bias needs to change. In B2B, AI voice has a critical use case we call AI MDR—Market Development Representative. Unlike traditional SDR roles, MDR's core job is not closing deals directly. It is lead cleansing, activation, and needs confirmation. AI voice is a natural fit for MDR workflows.

But there is a prerequisite—use it only with people who already have trust in you, or who expect the outreach. You cannot launch AI calling blindly. Cold AI calls to new prospects once or twice a year may be tolerable; do it too often and you create harassment. That is the wrong play.

• Event invitations—Prospects attended last year and the year before; this year's reminder is a helpful nudge, not a cold dial
• Pre-event confirmation—Confirming attendance ahead of important meetings; generally well accepted
• Lead cleansing—Confirm company, title, and email via voice, and sync updates back to CDP fields

I recently worked with a life-sciences customer. They use AI EDM with strong results, but some emails never get opened. They decided to try large industry conferences and use voice outreach for invitations. During the call, the AI might say: "I'll need to send you the invitation—can you confirm your current company, title, and whether this email is still correct?" Once confirmed, the invitation goes out—and in that process, the contact record gets refreshed.

The voice data generated by AI calling then feeds back into your CDP—updating contact fields and intent scores. That is a classic, high-value pattern.

"The key is not volume—it is precision. AI MDR is not about how many calls you can make in a day. It is about activating the right people, with existing trust, at the right moment."

3. The Evolution of SDR — From Volume to AI Augmentation

In my article on the 20-year paradox of the B2B marketing funnel, I analyzed a common mistake: the more data you accumulate at the top of the funnel, the less you have in the middle and bottom. At the top, marketing spends heavily on ads, content, and events—and leads pour in. But in the middle—SDR follow-up, lead scoring, and nurture—data dimensions shrink sharply. At the bottom—opportunities, closed deals, renewals—it becomes a black box.

That top-heavy funnel is the root problem. And AI + SDR is starting to fix it.

First shift: the return of data dimensionality

More customers are realizing that data assets matter. Once data richness becomes a priority, bringing voice data back into the stack is essential. We have customers—whether in-house or outsourced SDR teams—doing the same thing: piping SDR voice data back into the system.

One case stands out. A customer's voice data lived in an on-prem deployment and could not be extracted. Internally, they pushed through layer after layer of resistance. Why? Because security must serve the business. Data that cannot support these capabilities is not a strategic asset. Feeding voice back into the CDP makes the entire customer profile richer. That is absolutely worth doing.

Second shift: AI across the post-SDR workflow

The traditional pattern: SDR makes a call, and then… nothing. Today it is different. No answer → automatically trigger follow-up. Connected → use voice insights to trigger the next steps in the journey—touchpoints on day 3, day 5, and beyond.

Asking humans to do this goes against human nature. Call, no answer. Call again, no answer. Call again—how many SDRs keep going? AI can follow up persistently, without fatigue and without emotional drag.

I recently spoke with the CEO of a Hong Kong–listed ticketing company. He said something that stuck: "SDR is fundamentally a sales role, and the hardest quality in sales is persistent follow-up."

"AI handles persistent, tireless follow-up and lead nurturing. SDRs handle deep conversations on high-value leads—that is the right division of labor."

4. WeCom's Rise — An Underrated Marketing Channel

From what we see in the field, WeCom (WeChat Work) is becoming critically important. The underlying reason: WeChat Official Account feeds are being folded. Once that happens, the channel left in the WeChat ecosystem for ongoing customer engagement is largely WeCom. Tencent built it as an outbound channel—it integrates well with Channels, WeChat, and mini programs.

Folding rolled out gradually from 2024, scaled through large-scale gray releases in 2025, and by 2026 it is the norm. You may think your carefully planned campaign posts are being seen; in reality, most users never scroll to them.

Once customers are on WeCom, you can run many useful plays. Deploy an AI chatbot for pre-sales conversations and move prospects toward agreement. Push a mini program—when a prospect submits a form, binding rates rise because you know who they are, and your sidebar can surface relevant content.

There is another strong mechanism: when a lead comes in, assign it to an SDR for a call or to WeCom for chat. Completing either counts as success. Phone and chat run in parallel and complement each other. Some prospects hate calls but will chat; others pick up instantly but ignore messages. Cover both paths and conversion improves.

"WeCom was severely underrated before 2026. After Official Account folding, it became the only effective outbound channel in the WeChat ecosystem—messages reach customers proactively instead of waiting for them to find you in a buried feed."

5. Multi-Touch Strategy — 75–95% of Leads Need Long-Term Nurturing

Last week I had a long conversation with the marketing director at a leading B2B marketing services firm. She is a paid-media expert, and her company takes WeCom seriously. I asked her: adding an email field to a lead form often lowers conversion rates and makes acquisition harder—so as a media expert, why do so many of your leads still include email?

Her answer was illuminating. "Keep fields to a minimum—usually company name, name, and phone. Everything else gets filled in on follow-up." So when SDRs call back, they always confirm email.

I asked: if you need WeCom and email, do those conflict? She said: "No conflict. Customers need multiple touchpoints—you never know which channel will move them."

In other words, 75–95% or more of your leads are dormant accounts marketing must nurture over time. That number matters. Of 100 leads you paid for, only about 25 have a near-term path to opportunity, and only about 5 will close. What about the remaining 75–95? Do not discard them. They are assets—just temporarily asleep.

That is where multi-touch strategy earns its keep. Email, WeCom, AI voice, website chatbot, content pushes—each touchpoint activates dormant leads at different times, in different ways. A personalized email goes unopened today; a WeCom industry report next week might get a click. An offline event invitation next month might deepen the relationship.

"Multi-touch is not channel stacking—not blasting the same message everywhere. It is building an encircling outreach network around each lead—right time, right channel, right message."

6. CRM Is Not a Marketing Hub — Proving ROI Requires an M2L Closed Loop

Last week I met with another very large customer whose marketing platform was broken. They were using CRM as their marketing hub—creating forms in CRM and running nurture flows on a foreign platform. That setup was doomed from the start.

Using CRM as a marketing hub is like making slides in Excel—you can do it, but it is painful and the results are poor.

The AI era has arrived. Everyone talks about AI agents and how AI will change marketing. But structured foundations must come first. A simple truth: if your data foundation is weak, the smartest AI agent cannot help—it is analyzing incomplete, disconnected data. It is like giving a genius a map where only one-third of the landmarks are marked. Can they find the way?

Another related case: a customer told me they used a marketing product internally and the team liked it, but they could not prove value to leadership or the CEO. Why? Leads were not flowing through M2L (Marketing-to-Lead) logic. There was no backend integration with CRM to complete the M2L loop. Leads stayed in marketing—they never reached Sales Qualified Lead status or became opportunities. The full marketing ROI was invisible.

7. AI EDM — A Legacy Channel's New Surge

Over the past two-plus months, the new version of AI EDM has been live, and many customers are using it at scale. Volumes are large—some send 200,000–300,000 emails per week; others send 70,000–80,000. Results have been strong.

I have said for years that email will not die. People have been predicting its decline for ages—open rates falling, younger audiences ignoring it—but AI has brought AI email marketing back to life.

Data quality determines AI outcomes. Garbage in, garbage out. That is not an AI problem—it is an input problem. But if you have a solid CDP with years of accumulated data, that is a major advantage.

"Once data cleansing is in place, AI EDM becomes a high-efficiency marketing automation engine for lead generation. No need to write emails one by one—AI handles it, and you focus on results."

8. Content Production and GEO — Great Content Is Always the Foundation

GEO (Generative Engine Optimization) is one of the hottest topics in marketing in 2026. In 2023, researchers at Princeton University introduced GEO—aimed at getting content selected and cited by AI engines, not just indexed by traditional search. By 2026, as an estimated 30% of search requests bypass traditional search engines, GEO's practical value is clear.

But GEO has a real problem: ROI is hard to prove. You optimize for generative engines—how do you show how many leads or opportunities that produced? You cannot, easily. AI engines do not give you clean traffic source tracking like Google Analytics.

Still, the core opportunity in GEO is the same as always: produce great content. I spoke last week with an industrial software customer from Germany. Even with AI, they find content production difficult. Quality content is always hard. That is reality. No matter how far AI evolves, deep industry understanding, precise grasp of customer pain points, and real insight into business scenarios—AI can assist, but it cannot replace human judgment.

"Great content is the fuel for every outreach channel. Without it, more touchpoints and stronger AI are water without a source."

9. NBA — Let AI Decide What Happens Next

One customer uses a concept called NBA—Next Best Action. The logic: when a lead enters the system, AI can analyze what should happen next. Send an email? Message on WeCom? Human outreach? Invite to an event?

AI weighs behavioral data, interaction history, and industry attributes to recommend the optimal next step. This is not simple rule-based automation—"if they opened the email, send WeCom"—that is traditional marketing automation. NBA is genuine intelligent decision-making: AI evaluates all available data and picks the action with the highest probability of success from several options.

"From 'humans decide what to do' to 'rules decide what to do' to 'AI decides what to do'—that is the inevitable path of marketing efficiency."

Early Adopters Win

Many marketers ask me: the economy is tough—what marketing tactics should we use? My logic is simple: do the math. If a new approach costs less than the old one, use it boldly. On a call with a software company founder recently, I walked through exactly that calculation. If you can prove the new approach is cheaper than what you were doing before, you should experiment aggressively.

But you have to actually run the numbers—not gut-feel "this seems to work," but data: how much went in, what came out, how much you saved versus the old method. When the math is clear, execute. When it is not, test until it is.

Looking back at the first half of 2026: 96% of marketing teams are using AI, and 34% of enterprises have deployed AI agents. The easy gains from tooling are behind us. What matters next is who has clarified their methodology, built their data foundation, and established a multi-touch system. The gap between AI-enabled teams and those still watching will only widen. That is not alarmism—it is what is happening now.