Email Isn’t Dead. Your Generic AI bs nurture Is. A 2026 playbook for B2B emails that sound human and actually move pipeline.
The only thing about email that’s dead — or should be, is that at some point we all started copying each other and eventually B2B emails were just corproations emailing other corporations and the human element was completely abandoned. Today:
Open rates hovering in the 20–40% range depending on industry. HubSpot Blog
Filters more aggressive about look-alike, high-volume content. Litmus
A wall of LLM-flavored emails that could’ve been written by any vendor in your category.
The good news: email is not dead. It’s just crowded, and AI has made it cheaper than ever to ship mediocre messages.
The B2B teams who will win in 2026 are the ones treating email as a human, high-signal channel and letting AI play the role of diligent intern, not CMO.
Here’s how we’re thinking about email in the AI era at R&D.
1. Start with the name, not the subject line
Most marketers obsess over subject lines. Your buyers start somewhere else:
Who is this from?
→ Do I care?
→ Then subject line.
Eye-tracking and UX studies have shown for years that readers look at the sender name before the subject line. ScienceDirect In practice, that means:
Pick a house human.
“Rachael from R&D B2B Marketing” will usually beat “R&D Monthly Newsletter.” People reply to people, not brands.Create sub-identities by job.
Don’t send everything from the same bland address:Thought leadership:
Rachael at R&D B2B MarketingProduct / feature updates:
[Product] Team at [Brand]Events / community:
[Name] from [Program/Event]
Test the From name before you burn 10 more subject line tests.
Across different programs, we’ve seen sender changes swing opens by ~20–30% when the content behind that name is consistent and relevant. Superhuman Blog
2026 is the year to define your house humans for email and stick with them long enough to build familiarity.
If your own program is still sending from marketing@company.com, that’s a January fix.
2. Stop shipping AI-flavored nothing-burgers
We can all recognize the generic AI email now:
“We’re revolutionizing X with our innovative platform so you can unlock the power of Y and accelerate Z.”
You could swap logos, and nothing would break.
In an AI-saturated world, email differentiation comes from three things:
Sharp positioning
“We help [specific ICP] solve [1–2 painful problems] in [concrete way].”
If your product is for mid-market energy compliance teams, say that. Don’t write like you’re selling to “businesses of all sizes.”A clear point of view
What do you believe about your category that not everyone agrees with?“Most carbon dashboards are vanity reports, and your board knows it.”
“If your AI data center plan doesn’t start with power, it’s a wish list.”
Recognizable voice
A human can hear the difference between:“We help you unlock the potential of your data.”
“You already have the data. We help you stop guessing what to do with it.”
AI is a force multiplier only when it’s fed strong human source material: customer calls, win/loss notes, objection logs, stories from the field. Vague inputs = vague outputs at scale.
If you stripped your logo from your last email, could your closest competitor send it without editing a word? If yes, you don’t have an email problem, you have a positioning problem.
3. Treat the inbox like personal space
Your buyer’s inbox is closer to their kitchen table than to your homepage.
If someone invites you in, you need to know what role you’re playing:
Coach – “Here’s how to think about methane rules this quarter.”
Curator – “We read these 19 boring PDFs from COP 30 so you don’t have to.”
Co-pilot – “Here’s what we’re testing alongside customers like you.”
Intern – “I tried this in the wild, here’s what broke and what worked.”
One pattern we like: the “intern persona.”
On a few client programs, we framed a series as updates from “a very honest intern trying to make sense of this product and market.” It works because:
It’s inherently human and humble.
It creates a story arc you can build on over months.
It gives you permission to share experiments, not just polished wins.
Whatever role you choose, write to one person: the brilliant friend or favorite coworker version of your ICP. If your email wouldn’t survive as a 1:1 note to them, it’s probably too vague or self-absorbed.
And then: ask for replies.
Replies:
Send a strong positive signal to inbox providers. Litmus
Surface objections you’ll never see in click data.
Turn a “campaign” into a conversation.
If your program never explicitly invites replies, that’s a missed lever for both deliverability and relationship building.
4. How to say no to “email soup”
You’ve been in this meeting:
“Can we put the feature launch, the webinar, the podcast, the pricing change, and the new case study in the newsletter?”
Most humans give your email 5–10 seconds before they decide if it’s worth their time. Nine priorities equal no priority.
Give yourself language to push back:
“For this audience, the Q1 newsletter needs to do one job really well.”
“Let’s give execs the story and a single CTA, and use in-app nudges for end users.”
“We can run a follow-up campaign for the other asks over the next 10 days instead of stuffing them all into one scroll.”
As a rule:
1 email = 1 job.
9 jobs = 0 email.
5. Put AI to work, but don’t let it take the mic
AI is great at being a tireless junior. It’s terrible at being your head of marketing.
Here are ways we see it working in email today (that aren’t embarrassing):
Outbound assist for BDRs/SDRs
Pull firmographics, tech stack, recent funding, and hiring trends.
Suggest 1–2 relevant insights or customer logos.
Draft an intro paragraph that a human can tighten in 30 seconds.
Section-by-section support, not full email generation
“Give me 3 alternative intros for this story, 60–80 words each.”
“Tighten this feature explanation for a VP of Ops, but no buzzwords.”
“Turn this customer quote + bullet notes into a short use-case vignette.”
Mandatory human edit pass
Every draft goes through a human filter for:Weasel words (“innovative,” “game-changing,” “end-to-end”).
Specificity (do we name workflows, numbers, or just vibes?).
Voice (does this sound like us?).
ICP rehearsal with AI personas
Take your ICP doc and ask:“Where would you stop reading this email?”
“What’s confusing or irrelevant to you?”
“What’s missing for you to say yes to a demo?”
Adjacencies worth testing in 2026
Lead magnets with small “mini chat” assistants that answer questions about the asset.
Post-webinar follow-ups written from transcript highlights, not generic “thanks for attending” templates.
Treat AI as the intern who never gets tired. Your job is to be the editor and strategist, not the prompt that ships unedited.
6. Choose the right surface: email, in-app, SMS
Not every message belongs in an email.
A simple mental model:
Right message → right moment → right surface.
Email = relationship channel
Use it for narrative, thought leadership, and messages that benefit from reflection or forwarding inside an org.In-app = behavior channel
Perfect for active users:“You started this workflow, here’s how to finish it.”
“Power users who do X get 20% better Y.”
“This new feature is live in the part of the product you already use.”
SMS = sacred space
In B2B, treat text like a fire alarm, not a content channel:OTPs and security codes.
“Your workshop starts in 10 minutes” level reminders.
Critical service notices.
If you’re using SMS for weekly promos to B2B buyers, you’re probably burning trust.
7. Measure what matters, and know when to let go
Before you open a dashboard, answer one question:
What job is this email supposed to do?
Different jobs, different success metrics:
Thought leadership
Replies (“we’re dealing with this exact problem”).
Forwards / CCs inside target accounts.
Clicks to linked content.
Product updates
Feature adoption and usage.
Help center views and self-serve success.
Downward trend in related support tickets.
Sales enablement / outbound
Meetings booked.
Opportunity creation and movement.
Deal velocity and size.
Opens and clicks still matter, but they’re health indicators, not the whole story. Benchmarks suggest average open rates in the 30–40% range across categories, with CTRs around 2%. HubSpot Blog: Your job is to understand why your numbers look the way they do, given your audience, list age, and message mix.
On sunsetting:
Define a cooling-off path:
No opens in ~90 days or 8–10 campaigns? Move them to a low-frequency re-engagement track.
Still no engagement? Suppress from most promos; keep only mission-critical messages.
Be explicit internally:
List size is not the goal. Deliverability and relevance are. Clean lists, high engagement, and active pruning send better signals to inbox providers. Validity
Hoarding disengaged contacts is like bragging about a giant warehouse full of products nobody buys.
8. Tools and workflows actually worth your time
There’s no shortage of “AI for email” tools. A few foundational pieces we actually like:
ChatGPT / Claude
Great for:Drafting alt versions of copy.
Condensing long internal docs into email-sized insights.
Role-playing objection handling or ICP personas.
The key is feeding them your brand voice guide, ICP definitions, and real customer artifacts (call transcripts, win/loss notes, support logs) so they’re working from your reality, not the generic internet.
NotebookLM for deeper work from your own sources
Shines when you need to turn dense materials into usable email narrative:Upload RFPs, interview transcripts, research reports, and long slide decks.
Ask it to highlight 5–10 insights your buyers would actually care about.
Use it to draft FAQs or story angles grounded in your own material. Media Shower
Think of this stack as your research and drafting lab. Email becomes the distribution layer for what you learn.
The tools are not the strategy. They’re multipliers for the quality of your thinking and inputs.
9. So what do you do with all of this?
If you remember nothing else, make it this:
Treat sender names as strategic real estate.
Make your messages so specific that a competitor couldn’t steal them.
Respect the inbox as personal space.
Give each email one job.
Let AI be the intern, not the owner.
Measure what matters and prune ruthlessly.
Email isn’t dead. Bad, generic email is dead.
The rest is an opportunity.
