USE CASE: GenAI B2B Email Nurture: How a Web Design Agency Stopped Burning Trust

A real-world GenAI B2B email nurture use case: how a web design agency replaced a drip that spent its prospects' trust with a five-track long-cycle nurture, drafted with GenAI and validated by the people who sell. The lesson is not in the emails. It is in what the system decided not to send.

USE CASE: GenAI B2B Email Nurture: How a Web Design Agency Stopped Burning Trust
AI-generated illustrative image. No real client, brand or location is depicted.

Most writing on GenAI B2B email nurture is about writing better emails faster. This is about something else: using GenAI to reorganise a nurture programme around the buyer's stages, so that each email arrives when it is owed rather than when the calendar says. A web design agency did this with sequences it had been running, unread, for three years, and the part worth your attention is how little of the result was about copy.


The Context: a Web Design Agency Nurturing on a Six-week Timer for a Twelve-month Decision

A web design agency of around thirty people sells website rebuilds and redesign programmes to mid-sized companies: six-figure projects, typically decided over nine to fourteen months by a marketing lead, an IT or digital manager, a finance director and, at the end, the managing director. Its inbound engine worked well enough. A redesign readiness guide and a quarterly webinar produced a steady flow of contacts, each of whom dropped into a nurture sequence named after whichever asset they had downloaded: nine emails over six weeks, with a request to "book a discovery call" at emails three, six and nine, followed by the monthly newsletter.

The sequences reported healthy open rates. Nobody had read them end to end since they were built. And the agency's best-fit prospects, the ones who would eventually run a serious redesign, were the ones most likely to go silent by week four.


The Challenge: the Trust Was Being Spent before the Buyer Was Ready to Use it

A website redesign is bought on a trigger: a rebrand, a new marketing director, a platform reaching end of life, a sales team complaining the site loses deals. The agency cannot create that trigger. Its only job in the months before is to be the choice the buyer already trusts on the day the trigger arrives. The drip was doing the opposite. A marketing lead who downloaded the guide because her director had asked "do we need a new site?" was being asked for a discovery call eleven days later, while she was still trying to define the problem. By email six she had a view of the agency, and it was that the agency was not listening.

Two things made it worse. The sequences were addressed to a person, so when the IT manager from the same company joined the webinar four months later he started the same path from email one, and the account was now receiving two uncoordinated sequences. And the only way out was unsubscribe. Prospects whose projects had been parked for a budget year, a perfectly ordinary reason for silence, kept receiving asks, and some of them left for good. When a client who eventually signed said, in the kick-off, that they had "almost not called because of the emails", the agency stopped treating nurture as a copy problem.

The catch: Nurture does not lose the deal on the day of the ask. It loses it months earlier, on the fourth email the buyer did not need, and the loss is invisible because the buyer does not reply to say so. By the time the trigger arrives, the inbox has already decided.

The GenAI Workflow: from Source-Built Drips to a Long-Cycle Nurture Architecture

The agency's account director and its two salespeople ran the redesign in three passes, with GenAI doing the reading and drafting and the three of them doing the judging.

Pass one: inventory. They exported every active nurture email with its trigger and send timing, forty-one in total, and asked GenAI to classify each by the buying-stage question it actually answered: is this a problem, how do we define it, what kinds of answer exist, which supplier, can we defend the choice. The output showed what nobody had seen: twenty-six of the forty-one emails answered the first two questions, the first ask fell at a point where every account was still defining the problem, and nothing at all addressed the finance director or the managing director.

Pass two: structure. They rebuilt the programme as five Stage Tracks, one per buying question, enrolling accounts by behaviour rather than by download. Movement Rules were written in plain language with sales in the room: an account moves into Supplier Evaluation when a second role appears or when pricing and process content is engaged, not when a timer fires. Trust Rules set a cadence ceiling per account, not per contact, and removed every ask from the first two tracks. Exit Rules added a dormancy track for parked projects, a return path when a signal reappears, and a handoff to sales that carries the account's stage history.

Pass three: drafting. Only now did GenAI write emails. For each track and each committee role it drafted from the agency's own material: anonymised discovery-call notes, the questions prospects actually asked in proposals, the objections finance raised. The drafts went to the salespeople, who cut what no prospect had ever asked, flagged claims the agency could not stand behind, and rewrote the openings in the agency's own voice. The drafts were a starting point; the salespeople's edits were the product.

🖥️
The GenAI prompt:

You are a B2B marketing operations strategist redesigning the email nurture of a web design agency whose buyers decide over nine to fourteen months, by committee. You treat every email as a transaction that either earns the right to the next one or spends it.

INPUTS
[The full text of every active nurture email, with its enrolment trigger and send timing]
[The five buying-stage questions: is this a problem · how do we define it · what kinds of answer exist · which supplier · can we defend the choice internally]
[The committee roles: marketing lead, IT or digital manager, finance director, managing director]
[Anonymised discovery-call notes and the questions prospects asked during proposals]

TASK
1. Classify every existing email by the stage question it answers. Flag any email whose sequence enrols by download rather than by stage, and report where the first ask falls.
2. Propose five Stage Tracks. For each, state the question it answers, the content it should send, and the behaviour that moves an account to the next track. Mark which tracks ask for nothing.
3. For each track, draft one email per committee role, using only language and concerns found in the discovery notes. Do not invent client names, results or statistics.
4. Propose the Exit Rules: when an account passes to sales, when it enters dormancy, and what signal brings it back.

Flag every inference you are NOT confident about, and tell me which data point, email or note would confirm or kill it.

That final instruction carries the weight. GenAI is non-deterministic by design: it produces different drafts on different runs, and it will write a confident, plausible claim about what a finance director worries about with nothing behind it. The variable the agency controlled was not whether the first draft was right, but rather the quality of the material fed in, real call notes rather than assumptions, and the rigour of the salespeople's validation behind every email that went live. If you skip that step, you'll get a beautifully structured programme that says nothing a buyer recognises.


The Result: Fewer Emails, Each One Owed

The visible change was subtraction. Accounts now receive fewer emails over a longer period, and none of them ask for a call until the account's own behaviour says the question has changed. The first two tracks give and do not ask. The programme recognises an account rather than a contact, so a second role appearing from the same company is a signal that moves the account forward, not the start of a duplicate drip. Projects that go quiet for a budget year sit in dormancy, where the agency stays present at a low cadence and asks nothing until a signal returns.

Quietly the larger result: sales now receives an account with its stage history attached, so the first discovery call starts from what the committee has already read and asked, rather than from the beginning. The two salespeople describe the handoff as the first time nurture has told them something they did not already know. No invented percentages here; the change worth reporting is in the mechanism, and the metrics below are where it should show up over the next two or three quarters.


A nurture redesign is only worth doing if it moves something. These are the three to watch, with where the industry sits and the direction this work pushes them. The point is the direction of travel, not a promised number.

Unsubscribe Rate per Nurture Send

The share of recipients who opt out of a given email. It is the Trust Ledger's most honest line: every spike marks an email that was sent before it was earned. With asks removed from the early tracks and a cadence ceiling per account, this should fall steadily and, above all, stop spiking at the old ask points.

Benchmark: The median unsubscribe rate across industries was 0.22% in 2025, more than double the prior year after mailbox providers made one-click unsubscribe easier (MailerLite, 2025); under 0.5% is generally regarded as healthy, with higher rates signalling a content-audience mismatch (HubSpot, 2025).

Click Rate on Nurture Emails

Clicks as a share of delivered emails, read per track rather than per campaign. When each track answers only its own stage question, the content on offer matches what the account is trying to work out, and clicks drift upward. Read it alongside the first KPI: rising clicks with falling unsubscribes is the pattern that says the ledger is in credit.

Benchmark: Median click rate of 2.09% across 3.6 million campaigns, with a click-to-open rate of 6.81%; agencies as a sector record an open rate of about 40.5% (MailerLite benchmarks for 2026).

Stage Movements per Quarter

The number of accounts that fired a Movement Rule and advanced one track, counted by track. This is the metric the drip could never produce, and the one that measures the buyer rather than the email. Expect it to start small and build as the early tracks fill; a track with engagement but no movements is a track whose Movement Rule or content needs another look.

Benchmark: No reliable public benchmark; this is an internal, system-specific measure. Set your own baseline in the first quarter and track the trend by track.

Your own trend, quarter to quarter, matters more than any single benchmark; the benchmark is context, not a scoreboard.


Why this Transfers

Every long-cycle B2B business is running a nurture that was designed for a shorter decision than the one its buyers make. The job was never to write warmer emails, but rather to organise the programme around the buyer's stages, govern how much it sends, and design the exits before the buyer designs them for you. GenAI makes the reorganisation an afternoon's work instead of a quarter's, on one condition: real buyer material in, real sales judgement out.


The Nurture Architecture: How to Build Email Sequences That Move B2B Buyers Forward Without Burning Trust
B2B buyers do not stop reading your nurture because the emails are bad. They stop because every unearned email spends trust that does not refill. Here is the four-component architecture that moves an account forward across a long cycle, and knows when to go quiet.