Dave Schools is a B2B marketer building Singulate, an email marketing platform focused on AI-assisted segmentation and personalized campaigns. His posts explain how to use data and messaging strategy to make personalization more relevant—and how to avoid generic, unchecked AI-generated copy.
Most useful for practical guidance on AI-assisted email personalization, anti-slop rules, and where automation should stop.
Best for
B2B agency teams running email and lifecycle campaigns wanting a practical AI personalization workshop.
Not for
Paid endorsement campaigns with guaranteed impressions and performance targets for sales-engagement vendors.
Key takeaways
Personalize only sentences that need it and check generated copy against contact data.
Use examples to guide output and prefer ordinary segmentation where AI adds no value.
A company-reported split test across 46 campaigns reported more clicks for Singulate's version.
Check first: Confirm partnership terms, audience composition, and any comparable workshop attendance results.
Main ideas
AI slop is a cost and quality problem, not a matter of taste · Data is not the bottleneck, creativity is · Personalization is a build with five named decisions
Dave Schools is a B2B marketer whose LinkedIn posts focus on AI-assisted email personalization and Singulate, the email platform he discusses building.
He is known for arguing that useful personalization needs more than contact data: it also needs a clear messaging strategy and human review.
His posts offer practical guidance for building email agents, avoiding generic AI copy, and deciding where automation should stop.
They also show a broader interest in chat-based AI assistants and in making B2B marketing more distinctive.
Readers interested in the practical choices behind AI-powered email—and the risks of automating it without enough control—will find his work especially relevant.
Schools connects a sharp warning about AI-generated sameness to specific working practices.
In one post, he advises marketers to personalize only the sentences that need it, check generated copy against contact data, use examples to guide output, and prefer ordinary segmentation where AI adds no value.
In another, he describes a split test across 46 campaigns and reports more clicks for Singulate's version; that is a company-reported result, not independent confirmation.
His argument is not simply to automate more: he advocates combining agentic tools with transparency and direct editing.
That tension makes his work useful to readers weighing speed against relevance and control.
A personal account of using AI assistants to cancel subscriptions and book a restaurant shows the same interest in getting from a chat to a completed task, beyond email marketing.
Dave Schools writes mainly about AI-assisted email personalization, and the argument running through most of the posts is that the technology is not the hard part.
Read more
The hard part is deciding what to say with the data, and keeping a human in the loop so the result does not read like unchecked machine output.
That position shows up as a rule-based anti-slop method, a split test on 1.4 million emails, a five-decision build checklist, and a product argument that automation should stay editable.
A second, broader theme sits outside email: Dave Schools treats personal AI assistants as a shift in how work gets done at all, and argues that when everyone uses the same tools, standing out matters more rather than less.
AI slop is a cost and quality problem, not a matter of taste
The recurring argument is that AI-generated email copy has become cheap, fast and everywhere, and that most teams treat the problem as stylistic when it is really operational.
Read more
The fix is framed as a set of constraints a team can apply to live campaigns: prompt one or two sentences at a time rather than the whole email, never send AI messaging on autopilot without checking outputs against the contact's actual data, write the exact sentence and only swap in the personalized parts, avoid giant essay prompts because model variability makes them unreliable and expensive at scale, use ordinary segmented dynamic content where AI adds nothing, run a prompt over personalization tokens to clean and condition messy data, and give the model two or three examples of good output to keep it on the rails.
The list of phrases that "shriek AI": "I noticed that...", "I came across your...", "I saw on LinkedIn", and any line that tells the recipient their own job title, company or industry. The food analogy carries the whole argument: fast food and soda will not kill you immediately, but as a habit they make you noticeably unhealthy.
Dave Schools argues that personalization fails less because teams lack data than because they lack a messaging strategy for using it.
Read more
The two halves are stated explicitly: data is one side, how you instruct the agent to use that data is the other, and improving personalization keeps moving results until you reach the scale of Amazon or Netflix.
A test of 1.4 million emails sent through Singulate since mid-2024, run as a split inside the same email with the same day, subject line and sender.
Reported results: across 46 campaigns the personalized group clicked 74% more, roughly 37 clicks per 1,000 recipients against about 21, personalizing two parts of the email produced 3.7 times more clicks, and open rates were nearly identical, meaning the subject line drove opens and the body drove clicks.
The illustration of the principle: knowing someone's city is not a reason to drop in a city token, it is an opening to cite a landmark they pass on the way to work, reference local weather or news, or buy them lunch from a restaurant they recognize.
Personalization is a build with five named decisions
Rather than a single prompt, personalized email is presented as five choices a team has to make deliberately: the use case, kept specific; the model; the data to personalize against, such as form fill, industry or job title; the messaging strategy that turns that data into something worth reading; and where the personalized snippet sits inside the email.
Read more
Three follow-up steps are kept short: A/B test it so you can measure whether it works, make sure it does not sound like AI slop, and confirm it integrates with the existing stack.
The named use cases are deliberately narrow: a post-demo nurture, a webinar invite, a closed-lost re-engagement. A customer called Brendon is described as having built this setup end to end, and the post promotes a webinar where he walks through it.
The stated position is that both extremes fail: the blast email is dead, but full autopilot personalization only produces slop and then creates more cleanup work.
Read more
The proposed middle is agentic tools that build branded email quickly while keeping transparency, flexibility and direct edit control, with the platform deciding how to execute a messaging strategy the user has set. It is presented as a balance question, not a feature list, and the practical version for a reader is deciding which steps in their email workflow to automate and which to keep manual.
Studio is described as "like Claude Design but for marketing operations": build by chatting and commenting, or make direct edits yourself, with the claim that a solo user can now do targeting that previously took a large marketing team. The launch post frames the same point as six numbered comparisons, each built as "most tools do X; this does Y," covering drafting, deterministic personalization through Wow Tags and branching, data enrichment, fine-tuning, segment variants, and replacing a stack of separate data providers and models.
The argument moves from writing rules toward product design.
The anti-slop pieces ask marketers to constrain their prompts; later posts claim that deterministic tags, branching and fine-tuning remove the slop risk at the platform level instead.
Those are Dave Schools's own product claims, not independent findings.
This is the marketing philosophy under the anti-slop work: as AI tools spread, output converges into a sea of sameness, so differentiation through humor, shock, raw human storytelling and intensity becomes more valuable, not less.
Read more
The blunt conclusion is that playing it safe and predictably will not work in B2B marketing.
A recap from Exit Five Drive in Stowe, Vermont, which credits Dave Gerhardt's origin-story keynote and keynotes by Ann Handley and Seth Godin, who are reported as arguing for quality and thoughtfulness over speed and efficiency with AI; Kara Hardin's "Part of me" framework; and Jess Cook's Baby Got Brand workshop.
The post also repeats a coined term heard from a colleague, "Clauderboarding," for over-using and over-sharing Claude output.
Those views are attributed to the named speakers, not presented as Dave Schools's own research.
Outside email, the same thesis is applied to personal AI assistants: the change is not that tasks get faster, it is that the time it takes to note a task down is now roughly the time it takes to get the whole thing done.
Read more
Dave Schools frames this as a paradigm shift in personal workflow rather than a product claim, because an assistant with logins and tools completes things directly instead of leaving a reminder.
A few minutes of chatting with the assistants Hark and Muse: Hark cancelled five subscriptions for $872 a year plus a $22 refund and paused a lawn-care service at $1,163 for the year, access to Rocket Money surfaced further savings, and Muse researched and booked an OpenTable reservation and analyzed 7,000 LinkedIn connections for introductions into key accounts. In a side test of five image tools for a custom icon, ChatGPT was the only one that produced a clean transparent PNG, with Recraft and Nano Banana described as near useless.
A steady stream of posts doubles as a running account of how Singulate is built and sold, naming the tools and channels rather than describing the motion abstractly.
Read more
This is useful for a reader trying to judge whether the ideas come from hands-on operator experience or from commentary.
A "build month" update lists a founding account executive joining, HubSpot cleanup, stronger Claude automation, outbound and inbound nurture run through Singulate with HubSpot and Instantly, Heyreach and LinkedIn ads named as channels, attendance at Exit Five Drive and HubSpot Unbound, two marketing videos for a product update, a webinar with Brendon B. titled "How I Doubled Sales From Email By Building My Own AI Personalization Engine," and a customer result of a 40% open rate against a usual 25-33%. A separate post names the team members to look for on the Unbound floor and the questions they handle: scaling ABM, webinar invites, database re-engagement and personalized nurture journeys.
It answers the question people actually arrive with: what does anti-slop mean in day-to-day work?
Read more
This is the most direct test of whether the workshop topic has substance behind it.
Argues that personalized email has swung from dead blast emails to autopilot AI writing that produces "slop" — output that is cheap, fast and everywhere, and unhealthy as a habit. It then lists eight concrete rules: use one prompt for one or two sentences rather than a single prompt for the whole email; never send AI-generated messaging on autopilot without reviewing outputs next to the contact data; be prescriptive, writing the sentence you want with only the personalized parts swapped in; avoid gigantic essay prompts, because model variability makes them unreliable and expensive at scale; use segmented dynamic content where AI is not needed at all; apply a prompt to clean and adapt raw personalization tokens instead of dropping messy data straight in; cut phrases that announce AI, such as "I noticed that..." or "I came across your...", and drop any line that tells the recipient their own job title, company or industry; and give the model two or three examples of good output to constrain it.
For brands planning a campaign
Each rule converts into a workshop exercise: take a real email, find the slop patterns, rewrite the snippet. The eight rules also leave the audience with a checklist they can keep.
This is the post most likely to be quoted back at a workshop, because it makes a specific performance claim and describes the test behind it.
Read more
Read it before promising any numbers.
Reports a within-email split test — same day, same subject line, same sender — comparing a "singulated" version against a traditionally personalized one across 46 campaigns.
The stated findings: the singulated group clicked 74% more; about 37 clicks per 1,000 recipients against about 21; when two parts of the email were personalized, clicks rose 3.7 times; and open rates were nearly identical, so the difference showed up after the open.
The lesson drawn is that data is only one side of personalization and messaging strategy is the other.
Knowing a contact's city is not personalization on its own; it is an opening to reference a local landmark, recent local news or weather, or lunch from a local restaurant.
"Data is not the bottleneck, creativity is."
For brands planning a campaign
Supplies a reusable test design and a set of claims to build on or challenge. If the workshop quotes the percentages, present them as Singulate's own internal measurement.
A short, practical checklist answering "where would I even start?" — useful as the spine of a workshop agenda if you need one.
Read more
Lays out five choices for anyone building a personalization agent: pick a specific use case (post-demo nurture, webinar invite, closed-lost re-engagement); pick the model; pick the data to personalize with, such as form fill, industry or job title; pick the messaging strategy for that data; and pick which email the snippet goes into and where inside it.
Three tips follow: A/B test it so you can measure whether it works, make sure it does not sound like AI slop, and make sure it integrates with the existing stack such as HubSpot or Marketo.
A customer named Brendon is described as having built the same setup end to end.
For brands planning a campaign
Works directly as a session outline: five decisions, three tips, one case. It also signals what the audience already knows, which shapes how much explanation the workshop needs.
The product announcement shows where this person's own tooling sits and how it is pitched against other email platforms — context a workshop host needs before repeating any claim.
Read more
Announces Studio, a self-serve plan on the agentic newsletter platform, free for 14 days and $49 a month after. Six points frame it against other tools: the agent drafts, designs, segments and manages one-to-one personalization through chat rather than stopping at writing; deterministic personalization through "Wow Tags" and branching is said to remove hallucination risk on lists above 10,000; contact data can be enriched by the platform; fine-tuning is offered so the personalization strategy holds up at scale; the tool writes segment variants as well as the base email, turning "72 emails into what feels like 1"; and it replaces stacking a data provider such as ZoomInfo or Apollo with a separate model like Claude or ChatGPT.
For brands planning a campaign
These are the founder's own product claims, not independent results. They are useful in a workshop for a build-versus-buy discussion and for asking what "deterministic" personalization means in practice.
States the position the rest of the work argues from, in a single page.
Read more
Good for a reader who wants the argument before the tactics.
Argues that both extremes fail: blast email is dead, but autopilot personalization only creates slop and extra cleanup work.
The proposed middle is agentic tools for building branded email quickly, plus transparency, flexibility and control.
Studio is described as "like Claude Design but for marketing operations" — build by chatting and commenting, or edit directly — with the claim that a solo user can now do targeting that previously took a large marketing team.
The closing point is that the choice is not all-or-nothing between automation and manual building.
For brands planning a campaign
The balance framing is the most arguable claim in the set. Treat it as a pro-and-con segment rather than an uncontested premise.
Shows the wider agent thesis beyond email, and the personal, numbers-first storytelling style.
Read more
If a newcomer reads only one post for tone, start here.
Describes a few minutes of chat with the personal assistants Hark and Muse.
Hark cancelled five subscriptions for $872 a year plus a $22 refund, and paused a lawn-care service at $1,163 so far that year.
Giving Hark access to Rocket Money surfaced further savings.
Muse researched and booked an OpenTable reservation and analysed 7,000 LinkedIn connections to find warm introductions into key accounts.
A side test of five image tools for a custom icon found ChatGPT the only one producing a clean transparent PNG, with Recraft and Nano Banana described as near useless.
The takeaway: the time it once took to write a to-do now covers getting the whole task done — "action at the speed of chat."
For brands planning a campaign
A concrete, self-deprecating opening story for a workshop because it is not product-led. It also introduces the agent framing that connects back to email.
Explains the marketing philosophy behind the anti-slop position — why standing out matters more, not less, when everyone has the same tools.
Read more
A conference recap from a second year at Exit Five Drive in Stowe, Vermont.
It notes the coined term "Clauderboarding" for over-using and over-sharing Claude output, keynotes from Dave Gerhardt, Ann Handley, Seth Godin and Kara Hardin, and workshop highlights including Jess Cook's Baby Got Brand and a Customer.io session.
The loudest message taken from the event: when everyone uses the same AI tools and everything converges into a "sea of sameness," being weird, remarkable and off-the-wall intense is the differentiator, through humour, shock and raw human storytelling.
It warns that playing it safe and predictably will not work, and closes by asking attendees for their own takeaways.
For brands planning a campaign
Gives the workshop an argument that anti-slop is a commercial stance, not just a quality preference. The conference quotes also make a lively opening slide.
Anti-slop as a repeatable method, not a style preference3
Building and marketing a company in public3
Where email and AI are heading2
Building an AI email personalization agent, decision by decision1
The 1.4 million email split test1
Agentic automation with a human hand on the controls1
Launch posts as structured comparisons1
Distinctiveness over sameness in B2B marketing1
Building an AI email personalization agent, decision by decision
Dave Schools breaks personalized email into five choices a team has to make: the use case (a post-demo nurture, a webinar invite, a closed-lost re-engagement), the LLM, the data to personalize against, the messaging strategy that turns that data into something worth reading, and where the personalized snippet sits inside the email.
Read more
The post then adds a short checklist: test it, make sure it does not sound like AI slop, and confirm it integrates with the stack, naming HubSpot and Marketo. The through-line is that personalization is a build with named parts, not a single prompt.
Names each decision in order, keeps each one narrow ( a post-demo nurture, rather than all email ), and pairs the framework with a practical checklist.
For brands planning a campaign
A workshop can run a team through these five decisions using the client's own sequence as the working example, then leave with the checklist applied to one live campaign. Ask which of the five decisions the client has never explicitly made.
Anti-slop as a repeatable method, not a style preference
The clearest teaching in the corpus is a list of eight fixes for AI-generated email copy: prompt one or two sentences at a time instead of the whole email; never send on autopilot; tell the model to write an exact sentence and only swap in the personalized parts; avoid giant essay prompts because they are unreliable at scale and expensive; use ordinary dynamic content where AI adds nothing; clean and condition personalization tokens with a prompt rather than dropping in raw data; cut phrases that shriek AI, such as I noticed that, I came across your, and I saw on LinkedIn ; and give the model two or three examples of what good looks like.
Read more
Slop is framed as a cost and quality problem, not a matter of taste, and the food analogy makes the point that occasional use is survivable but habit is not.
Rule-by-rule list where each rule is a constraint a team can test against its own live emails.
For brands planning a campaign
This is the most directly reusable material for a workshop: the banned-phrase list, the sentence-level prompting rule, and the token-cleaning step can all be applied to the client's own campaigns during the session. A before-and-after review of the client's current emails would demonstrate the method rather than describe it.
A study of 1.4 million emails sent through Singulate since mid-2024, run as a split inside the same email (same day, subject line, and sender), compared a personalized version against a traditionally personalized one.
Read more
Across 46 campaigns the personalized group clicked 74% more; per thousand recipients that was roughly 21 clicks against 37; personalizing two parts of the email produced 3.7 times more clicks; open rates were nearly identical.
The conclusion is specific: the subject line drove opens, the body drove clicks, and the gap appeared only after people opened and saw the email was about them.
Data is one side of personalization; how you instruct the agent to use the data is the other.
Reports the comparison conditions so the numbers can be weighed, and separates open-rate effects from click effects.
For brands planning a campaign
A brand can evaluate this as evidence before committing budget, and a workshop can propose running the same split test against the client's own list with these figures as the benchmark. The city example knowing someone's city is an opening to cite a local landmark, not to drop in a token gives the session a concrete exercise.
Agentic automation with a human hand on the controls
The argument here is that the blast email is dead but unchecked autopilot personalization is no better, because it produces slop and then creates more work to clean up.
Read more
The proposed middle path is agentic tools that build branded emails quickly while keeping transparency, flexibility, and direct edit control. Studio is described as Claude Design but for marketing operations : build by chatting and commenting, but also edit anything yourself, with the platform choosing how to execute the messaging strategy while the user keeps full control.
Poses the choice as two extremes, then defines the balance as control rather than as a feature list.
For brands planning a campaign
This shifts a workshop from which tool to where does the human review sit. Ask the client which steps in their email workflow they want automated and which they want to keep manual, then design the review process around the answer.
Related posts widen the frame: email marketing is presented as sitting at a fork between generic blasts and unchecked AI personalization, with the middle path as the answer.
Read more
A separate experiment post extends the same thesis beyond email, describing personal assistant apps (Hark, Muse, and others) canceling subscriptions, pausing a lawn service, researching restaurants, and analyzing a LinkedIn network, and arguing that the time it takes to write down a to-do is now about the time it takes to get the whole task done action at the speed of chat.
Treats personal assistants and email agents as the same shift, with a list of completed tasks rather than a product claim.
The assistant-app post is about consumer and prosumer AI tools, not email marketing or Singulate; it shows the thesis in a different domain and should not be read as email evidence.
For brands planning a campaign
Useful as an opening for a workshop aimed at a client still deciding whether AI-era personalization matters at all, before the email-specific material begins.
The Singulate posts double as a running account of the company's own go-to-market: a founding account executive joining, cleanup of HubSpot, Claude automation work, outbound and inbound nurture run through Singulate with HubSpot and Instantly, Heyreach and LinkedIn ads named as channels, attendance at Exit Five Drive and HubSpot Unbound, two marketing videos for a product update, a webinar with Brendon B., and a customer result of a 40% open rate against a usual 2533%.
Read more
Names the tools and channels in the stack rather than describing the motion abstractly.
For brands planning a campaign
A brand can judge whether this hands-on operator experience fits its brief, and the named stack gives the workshop concrete integration questions: does the client use HubSpot, Instantly, or a sales platform that changes the plan?
Both product announcements are written as numbered comparisons, naming the common alternative first and the difference second: most email agents stop at writing and design, most AI tools produce snippets that break at scale, most platforms do nothing with your data.
Read more
The Studio post carries six of these points and ends with the terms free for 14 days, $49 a month.
Six-point comparison format, each point built as most tools do X; this does Y.
For brands planning a campaign
The format is itself a template a brand could reuse for its own launch or comparison content. A workshop can ask whether Dave Schools would run a version of this structure for the client's product.
An event recap from Exit Five Drive carries the argument that sits behind the anti-slop work: when everyone uses the same AI tools, everything converges into a sea of sameness, so now is the time to be weird, remarkable, and intense using humor, shock-and-awe, and raw human storytelling.
Read more
The post credits Ann Handley, Seth Godin, and Dave Gerhardt by name and repeats the coined term Clauderboarding for over-using and over-sharing AI output. Its summary judgment is blunt: playing it safe and predictably will not work in B2B marketing.
Conference takeaways distilled into one message, with the speakers and sessions named so the source of the argument is visible.
For brands planning a campaign
This is the motivational layer for a workshop on personalization and anti-slop, and the strongest material for a client skeptical that differentiation still matters in the AI era.
Dave Schools's retained posts describe one business relationship rather than a set of brand deals.
Read more
He writes as part of the Singulate team, and the same posts show Singulate operating as a HubSpot app partner. Other brands appear only as tools Dave or Singulate use or as event mentions, with no stated commercial relationship, so they are not listed here.
Singulate
Employment
Dave's posts speak from inside Singulate, not about it from outside: "Today, we're announcing Singulate Studio," "we're sending out our own outbound automation," "our GTM team at Unbound," and "I wish I could be there to rep Singulate as well."
Read more
The posts cover product launches, hiring, customer results, and internal tooling, and do not state a title or ownership stake.
Singulate, the company Dave posts for, states it is a HubSpot App Partner and uses HubSpot in its own outbound and nurture stack alongside Instantly.
Read more
Dave lists HubSpot as one of the marketing automation platforms a personalization agent should integrate with. Whether Dave personally holds any agreement with HubSpot is not stated.
Dave Schools's work is easiest to explore by starting with the practical email material and branching out from there.
Read more
The anti-slop post reads as a checklist a team can run against live campaigns, listing the phrases and prompting habits that make personalization sound machine-written, and the reported split test across 1.4 million emails shows the kind of difference the method is meant to produce.
From that point the build post lays out the five decisions behind a personalized email agent, and the product launch posts show where the argument goes when it becomes software: automation that stays editable, rather than a choice between generic blasts and unchecked autopilot.
Two posts sit outside email — a personal experiment with AI assistants that cancel subscriptions and book reservations, and a conference recap arguing that when every team uses the same tools, being distinctive matters more rather than less.
The similar creators below write on adjacent parts of the same shift: how AI changes marketing work, and where human judgment still has to sit.
Both write first-person posts about agents taking over work that people used to do, and both explain what that change means for a B2B audience rather than treating it as a purely technical event.
Read more
Ari Yablok's posts argue that agents do human work without following human rules or boundaries, and that the answer is to extend the same controls used on people to the agents doing the work.
Ari Yablok writes about enterprise governance, monitoring and security for agents, and about the product and brand story built around that idea. Dave Schools stays closer to campaign execution: how personalized email gets written, tested and kept editable inside a marketing team's stack.
Both translate what AI is doing to a specific marketing discipline into practical guidance for the people doing the work, and both use conference takeaways as a teaching format.
Read more
Olivia Ward's posts turn travel-planning behavior and tourism conference sessions into concrete advice for destinations, much as Dave Schools turns email results into rules a marketing team can apply.
Olivia Ward writes for destination and tourism marketing, with tourism economic data, trip-planning trends and trust statistics as the evidence base. Dave Schools works on B2B email personalization and messaging strategy.
The central claim is the same: AI speeds up marketing delivery but still needs a human to brief it, monitor it and check the output.
Read more
Both write short, practical posts built around a concrete example of catching an AI mistake before it reached a live campaign.
Jake Surrey's examples come from paid search builds, analytics and CRO audits, and the argument is about who should make in-platform changes. Dave Schools works on email personalization, using data and messaging strategy rather than paid media accounts.
The only measured audience number available for Dave Schools is a LinkedIn following of about 12,949, observed on 8 October 2026.
Read more
There is no supplied YouTube audience export, no audience survey, and no platform data that breaks the following down by country, age, or gender, so this section can speak to the size and apparent professional interests of the audience but not to its measured composition.
For an agency planning a workshop on AI personalization and anti-slop, that is a meaningful limit: the follower count tells you how many people could see a promotion, not who they are.
What the writing itself indicates is a practitioner audience oriented to email and go-to-market execution.
The most detailed posts are operational walkthroughs aimed at people who run campaigns: a step-by-step for building an AI email personalization agent, a recipe for stripping AI slop out of hyper-personalized email, and a launch post for a self-serve newsletter product pitched at founders, consultants, and agencies with a list to grow.
A split-test writeup reporting click differences across 1.4 million emails and 46 campaigns is written for someone who tests and measures sends, not for a general marketing reader.
One post names the go-to-market team and the topics they field at a conference 2demail 2dscaling ABM, webinar invites, database re-engagement, personalized nurture, and another recaps a B2B marketing community event, including what attendees took away about standing out when everyone uses the same tools.
The through-line is hands-on campaign work rather than, say, brand or creative strategy.
For the workshop brief, that content profile is a reasonable signal that Dave Schools 2ds material lands with marketers who own email and lifecycle campaigns, and the anti-slop angle has an obvious hook in the most instructional posts.
Treat it as appeal, not proof of attendance or seniority.
A concrete next step is to ask for three things before committing: audience survey results covering job function, seniority, and company type; recent results from a comparable session or webinar, with registration-to-attendance numbers; and the topics, formats, and depth that drew the strongest response, rather than the total reach of the account.
If a survey is not available, the buyer should price the engagement on content fit and the creator's own reach, not on an assumed room of decision-makers.
No partnership offer is documented in the material reviewed for Dave Schools.
Read more
His posts promote Singulate, the email marketing platform he is building, and explain its approach to AI-assisted personalization and segmentation; they are product and market commentary rather than a published sponsorship package.
The evidence contains no media kit, no rate card, no stated content formats a brand could buy, and no partnership inquiry route.
That is a gap in what has been reviewed so far, not a finding that he never takes on sponsored work; no such statement appears in his material.
A brand planning a paid partnership should treat the commercial terms as unconfirmed until a direct route and terms are located.
Good fit, on content grounds, for a working session of a few hours. The evidence is unusually specific about what the session would contain.
Read more
Brief
A B2B marketing agency wants Dave Schools to run a paid workshop for its email and lifecycle team on AI personalization and anti-slop strategy. The agency's own team runs campaigns in HubSpot and wants practical method rather than a keynote.
Reasons
The material is already shaped like a workshop. One post sets out five decisions for building a personalization agent — use case, model, data, messaging strategy, and where the snippet goes in the email — plus three checks: test it, make sure it does not read as AI slop, and confirm it integrates with the existing stack, naming HubSpot and Marketo. That is an agenda a team can run against its own live campaigns.
The anti-slop position has concrete rules behind it, not just a preference. The eight-point list includes prompting one or two sentences at a time instead of a whole email, never sending AI-generated messaging on autopilot, writing the exact sentence and swapping only the personalized parts, avoiding giant essay prompts, using ordinary dynamic content where AI adds nothing, cleaning personalization tokens with a prompt instead of dropping in raw data, cutting phrases like "I noticed that…" or "I saw on LinkedIn," and giving the model two or three examples of good output. Each rule converts directly into an exercise on the client's own emails.
There is a test design to borrow. The within-email split test — same day, same subject line, same sender — compared a personalized version against a traditionally personalized one across 46 campaigns, and reports 74% more clicks, roughly 37 clicks per 1,000 recipients against about 21, a 3.7-times lift when two parts of the email were personalized, and nearly identical open rates. A workshop can propose running the same design on the client's list, using the reported figures as a benchmark to beat rather than a promise.
The underlying argument gives the session a why. The Exit Five Drive recap makes the case that when everyone uses the same tools and output converges into a sea of sameness, differentiation comes from humour, human storytelling, and being deliberately off-pattern rather than playing it safe. That answers a skeptical team asking why anti-slop matters commercially.
Evidence limits
The 74% and 3.7-times figures are Dave Schools's own reported measurements for Singulate's own sends, not independently verified benchmarks. Present them as a stated internal result and a test design, not as an expected outcome.
No session record, webinar registration-to-attendance figure, or attendee feedback from a comparable workshop was supplied. Fee and format should be agreed on content fit, with any reach or attendance claim held back until there is evidence for it.
When to choose a specialist
Poor fit on the evidence reviewed, for two separate reasons — one about commercial terms, one about the audience.
Read more
Brief
A B2B sales-engagement vendor wants to buy a paid endorsement campaign: a run of sponsored posts and a short video in which Dave Schools promotes the vendor's AI outreach tool to a defined audience of senior demand-generation buyers, with guaranteed impressions and a performance target attached.
Reasons
No partnership offer is documented in the material reviewed. The posts promote Singulate, the email platform he is building, and explain its approach; none is a published sponsorship package. There is no media kit, rate card, stated format, or inquiry route in the evidence. That is a gap in what has been reviewed, not a finding that he never takes sponsored work — but it means terms, deliverables, and availability are unconfirmed and cannot be planned around.
The audience cannot be checked against the vendor's target. The only measured audience figure is roughly 12,949 LinkedIn followers observed on 8 October 2026, with no breakdown by job function, seniority, company type, or country. A campaign that needs to reach senior demand-generation buyers specifically cannot be underwritten by a follower count, which says how many people could see a promotion, not who they are.
The product being bought sits on the wrong side of the person's own argument. The most consistent thread in the material is that autopilot AI personalization produces slop and then extra cleanup work, and that the fix is controlled, reviewed, sentence-level generation. A campaign asking for high-volume automated outreach promotion would put the endorsement in tension with the position the audience follows him for, which weakens the very credibility being paid for.
The material gives no evidence of past brand-sponsorship work to point to. Nothing in the reviewed posts shows a paid third-party placement, a sponsored series, or a disclosure. A buyer should treat this as unproven on the current evidence rather than assume capability from product-marketing posts.
Evidence limits
This judgment rests on nine LinkedIn posts and one account record. A sponsorship policy or rate card might exist elsewhere; if a direct route and terms are located, the commercial half of this assessment should be revisited. Nothing here should be published as a claim that Dave Schools never accepts sponsorship.
A worked campaign example
Angle
Run the session as an audit, not a lecture: use the client's own live emails as the working material, benchmark them against the reported split test, and leave with a rewritten sequence and a review process the team can keep using after the day ends.
Brief
The agency workshop brief, taken one step further into a costed plan.
Leaves behind
A written review process covering where human sign-off sits in the send workflow, which the team keeps after the session.
The slop-pattern checklist derived from the eight rules, applied to a named set of live campaigns.
A running split-test template with the client's own measure defined before the first send.
Next steps
Ask for audience survey results covering job function, seniority, and company type before pricing the engagement, since the current audience picture is a follower count only.
Ask for a webinar or session attendance record if the agency wants evidence of how the material lands in a room; none was supplied.
Confirm scope from the recorded engagement rather than the follower count, and treat any performance figure quoted from the posts as Singulate's reported internal measurement.
Sessions
Title
Where the plan is missing pieces
What
Work through the five decisions against one live client campaign: the specific use case in scope, the model, the data being personalized against, the messaging strategy that turns that data into something worth reading, and where in the email the personalized part sits. The point of the exercise is to surface decisions the team has been making by default rather than on purpose.
Title
The slop audit
What
Take three current client emails and mark them against the eight rules: whole-email prompts, autopilot sends, unprescriptive instructions, oversized prompts, AI used where segmented dynamic content would do, raw tokens dropped in uncleaned, phrases like "I noticed that…" and "I came across your," and missing examples of good output. Then rewrite one snippet live using the sentence-level method. A before-and-after on the client's own copy demonstrates the method far better than describing it.
Title
Design the test and set the rule
What
Set up the within-email split — same day, subject line, and sender, one personalized cell against one traditionally personalized cell — and agree the measure in advance. The reported benchmark is 74% more clicks across 46 campaigns with near-identical open rates, which is useful as a target. Agree which steps stay automated and which keep a human review, since the argument is that full autopilot creates more cleanup than it saves.
Choose Ari Yablok when the agency wants the workshop framed as governance and oversight of AI work rather than as email mechanics.
Read more
The posts argue that the answer to agents doing human work is not a separate tool that governs agents, but extending the control plane that already governs people to cover agents too, across every work surface, with the same policy and the same monitoring.
That is the right frame for a marketing-operations or ops-leadership audience that has to decide who is accountable for agent output and how it gets reviewed.
The Exit Five Drive write-up adds a broader B2B marketing craft angle — human connection, story, and marketing that does not announce itself as marketing — which is useful for a session that needs to argue why anti-slop matters commercially.
The tradeoff is depth on the email itself.
These posts cover the enterprise control-plane argument and the marketing-craft argument; they do not give a sentence-level writing method, the eight anti-slop rules, model and data selection, or a within-email split test.
A team that needs to leave the room with rewritten live emails and a review process would still have to supply that half, or bring both speakers.
Choose Olivia Ward when the agency's clients sit in travel, tourism, or place marketing and the workshop should be about how destination content gets surfaced by AI and how a brand earns trust there.
Read more
The posts lay out three AI travel trends — hyper-personal itinerary planning, AI used for logistics as well as inspiration, and low trust, with only 6% of users reporting they fully trust AI answers — and give optimization steps against each.
A conference recap adds the position that the job is teaching the robots rather than competing with them, and that tourism data needs context when it is presented to stakeholders.
That maps onto a destination team's actual decisions in a way generic email advice does not.
The tradeoff is channel and audience.
This is not an email personalization or lifecycle workshop: there is no prompting method, no segmentation, no snippet placement, no send-time review process, and no split-test design.
It helps only when the team's own campaigns are travel or destination content and the goal is AI-driven discovery and trust; for HubSpot lifecycle email it will not carry the session.
Choose Jake Surrey when the practical method the team needs is for paid media rather than email.
Read more
The posts give two worked examples — a paid search build where Claude's output contained misaligned negative keywords that would have blocked important keywords, and a large CRO audit combining the GA4 API with Clarity data where a data anomaly was caught midway and would otherwise have distorted the recommendations.
The argument that follows is that AI should not optimize campaigns directly any more than Google or Meta should, and that a human with full account context should make in-platform changes, with trust-but-verify on AI output.
That is the same human-in-the-loop discipline the email anti-slop position rests on, applied to paid search and CRO.
The tradeoff is the channel.
These examples are paid search, GA4, Clarity, and in-platform campaign controls; there is no email copy method, no anti-slop phrase list, and no segmentation or personalization-token guidance.
If the brief is genuinely lifecycle email in HubSpot, these examples will not transfer directly — only the QA principle does.
This profile rests on nine LinkedIn posts published by Dave Schools and one LinkedIn account record observed on 8 October 2026.
All nine posts were read in full as post text: an eight-point anti-slop guide, a split-test writeup covering 46 campaigns, a five-decision checklist for building a personalization agent, two Singulate product launch posts, a post on where email automation sits between generic blasts and unchecked autopilot, a personal experiment with AI assistants, a conference recap, a monthly build update, and a team post.
The supplied posts carry no publication dates, so the work is described by theme rather than in sequence.
The performance figures — 74% more clicks, roughly 37 clicks per 1,000 recipients against about 21, a 3.7-times lift, and a customer's 40% open rate — are Dave Schools's own reported measurements and have no independent confirmation; they are presented as stated internal results.
Views attributed to named conference speakers are theirs as reported.
No material from other channels, no interviews, transcripts, videos, or articles, and no partnership terms or rate card were supplied, so the profile does not establish whether earlier career work or other signature projects explain additional name-search interest.
Recent engagement
Across their last 8 LinkedIn posts, the median post got 26 reactions and 4 comments. Checked 2026-11-02.
