Alfie Marsh is an entrepreneur who presents Station as an AI agent that helps non-technical teams carry out work across apps, such as researching leads, preparing meetings, and drafting outreach.
Alfie Marsh is most useful for promotional demonstrations of AI agent workflows for non-technical teams across apps.
Best for
A B2B SaaS agent platform for non-technical teams wants proof-of-work workflow content.
Not for
AI-detection vendors campaigning against unreviewed AI drafting, because Marsh argues the opposite.
Key takeaways
Station found pet stores, built a lead list, researched personalisation angles, and sent an email campaign.
6am agent briefing checks calendar, researches guests, pulls emails, calls, CRM notes, then posts to Slack.
Marsh says AI slop made LinkedIn better for people with strong ideas and weak writing.
Check first: Before contacting, confirm audience demographics, typical post results, commenter seniority, and sponsorship terms.
Main ideas
You should not need to be an engineer to use AI agents · Agents are personal leverage, best shown as a working routine · AI-generated writing has lowered the barrier to good ideas
Alfie Marsh is an entrepreneur who presents Station as an AI agent for helping non-technical teams get work done across apps.
His posts show what that means in practice: one describes Station finding pet stores, building a lead list, researching personalisation angles and sending an email campaign; another lays out an automated morning briefing that gathers meeting context and sends it to Slack and a phone.
Marsh also argues that AI writing can help more people share good ideas, even if the resulting posts show familiar AI-writing habits.
His feed is not all product explanation: it includes founder-focused encouragement and personal reflections on the people who support a business builder.
Follow his work for practical, promotional demonstrations of agent workflows and a public, optimistic view of AI’s role in knowledge work.
Marsh connects a broad claim—agents can act as digital coworkers—with step-by-step examples of tasks they are meant to handle.
The Station launch post makes the intended audience and workflow concrete, while the meeting-briefing post shows how several information sources and apps fit into one routine.
That makes his writing useful for readers trying to picture where agent tools might fit into everyday work, rather than only hearing abstract claims about AI.
The posts are also promotional, so product capabilities and comparisons such as replacing Chief of Staff work should be read as Marsh’s claims, not independent evaluations.
His separate posts on AI-written content and founder life show a more opinion-led and personal register alongside the product demonstrations.
The main one argues that AI agents are becoming a practical form of digital labor that non-technical teams can direct in plain language, and that the payoff is personal leverage: research, lead generation, meeting prep and outreach handled while you do something else.
The second track is more opinionated and less commercial.
Marsh argues that AI-written "slop" has actually improved LinkedIn by letting people with good ideas publish them, and he writes about founder discomfort and the supporters behind people who succeed.
Across the posts, the through-line is access: removing the engineering barrier to agents, and removing the writing barrier to ideas.
You should not need to be an engineer to use AI agents
Marsh's central argument is that agent tools have been built for developers and that the real opportunity is everyone else.
Read more
Announcing Station 2.0, he describes it as a "General Super Agent for Knowledge Work," placed beside developer-facing tools such as OpenClaw and Claude Code but pitched as "stupidly simple to use and specifically built for non-technical teams."
He frames agents as "a new form of digital labor" that takes action and executes tasks end to end, and says the user simply describes the task and the agent works out how to do it.
He closes the argument with the line "You shouldn't have to be an engineer to leverage agentic labor."
The stated capabilities are executing tasks across 2,500+ apps, searching for the skills, tools and APIs a task needs, scheduling background tasks, triggering workflows from app events, and reaching the user by email, SMS and WhatsApp.
The demonstration Marsh gives is deliberately small-business rather than enterprise: his wife asked Station to help sell her handmade dog bow ties. In a couple of minutes it found pet stores in the Bay Area, enriched a lead list in Google Sheets, researched their Instagram accounts for personalization hooks, drafted an email campaign and sent it from Gmail.
Agents are personal leverage, best shown as a working routine
Where the launch post states capability, Marsh's day-in-the-life posts argue by walkthrough.
Read more
The claim is that agents quietly do the work of expensive human support, and that setting one up is trivial.
In the strongest example he describes an agent running at 6am that checks Google Calendar for external meetings, researches every guest, pulls email, call recordings and CRM notes, researches LinkedIn, company news, social media and podcasts, writes an executive brief and agenda into a Google Doc, posts it to a Slack channel and sends an SMS.
He says it took about 60 seconds to set up and was free, and compares the output to work that "used to cost $60k/year and hiring a Chief of Staff," calling agents "a cheat code for life."
The launch post repeats the pattern with a list of everyday uses: daily executive briefings, meeting intelligence prep, company research, SEO and content creation, viral content ideas, running a CRM on autopilot, and scoring and researching inbound leads.
The 6am executive briefing is the clearest instance: a named, step-by-step pipeline with a stated setup time and a cost comparison against hiring a chief of staff.
The earlier framing is product capability; the later posts turn that into a documented personal routine, moving from what the agent can do to what it does before Marsh wakes up.
AI-generated writing has lowered the barrier to good ideas
Marsh takes a deliberately contrarian position on AI content.
Read more
The argument is that LinkedIn's old problem was not a shortage of interesting people but that good thinking arrived badly written: "Great idea.
Horrible execution."
AI flipped that, so now "bad writers turn great ideas into good posts."
Marsh concedes the familiar tells of machine writing, naming too many em dashes and phrases like "here's the kicker," and still concludes that the trade-off is worth it because people who previously would not post now do.
The post is framed as an invitation to argue, closing with "Prove me wrong."
The examples are the tells themselves: overused em dashes and stock phrases such as "here's the kicker," offered as the admitted cost of a lower barrier to entry.
This sits in tension with Marsh's own product posts, which use the same short-line, bulleted LinkedIn format the argument is defending. Marsh is not arguing AI writing is stylistically better, only that the ideas now get through.
Growth happens outside comfort, and success has unseen supporters
Away from product content, Marsh writes founder-to-founder about discomfort and gratitude.
Read more
One post admits that taking "silly selfies for strangers on the internet" sometimes feels idiotic, then argues that this discomfort is the job: entrepreneurship forces you to do things while others look and laugh, and that willingness becomes a competitive advantage. A second post turns the same attention outward, arguing that society celebrates the people who succeed but rarely the people who made it possible.
Marsh thanks his wife Adriane, in their tenth year together, for listening to hours of pitch practice, reviewing monthly investor updates and consoling Marsh after failed investor meetings. The post widens into a general thank-you to the friends, colleagues and family who support people who are "daring to be great," and supplies the fundraising context the product posts leave out.
This is the clearest single statement of what Marsh builds and who it is for.
Read more
If you only open one post, open this one: it names the product, the positioning, and the audience in the first three lines.
Marsh announces Station 2.0 as a "General Super Agent for Knowledge Work," explicitly placed next to developer-oriented tools (OpenClaw, Claude Code) but described as "stupidly simple to use and specifically built for non-technical teams."
The post lists capabilities: executing tasks across 2,500+ apps, searching for tools and APIs needed to finish a task, scheduling recurring background work, triggering workflows from app events, and communicating by email, SMS and WhatsApp.
The concrete demonstration is a small-business one, not an enterprise pilot: Marsh's wife asked Station to find pet stores in the Bay Area for her handmade dog bow ties, and within a couple of minutes it produced a lead list in Google Sheets, researched Instagram accounts for personalisation angles, drafted an email campaign and sent it from Gmail.
Marsh also lists personal uses including daily executive briefings, meeting prep, company research, SEO and content creation, CRM management, and scoring inbound leads.
For brands planning a campaign
For a B2B SaaS company weighing Marsh as a practitioner voice, this is the post that shows both product-marketing craft and a habit of proving the tool with a visible, checkable workflow.
A collaboration could reasonably ask for a documented walkthrough of one workflow with its inputs and outputs, rather than a generic endorsement.
Note that the capability list is a product claim made by the person promoting the product, not an independent test.
It answers the practical question a buyer asks first: what does an agent actually do in a working day?
Read more
The post is a step-by-step workflow, not a theme.
Marsh describes an agent that runs every morning at 6am: it checks Google Calendar for external meetings, researches every guest, pulls company history from emails, call recordings and CRM notes, researches LinkedIn, company news, social media and podcasts, writes an executive brief and meeting agenda into a Google Doc, posts it to a Slack channel, and sends an SMS. Marsh frames this as replacing the kind of work that "used to cost $60k/year and hiring a Chief of Staff," describes setup as taking about 60 seconds, says it was free, and closes by asking readers to comment "STATION" to receive a link by direct message.
For brands planning a campaign
The workflow maps directly onto revenue and sales-operations pain, which makes it the strongest post for a meeting-intelligence or CRM integration brief.
Two things a brand should plan around: the cost-replacement comparison is a promotional claim rather than a measured case study, and the comment-to-DM close is an engagement mechanic that funnels attention into a private channel.
That is useful if the campaign is designed around comment-driven demand capture, and a mismatch if the brand wants neutral evaluation content.
It tells you where Marsh stands on the most contested topic in the AI content market, in Marsh's own words.
Read more
Read it before deciding whether the voice fits an AI or content brief.
Marsh argues that LinkedIn's problem used to be well-informed people writing badly; AI has flipped that so "bad writers turn great ideas into good posts."
The post concedes the tell-tale signs of AI writing, naming too many em dashes and phrases like "here's the kicker," then argues the trade-off is worth it because people who previously would not post now do.
It closes with "Prove me wrong," inviting disagreement in the comments.
For brands planning a campaign
A brand running an AI-tooling campaign gets a writer who will defend AI-generated content publicly, which can be an asset for a debate-style campaign and a liability for a brief that promises human-crafted editorial quality. The post also functions as an existing conversation starter: replies to it are a ready-made audience of people who already argue about AI content quality.
It shows the tone Marsh uses when not selling: founder-to-founder encouragement, first-person, short lines, one idea per paragraph.
Read more
Useful if the intended collaboration is a founder-journey narrative rather than a product demo.
Marsh describes being a 34-year-old entrepreneur who takes "silly selfies for strangers on the internet" and admits sometimes feeling like an idiot, then argues discomfort is what entrepreneurship requires.
The stated lesson is that "growth happens on the other side of comfort" and that the willingness to work outside it becomes a competitive advantage.
The post is framed as advice for the coming year and closes with an exhortation not to let comfort become a prison.
For brands planning a campaign
This establishes a motivational register that suits founder-community, resilience and personal-brand briefs, but carries no product detail and is a poor vehicle for technical evaluation. It also shows the posting habits a partner would inherit: selfies, first-person vulnerability, aphorism-based closers.
It supplies the fundraising context that the product posts leave out, and shows how Marsh handles personal material in a professional feed.
Read more
Marsh thanks his wife Adriane, in their tenth year together, describing how she listened to hours of pitch practice, reviewed monthly investor updates and consoled Marsh after failed investor meetings. The post widens into a general thank-you to the friends, colleagues and family members who support people who are "daring to be great," noting that society celebrates the people who succeed but rarely the people behind them.
For brands planning a campaign
The fundraising references show an early-stage context that is useful background when judging which business audiences the voice can credibly reach.
This is a weak fit for a commercial integration and a strong fit only for a human-interest or founder-support narrative.
Treat the investor meetings as described personal experience, not as evidence of named backers or commercial relationships.
Alfie Marsh launched Station 2.0 as a "General Super Agent for Knowledge Work," describing it as similar in capability to developer-facing agent tools but "stupidly simple to use and specifically built for non-technical teams."
Read more
The launch post frames agents as "a new form of digital labor" that takes action and runs tasks end to end, and lists practical capabilities: executing tasks across 2,500+ apps, searching skills, tools and APIs, scheduling recurring background tasks, triggering workflows from app events, and messaging the user by email, SMS and WhatsApp.
Product launch post built on a contrast with engineering-oriented tools, followed by a capability list and named personal use cases.
For brands planning a campaign
A SaaS or agent-tooling brand gets a practitioner who speaks directly to the buyer question of whether non-technical staff can actually run agents. The strongest angle is a demo-led or comparison format that tests the accessibility claim rather than a general awareness post.
Walkthroughs of agent workflows run by the creator
Marsh documents a daily executive briefing built on an agent that runs at 6am: checking Google Calendar for external meetings, researching guests, pulling email, call recordings and CRM notes, researching LinkedIn, company news, social media and podcasts, writing a brief and agenda into a Google Doc, posting it to Slack and sending an SMS.
Read more
Marsh says the setup took 60 seconds and compares the output to work that would otherwise require a chief of staff. The launch post adds comparable examples, including a lead-generation run for a handmade pet bow-tie shop that found Bay Area pet stores, enriched a lead list in Google Sheets, researched Instagram accounts for personalization and sent campaigns from Gmail.
Numbered, step-by-step accounts of a personal workflow, ending with a concrete time or cost comparison.
For brands planning a campaign
This is the most directly usable material for a campaign that needs to show an agent doing real work. A collaborator brief can ask for a documented walkthrough of one specific pipeline with inputs, steps and output, which suits tools that need proof rather than positioning.
Beyond the flagship examples, Marsh names recurring applications: daily executive briefings, meeting intelligence preparation, company research, SEO and content creation, viral content ideas, running a CRM on autopilot, and scoring and researching inbound leads.
Read more
The pattern is operational and personal rather than enterprise-transformation framing — agents described as a "cheat code" and an "agentic co-worker" that the user simply describes a task to.
For brands planning a campaign
Useful for content marketing, sales-ops and CRM vendors whose products appear inside these workflows. The implication for a brief is to anchor the collaboration in one named use case with a verifiable before-and-after, not a broad list of possibilities.
Marsh argues that AI-generated "slop" has made LinkedIn better, because people with strong ideas and weak writing can now publish readable posts.
Read more
The post concedes the familiar tells — too many em dashes, overused phrases — but argues the trade-off is worth it because it lowered the barrier to entry for good ideas.
Opinion post with a stated counterintuitive thesis, an acknowledged objection, and a direct challenge to readers.
This post carries no product tie-in and is not about Station.
For brands planning a campaign
Signals a contributor willing to take a contrarian public position on AI content, which suits writing, content-ops and publishing tools. A brand should expect a polarizing stance and decide deliberately whether that disagreement is on-message; it is a poor fit for vendors whose positioning depends on AI-detection or anti-AI-drafting claims.
One reflects on being a 34-year-old first-time founder doing uncomfortable public posting, with the argument that growth happens outside comfort. The other is a tribute to Marsh's wife Adriane, describing ten years together and her role listening to pitches, reviewing monthly investor updates and pushing for lunch breaks and sleep, framed as recognition for the unsung supporters behind people who succeed.
Personal narrative structured around a single takeaway line and a direct thank-you to the reader's own supporters.
Neither post discusses agent tooling, and neither should be read as evidence of commercial partnerships.
For brands planning a campaign
Relevant for founder-brand and employer-brand work, and useful context for judging tone, but it does not support an AI-agent-tooling brief. A brand evaluating Marsh as a practitioner voice should separate these product-free posts from the tooling content rather than treating them as category evidence.
The retained posts document one brand connection: Station, the AI agent product Marsh announces and promotes.
Read more
The material does not state Marsh's role in the company or whether the relationship is commercial, so it is recorded as a disclosed partnership rather than ownership or employment.
Station
Disclosed partnership
Marsh announced Station 2.0, described as a general super agent for knowledge work that runs tasks across 2,500+ apps, and invited readers to create a free account at GetStation.ai.
Read more
In a separate post Marsh described setting up an agent that compiles a 6am executive briefing from calendar, email, call recordings, CRM notes and web research, then offered a DM link to the same product. Marsh writes about Station as an insider, but no reviewed post states whether Marsh owns, works for, or is otherwise contracted by the company, so the relationship is unstated.
Alfie Marsh writes about AI agents as everyday labor for people who are not engineers.
Read more
The clearest entry point is the Station 2.0 launch post, where Marsh describes a "general super agent" for knowledge work that runs tasks across thousands of apps and is meant to be simple enough for non-technical teams.
From there, the workflow posts are the most concrete: a step-by-step account of an agent that prepares an executive briefing at 6am, gathering calendar, email, call and CRM context before the day starts, and a lead-generation run that found local pet stores, built an enriched list and drafted an email campaign.
If you want the argument rather than the product, Marsh also writes contrarian opinion pieces, including the case that AI-written posts improved LinkedIn by letting people with strong ideas publish them despite weak writing.
A smaller set of founder-reflection posts, about discomfort, failure and the people who support a founder, shows a different side.
Read the launch post and the briefing walkthrough first: together they set out the capability claim and the working proof.
The topic tags below are the fastest way to move through the corpus.
Kirkpatrick also writes about removing manual steps from client-facing business processes with automation, and does it the same way: naming the tools, listing the exact handoffs and showing what runs automatically.
Read more
One post breaks down a client onboarding chain across CRM, drive folders, project boards, email and Slack, and counts the manual steps it replaces; another lists buildable AI workflows for advisory practices, from document filing to meeting prep and follow-up drafting.
Kirkpatrick writes from an agency and consulting position, describing workflows built for clients and inviting a discovery call, and draws on earlier wealth-management experience. Marsh's posts centre on his own agent product and on his personal use of it, so the reader gets a founder demonstrating a tool rather than a builder selling an implementation service.
Seshan covers the same shift Marsh writes about: agents moving from assistants that surface information to teammates that research, prioritise prospects, draft outreach and take action inside a seller's existing tools.
Read more
Both explain autonomous sales and research work by naming the agent and listing what it does, and both tie that to how selling time should be spent.
Seshan speaks for a large enterprise platform, describing cloud-scale products and keynote announcements aimed at sales organisations already inside that ecosystem. Marsh writes about a small startup tool aimed at non-technical teams and illustrates it with personal errands, such as a family shop's outreach run, rather than enterprise deployments.
Mirich shares Marsh's interest in everyday automation and writes it as a short, teachable recipe: name the repetitive task, list the trigger, the automated action and the outcome, then ask the reader what they would automate.
Read more
The examples are deliberately ordinary, covering follow-up reminders and personalised video replies that would otherwise be done by memory.
Mirich stays tool-level and brand-neutral, explaining how to wire up an existing platform (Zapier) for small wins, with none of the agentic AI framing, product launch material or founder narrative that runs through Marsh's posts. The practical effect is a how-to voice rather than a practitioner building and promoting an agent product.
Alfie Marsh's measurable audience is a LinkedIn following of 33,547 at the last profile observation on 8 October 2026.
Read more
That is a reach figure and nothing more: the reviewed material contains no demographics, no engagement numbers, and no information about the roles, employers, or seniority of the people following.
A buyer can use it to judge scale, not to conclude the audience is made up of operations leads, founders, or any other specific group.
The five reviewed posts show what that following is being asked to pay attention to.
Selected examples: the Station 2.0 launch, framed as a general agent for knowledge work built "specifically for non-technical teams," with a worked example of a small handmade-goods shop finding pet stores, building a lead list, and sending outreach; a 6am agent-generated executive briefing that researches meeting guests from calendar, email, call recordings, and CRM notes; an argument that AI-assisted writing has improved LinkedIn by letting people with good ideas who write badly post anyway; and two founder-personal posts, one on growth happening outside comfort zones and one thanking his wife for years of support through failed investor meetings.
For a B2B SaaS company weighing Marsh as a practitioner voice on AI agents for non-technical teams, the product and workflow posts match the brief's subject and framing, while the personal posts show the audience also receives founder-story content.
What the evidence does not settle is whether those followers are buyers in that market or a general founder-and-creator following.
There is no way to tell from this material whether the reach converts into attention from the people a B2B SaaS campaign needs to reach.
Treat the follower count as scale, and treat audience relevance as unverified.
Concrete next step: ask Marsh for any audience survey results or platform demographic export, plus typical results on posts covering agent workflows for non-technical users — impressions, comments, and, if available, the seniority or function of commenters.
The demographics export would answer geography, age, and gender; it would not answer employer, occupation, or job seniority, so ask specifically whether a survey or other documented source covers those professional attributes for this brief.
Alfie Marsh's reviewed LinkedIn posts contain no published partnership or sponsorship offer.
Read more
The commercial activity they document is promotion of Station, the AI agent product Marsh is launching and using, including a free-account signup route and a post that invites readers to comment for a direct message with a link.
No rate card, media kit, sponsored-post example, disclosure practice, audience or format specification, or inquiry address appears in the material reviewed, so nothing in this section can state what a paid collaboration would include, cost, or require.
The practical consequence for a buyer's plan is that any Station-related collaboration would currently have to start from a direct, individual approach rather than a published process, and that Marsh's public emphasis is on product adoption and agent workflows for non-technical teams rather than on creator-side commercial terms.
Because no official partnership route is verified, no inquiry link is given here.
The actionable gap is an unpublished partnership policy: formats, deliverables, pricing basis, disclosure and approval requirements, and a contact route all still need to be established directly with Marsh before a plan can be costed or scheduled.
Take it. This is the brief Alfie Marsh's material is already built for, and the collaboration should be demo-led rather than an awareness post.
Read more
Brief
A B2B SaaS company selling an agent platform for non-technical teams wants a practitioner voice to produce proof-of-work content: a documented workflow in which someone without an engineering background completes a real task end to end, with the vendor's product inside it.
Reasons
Marsh's central product claim is aimed at exactly this buyer.
The Station 2.0 launch describes a general agent for knowledge work as similar in capability to developer-facing tools but "stupidly simple to use and specifically built for non-technical teams," and frames agents as digital labour that takes action and finishes tasks rather than assisting with them.
A vendor whose product sits in that category is not borrowing credibility from an unrelated creator; it is hiring someone already speaking to the same question.
The corpus contains the format the brief needs.
Marsh's meeting-briefing post runs a single workflow step by step: at 6am an agent checks the calendar for external meetings, researches each guest, pulls emails, call recordings and CRM notes, writes a brief and agenda into a Google Doc, posts it to Slack and sends an SMS.
The launch post adds a small-business example where a handmade dog bow-tie shop found Bay Area pet stores, built an enriched lead list in Google Sheets, researched Instagram accounts for personalisation and sent campaigns from Gmail.
Both are inspectable sequences with named inputs and outputs, which is the raw material for a walkthrough, not just a theme.
One qualification sits on the vendor side of the table.
The Station posts are product marketing written by the person promoting the product, so the capability list is a claim, not an independent test.
A collaboration should ask for a documented before-and-after the vendor can stand behind, and the contract should be clear that the deliverable is a walkthrough, not a measured case study unless measurements are supplied.
There is also no published partnership route for Marsh, so scope, deliverables, pricing basis, disclosure and approval all have to be agreed directly and in writing before anything is scheduled.
When to choose a specialist
Decline. The brief asks Marsh to argue against a position Marsh has already taken in public.
Read more
Brief
A vendor selling AI-detection and human-authorship verification wants Marsh to make the public case that unreviewed AI drafting is degrading professional content, and to front a campaign positioning its product as the quality safeguard.
Reasons
Marsh's post on AI and content quality concedes the familiar tells of AI writing — too many em dashes, overused phrases — and then argues the trade-off is worth it, because people with strong ideas and weak writing can now publish readable posts.
The headline position is that AI slop has made LinkedIn better.
A detection vendor's campaign needs the opposite conclusion, and a creator who has publicly disagreed with it will either be asked to reverse a stated view or to soften it until the endorsement means nothing.
Both outcomes damage the brand and the creator.
The mismatch is visible to the audience.
Marsh's post closes by inviting disagreement, so the people who read and reply are already arguing about AI content quality.
A campaign built on the assumption that AI-assisted drafting is the problem would arrive in front of a comment thread that has been having that argument for months.
A second brief fails for different reasons and is worth recording here.
A vendor needing enterprise procurement sign-off — security review, integration certification, independent performance evidence, named customer references — cannot be served by this corpus.
The evidence supplied is short-form LinkedIn posts plus a follower count of 33,547 observed on 8 October 2026, with no demographics, no engagement figures and no independent testing of any product claim.
Reach is documented; the professional composition of that reach is not, so nobody can yet say the audience contains the security or platform buyers such a deal needs.
For either brief, the practical obstacle is the same: no rate card, media kit, deliverable specification, disclosure practice or inquiry address appears in the reviewed material. Any rejection should be followed by a direct enquiry if the brand still wants to work with Marsh on a different angle, rather than a silent no.
A worked campaign example
Angle
"One workflow, run by someone who is not an engineer." Instead of a list of possible uses, each piece documents one task from start to finish and ends with the same instruction: here is what to ask your own agent, and what a good output looks like.
Brief
Agent platform vendor, content collaboration for the non-technical-teams brief above.
Deliverables
Three posts, one workflow each, written in Marsh's observed first-person, numbered-step style: a lead-generation run (the pet-store example shows the shape), a meeting-preparation run (the 6am briefing), and one operational run such as inbound lead scoring or CRM upkeep. Each post carries a companion screen recording of two to three minutes showing the setup, so a reader can see how long it actually takes rather than trusting a claim. Each post ends with one concrete prompt the reader can paste, and a single route to the product.
Open gaps
Before the collaboration is costed, ask Marsh directly for partnership terms — formats, deliverables, pricing basis, disclosure and approval requirements, and a contact route — none of which is published. Before promising reach to the vendor, ask for any audience survey or platform demographic export, plus typical impressions and commenter seniority on posts about agent workflows for non-technical users; the follower count establishes scale, not audience relevance. If the first deliverables are to be presented as results rather than demonstrations, ask for the measured before-and-after behind them, because the existing posts report Marsh's own claims about the product rather than independent testing.
Success test
A reader who has never configured an agent can copy the prompt, run the workflow and get a usable output without contacting support. Measure it two ways: the vendor's form submissions and set-up completions, and a small test in which three non-technical staff outside the company follow one post and report where they got stuck. If the testers stall at the same step, the post needs that step written out, not more promotion.
Why this shape
The evidence already shows Marsh works best when the post is a sequence with visible inputs and outputs. The two strongest posts in the corpus are exactly that: the 6am briefing and the found-leads-and-email campaign. The capability list and the "2,500+ apps" claim are weaker because they can only be taken on trust. Building the campaign on the demonstrated shape plays to what is proven and gives the vendor something buyers can repeat.
The vendor should also settle the demand-capture mechanic deliberately. Marsh's own close asks readers to comment for a direct message with a link, which funnels attention into a private channel and suits comment-driven capture. A vendor that needs trackable pipeline should replace it with a landing page the vendor owns, while keeping the conversational close.
Choose Brady Kirkpatrick when the vendor needs the collaboration to end in a built automation inside the client's own stack, not a demonstrated run.
Read more
The retained posts show a builder's sequence: one walks a new client from signed contract or payment through a seven-step handoff — CRM record moved to active, Google Drive folders created with permissions, project board spun up from a template, personalised welcome email sent, Slack notification, kickoff booked — and another lists four automations to add to an advisory practice this month, including document filing, post-meeting recaps and life-event triggers.
Kirkpatrick also writes from a specific vertical, stating a past in wealth management and describing the 30–45 minutes per client that meeting prep consumes there.
A vendor selling into financial advisory, accounting or professional services gets a credible person to demonstrate a working build in that exact setting, and Kirkpatrick publishes a discovery-call route, so the commercial conversation has an obvious entry point.
The posts are an agency offer, not an independent product test.
Kirkpatrick closes one with a Notion document and "if you want my team to build this for your business" plus a discovery-call link, so the content reads as a services pitch and the buyer is buying hours as much as an endorsement.
Nothing in the retained material addresses agents for non-technical teams or an agent platform; the automation logic is workflow plumbing across CRM, Drive, project boards and Slack.
There is also no documented before-and-after measured on a client, and no dates or figures beyond the two posts, so a campaign promising results would be over-claiming.
If the vendor's product is a general agent that non-technical staff are meant to run themselves, Kirkpatrick's build-for-you framing answers a different question than the one the campaign is asking.
Choose Arvind Seshan when the brief needs enterprise CRM credibility and a selling-organisation audience rather than a self-serve operator.
Read more
The retained posts come from a platform keynote context: one introduces three pieces of a sales vision — an agent that turns calls, emails, texts and meetings into durable context, an outbound sales agent that researches and prioritises prospects and personalises outreach, and CRM capability delivered inside a third-party assistant for meeting prep, pipeline review and governed actions.
The other is a focused piece on that outbound agent learning how a team already sells and applying it across the workflow.
A vendor that needs to be seen alongside a large established platform, or that sells to sales leadership at companies with existing CRM investment, gets someone who can speak to the enterprise version of the question: how agents change the shape of a revenue organisation.
The material is platform announcement writing, so a collaboration lands as corporate product marketing, not a practitioner's proof-of-work.
The vision is aimed at sellers inside large organisations, the opposite of the non-technical-teams buyer in this brief, and there is no walkthrough a reader could copy or run — no inputs, outputs, prompts or setup shown.
Claims such as the outbound agent learning a team's winning motion are stated as product direction with a keynote recording linked, not as a documented test, so credibility transfers from the platform rather than from demonstrated use.
Two posts also cannot establish a body of independent work, and no audience composition, engagement or commercial terms are published, so reach and buyer relevance remain unknown.
Choose Cooper Mirich when the campaign wants an accessible, immediately copyable automation audience rather than a buyer evaluating an agent platform.
Read more
The retained posts work at recipe level: one sets out a three-step flow that adds a follow-up reminder to a calendar or task manager when a CRM action completes, and the other explains how to use a form submission to trigger a templated personalised video sent through a video tool.
Both are single-tool, low-cost ideas an operations generalist can set up without engineering help, and both are written to prompt reader replies and shares.
For a vendor whose product is a lightweight automation tool and whose buyer is an SMB or ops generalist already comfortable wiring apps together, this is a closer audience fit than an agent-platform narrative.
The content is generic productivity-hack writing, and the automation layer is the tool the post names, not an autonomous agent completing a task end to end.
The scope is deliberately small — reminders and video follow-ups — so it cannot support a claim that the audience evaluates or buys agentic labour.
The posts also predate the rest of the comparison material by more than a year, and nothing in the retained work shows current engagement with agent platforms, so using Mirich to speak to the 2026 agent-tooling question would be a stretch the evidence does not support.
As with the other comparators, no engagement figures, audience composition, rates or deliverable terms are published, so a vendor would have to establish those directly before committing budget.
This page draws on five LinkedIn posts read as full text: the Station 2.0 launch, an agent-run morning briefing, an argument that AI writing has improved LinkedIn, a founder reflection and a personal tribute to Marsh's wife.
Two of the five carry dates, 2 and 3 March 2026; the other three came without publication dates, so nothing here is presented as a timeline of Marsh's output.
The account record behind the follower figure was observed on 8 October 2026.
Gaps that could change a reader's decision are flagged in the relevant sections: no transcripts, no long-form writing and no independent testing of the product claims, which are reported as Marsh's own.
Partnership terms, audience demographics and engagement figures were also not part of the evidence.
Similar-creator comparisons rest on two full posts each.
Recent engagement
Across their last 6 LinkedIn posts, the median post got 39 reactions and 15 comments. Checked 2026-11-02.
