Othmane Khadri builds AI-assisted go-to-market workflows for B2B sales and marketing, sharing practical examples of outbound campaigns, CRM follow-up and deal recovery. Khadri describes this work through the agency Earleads and the GTM operating system Yalc.ai.
Fully automated outbound platforms promising SDR replacement and no human review do not fit.
Key takeaways
A service sells taste, context, execution, accountability, and outside feedback that an agent cannot.
Lost deals should be reworked monthly and reopened with the specific reason they died.
Judge outbound campaigns by good calls per 100 leads and watch for ICP drift.
Check first: Confirm rate card, audience demographics, and engagement on lost-deal recovery posts.
Main ideas
A service sells what an agent cannot: taste, context, execution, accountability, and outside feedback · AI is strong at deterministic, math-like work, and the last part of a job stays human · Lost deals should be reworked on a monthly cadence and reopened with the specific reason they died
Othmane Khadri shares practical examples of using AI in B2B sales and marketing, especially outbound targeting, CRM follow-up, and recovering lost deals.
The work pairs concrete workflows with a recurring argument: AI can produce useful outputs, but people still need to supply judgment, context, execution, and responsibility for results.
In one post, Khadri describes an agent that reviews lost deals, identifies recurring objections, and drafts tailored follow-up; another lays out five skills for teams using lemlist and Claap.
Khadri also publishes more provocative work, including a satirical agent aimed at people posting AI-generated comments.
Readers interested in applied AI for go-to-market work can follow these examples for both the workflows and the arguments about where human judgment matters.
Khadri makes abstract claims about AI concrete by showing the work around the model: which CRM fields and call notes an agent uses, how its findings are applied to open deals, and what a human does with the resulting draft.
The lost-deal post is a clear example: it describes reviewing closed-lost records, grouping objections, spotting similar patterns in open deals, and writing a personal re-engagement message.
In another post, Khadri argues that taste, private context, execution, accountability, and feedback from the market distinguish a paid service from access to an AI model.
That claim gives readers a useful lens for assessing automation, although it is Khadri's argument rather than independently demonstrated proof.
The tone is not uniformly instructional: the satirical agent targeting AI commenters shows a deliberately provocative side, while a short content post offers an opinion rather than a documented workflow.
A testimonial for an accounting firm appears in the reviewed posts but does not establish Khadri's own expertise in accounting.
The supplied material does not establish whether books, shows, or earlier signature projects are important to Khadri's public identity; targeted research into those name-search results is needed before treating this as a complete account of what Khadri is known for.
Khadri's posts return to a small set of arguments rather than a broad range of topics.
Read more
The central one is that when a buyer can run the same AI models, a paid service has to sell taste, hard-won context, the last mile of execution, human accountability, and an outside feedback loop rather than the deliverable itself. Around that sit practical positions: revise lost deals on a recurring cadence and reopen them with a specific question, judge outbound by good calls per 100 leads instead of reply rate, lead content with the best achievable outcome instead of efficiency, and release agents and skill libraries publicly as part of the work rather than as marketing attached to it.
A service sells what an agent cannot: taste, context, execution, accountability, and outside feedback
Khadri argues that once a client can point their own agents at the same models and sources, the deliverable stops being the product.
Read more
What is left to sell is judgment about which outputs to discard, private context that was never published online, the last stretch of work after the deliverable is handed over, a human who owns the outcome, and a loop in which real market results rather than self-grading improve the work. Khadri treats a model's willingness to ship average work as the weakness to sell against.
In an account of a six-figure upsell that nearly collapsed, a client asks: "Why would i buy your service knowing you'll use agents if i can just use my agents and ask them to do what i'll pay you to do."
Khadri answers with the five points and writes that the client said yes before the call ended.
The same post names the agency Earleads as the source of context and Yalc.ai as where the market-graded feedback loop lives.
AI is strong at deterministic, math-like work, and the last part of a job stays human
Khadri's underlying view is that models excel where the problem can be reduced to probability and math, which is why coding gets absorbed fastest, while a layer of taste, judgment, or consciousness is not close to solved.
Read more
In the sales version of the argument, roughly the first slice of a job gets automated and whatever sits between that and the paid outcome is owned by a person. Khadri cites Roy Amara, former president of the Institute for the Future: people overestimate a technology in the short run and underestimate it in the long run.
Khadri traces the position to a 2022 thesis on how human bias drives financial markets, with an added AI section that found the models carried similar biases. The later account of why a client should pay for human work puts the same idea in commercial terms: the deliverable is perhaps 20 to 30 percent of the work, and the remaining 70 percent is the painful, human, last-mile portion the buyer actually feels.
The position starts as an observation about AI spreading and hardens into a boundary claim. In 2022 the takeaway is that the technology would compound; by the later posts it is that compounding does not remove the need for judgment, and that shipping the average output is a choice a service can refuse.
Lost deals should be reworked on a monthly cadence and reopened with the specific reason they died
Khadri treats closed-lost deals as a standing backlog rather than a write-off, and argues the reopening message works when it names the concrete worry that stalled the deal instead of asking generically for another chance.
Read more
The workflow Khadri describes pairs that human habit with an agent that reads the record and clusters the objections, so the ask going out is informed by patterns across many losses rather than one memory.
A Yalc.ai agent run through Claude Code pulls every Closed Lost deal from Attio over twelve months, reads the lost-reason field, the last three activities before the stage changed, and attached notes or transcripts, clusters objections against buyer attributes and the stage where deals died, scores open deals that match a losing pattern, and drafts the follow-up.
The post quotes the message sent to one prospect, which asks whether the deal was lost because a head of sales and a CRO were never reconciled in the same room, and closes with a phone number.
Khadri reports the prospect replied: "You're the first company I turned down who came back asking something like this."
Judge outbound campaigns by good calls per 100 leads, and watch for ICP drift list by list
Khadri argues reply rate flatters campaigns that produce conversations without fit, so the more useful measure is how many good calls a campaign yields per hundred leads loaded.
Read more
A related warning is that targeting drifts one list at a time, as job titles quietly enter an audience nobody decided to include, so the drift is invisible unless something checks for it. Closed deals become a test of whether the rest of a campaign shares the buyer's context.
Khadri published five skills to run on Claude with lemlist and Claap connected: scoring every lead across the last five campaigns against the ideal customer profile to surface creeping job titles; following replies into booked calls to rank campaigns by good calls per 100 leads; checking a closed deal's buyer context against the rest of its campaign and, where it does not match, writing new list filters; drafting a short pre-call note from a similar company's past call, without sending it; and turning the exact words prospects used in calls that progressed into five opening lines. The library is described as open source and free, and as part of the Yalc.ai GTM operating system.
Content should be led by the best achievable outcome, not by efficiency
Khadri argues against making "how do I do this as efficiently as possible?" the guiding question for content, calling content the most asymmetrical-return marketing channel available and urging the best available outcome as the standard instead.
Read more
The claim is presented as an opinion about how the channel should be run rather than as a documented program.
A short post aimed at people who lead their content with the efficiency question: Khadri asks them not to, states the asymmetric-return claim, and proposes leading with the best outcome instead.
Building in public is part of the work, including open-source agents
Across the posts, agents and skill libraries are released openly as they are built, tied to Yalc.ai as a GTM operating system or harness and to the Earleads agency.
Read more
Khadri presents the release as the work itself rather than a marketing layer on top of it, and one release is deliberately provocative: an agent built to waste the time of the automated commenters Khadri had argued were bad for business.
After an earlier post arguing that AI-generated comments damage businesses reached 200,000 views, Khadri built and open-sourced an agent called Naughty that detects AI comments, reads the commenters' headlines, messages them as if wanting to buy their service, and keeps them answering questions. Khadri reports that 10 of 11 people replied.
What is actually being sold when a client can run the same models themselves?
Read more
Start here for Khadri's answer, given to a real buyer who asked that question mid-deal.
Khadri describes a six-figure upsell that almost died when the client asked why they should pay for a service the client could hand to their own AI agents.
The post answers with five things Khadri says a model cannot replace: taste (knowing which outputs to throw away), context (knowledge never published online, drawn from agency work at Earleads), execution (the last 70% of the work after the deliverable), accountability (a human who owns the outcome), and an external feedback loop (agents at Yalc.ai graded by the market rather than by themselves).
The client said yes before the call ended.
For brands planning a campaign
For any brief about objection handling or differentiating an AI-assisted service, this is a ready-made workshop segment: the five-part frame can be used as an exercise where a sales team writes its own answer to the 'why not just use my agents' question.
Start here for the full workflow, including the exact message that brought a lost prospect back.
Khadri describes a monthly routine of going through every lost deal, then automating it: a Yalc.ai agent run through Claude Code pulls every Closed Lost deal from Attio over the last 12 months via the CRM's connector, reads the lost-reason field, the last three activities before the stage changed, and any attached transcripts or notes, clusters objections against buyer attributes and the stage where deals died, scores open deals sitting on a losing pattern, and drafts a personal re-opening message.
The post quotes the message sent to one prospect, which asks a specific question about what went wrong and offers a phone number.
The prospect responded, and the deal restarted.
For brands planning a campaign
This is the closest match to a lost-deal-recovery brief. The five agent steps can be used as a live demo script, and the quoted outreach message works as a starting template a team can adapt.
Start here for five named skills a sales team can install, with what each one reads and returns.
Khadri published a library of five skills to run on Claude with lemlist and Claap connected: one scores every lead in the last five campaigns against your ideal customer profile and shows which job titles crept in; one follows replies into the booked calls and ranks campaigns by good calls per 100 leads instead of reply rate; one takes a closed deal and checks how much of the campaign shares that buyer's context, and writes new list filters if it does not; one reads a prospect's reply, finds a similar company's past call, and drafts a short pre-call note that it never sends on its own; and one pulls the exact words prospects used in calls that progressed and writes five opening lines. The library is described as open source and free, and Khadri says it is part of the Yalc.ai GTM operating system.
For brands planning a campaign
Useful for a session on outbound instrumentation. It also makes the data requirements concrete: the skills only work with a CRM plus recorded calls, which is a qualification question for any team considering the approach.
What does Khadri do outside straight B2B workflow building?
Read more
Start here for the satirical, public-facing side of the work.
After an earlier post arguing that automated AI comments are bad for business reached 200,000 views, Khadri built an open-source agent called Naughty.
It reads recent posts, detects AI comments, reads the commenters' headlines, messages them pretending to want to buy their service, and keeps them in a loop of questions.
Khadri reports that 10 of 11 people replied.
The agent was built on the same Yalc.ai harness.
For brands planning a campaign
A brand-safety signal as much as a creative one. Khadri is willing to publish and open-source a provocative tool aimed at a named group of people, which matters when matching a creator to a conservative enterprise audience.
How did Khadri come to this work, and what is the underlying view of AI?
Read more
Start here for the origin story that explains the recurring 'taste' argument.
Khadri traces the interest to 2022, writing a thesis on how human bias drives financial markets and adding an AI section to test whether models carried the same biases.
Khadri says they did, and that this made clear the technology would spread.
The post cites Roy Amara's line about overestimating a technology in the short run and underestimating it in the long run, argues that coding is math and therefore the easiest target for AI, and holds that a layer of taste, judgment, or consciousness will not be solved soon.
For brands planning a campaign
Background material for a talk framing. The taste-and-judgment thesis set out here is the same argument used to close the six-figure deal, so the two pieces work as a pair in a workshop.
What does Khadri think content marketing is worth?
Read more
Start here for the shortest version of the argument.
A short post aimed at people who lead content with the question 'how do I do this as efficiently as possible?' Khadri calls content the most asymmetrical-return marketing channel and argues the guiding question should be how to get the best available outcome instead.
For brands planning a campaign
A quotable one-liner for a content-strategy segment. It is an opinion, not a workflow, so it should not be used as evidence that Khadri runs content programs at scale.
AI spam, automation ethics, and playful retaliation1
Direct recommendations to founder audiences1
AI-assisted go-to-market workflows
Othmane Khadri's posts repeatedly show AI agents wired into real B2B sales and marketing work rather than discussed abstractly.
Read more
Examples include outbound campaign analysis, call follow-up, CRM deal review, content, and open-source agents. The named tools include Claude, Claude Code, lemlist, Claap, Attio, and Yalc.ai.
For brands planning a campaign
A brand selling sales engagement, CRM, or AI workflow software can evaluate Khadri for a practical build-along, workflow teardown, or implementation-focused session. The material suits buyers who want to see how agents fit existing tools, not a broad AI futurism pitch.
One post describes a monthly review of every lost CRM deal, then an agent that pulls Closed Lost records from Attio, reads lost reasons, recent activities, and transcripts, clusters objections, scores open deals against losing patterns, and drafts personalized reactivation messages.
Read more
Another frames the same sales problem around what a human service sells when buyers can use their own agents: taste, context, execution, accountability, and an external feedback loop.
Review lost deals on a recurring cadence.
Use CRM notes, call transcripts, and objection clusters to decide which open deals to save.
Write a specific, human follow-up that names the likely reason the deal stalled.
For brands planning a campaign
Direct fit for the supplied brief.
Khadri can speak to lost-deal recovery and objection handling with a named workflow and example message.
A workshop could focus on building the agent, reading its objection clusters, and keeping human accountability for the final ask.
A post aimed at lemlist and Claap users presents five open-source Claude skills: one scores leads against an ideal customer profile across recent campaigns to expose job-title drift, one follows replies into booked calls to rank campaigns by good calls per 100 leads, one checks whether a closed deal's buyer context appears across a campaign, one drafts pre-call insight notes, and one turns prospect language from successful calls into first lines for new campaigns.
Read more
Connect lemlist and Claap to Claude through MCP, run the skills against recent campaigns, and inspect the resulting campaign, call, and list outputs.
For brands planning a campaign
A strong fit for brands in outbound, sales engagement, call intelligence, or sales operations that want a concrete demo of campaign diagnostics. The open-source library gives a workshop or co-marketing angle with tangible artifacts rather than a generic talk on AI in sales.
Khadri argues that AI is strong at deterministic, math-like flows but does not supply taste, judgment, or consciousness.
Read more
The argument appears in a post about getting into AI in 2022 through research on human bias in financial markets and in a sales post about why a client should pay for human service when agents can produce a deliverable. Khadri uses Roy Amara's saying about overestimating technology in the short run and underestimating it in the long run, and says the last mile of execution is human and expensive.
For brands planning a campaign
Useful for a thought-leadership panel or content collaboration on where AI helps and where human judgment still carries the outcome. The argument is more useful to brands selling services, judgment-heavy work, or accountability than to brands selling fully automated AI outcomes.
One post advises against leading content with the question, 'How do I do this as efficiently as possible?' Khadri says content has the most asymmetric return on investment as a marketing channel and urges leading with the best available outcome instead.
Read more
For brands planning a campaign
A brand evaluating creator-led growth or content marketing can use this as a discussion point about quality-first content. It fits a workshop or talk on why efficiency-led content underperforms, with examples from Khadri's own posts.
AI spam, automation ethics, and playful retaliation
Khadri built an agent named Naughty that detects AI comments, reads commenters' headlines, DMs them pretending to want to buy their service, and keeps them in a loop of questions.
Read more
The post says 10 out of 11 people replied and that the agent is open source.
The post describes a deliberate time-wasting prank, not a standard outbound playbook. It shows a confrontational, humorous response to AI-generated comments rather than a neutral marketing method.
For brands planning a campaign
A brand in anti-spam, community tools, or creator-authenticity work could evaluate this as a cultural talking point. B2B SaaS brands with conservative buying audiences should account for the prank framing and possible tone risk before using it in a campaign.
Across several posts, Khadri ties the examples to Yalc.ai as a GTM operating system or harness and to the agency Earleads.
Read more
Khadri releases agents and skill libraries openly, including the five lemlist and Claap skills and the Naughty agent. The posts present building in public as part of the work, not a separate marketing layer.
For brands planning a campaign
Brands considering community-led growth, developer marketing, or agency-led implementation can evaluate Khadri for build-in-public collaborations. The open-source angle gives a campaign a reusable artifact that an audience can inspect and test.
One post recommends Gradient's accounting team to founders who need accounting services, naming Rohan Krishan and saying Khadri has worked closely with the team for months.
Read more
It is a direct endorsement rather than a workflow breakdown.
For brands planning a campaign
This shows Khadri will use a social post for a direct service recommendation to a founder audience. It is weaker evidence for a campaign brief than the workflow posts, but it can support a referral-style or founder-services collaboration if the brand accepts a short, plain endorsement format.
Othmane Khadri's posts document two businesses as his own: the GTM agency Earleads and the GTM operating system Yalc.ai.
Read more
The remaining relationships are tool or vendor references inside those workflows — Claude and Claude Code, lemlist, Claap, and Attio — plus an endorsement of the accounting firm Gradient. None of the reviewed posts discloses a sponsorship, affiliate arrangement, client contract, or investment.
Earleads
Owned business
Khadri states that Earleads is his agency and that he built it in 2023.
Read more
He describes its GTM offer as selling outcomes and owning full execution, and says it gives him access to client-specific knowledge he does not publish.
Khadri calls Yalc.ai his GTM operating system and refers to its agents as "our" agents, describing it as the harness behind the lemlist-Claap skill library and the agent he wired through Claude Code to review lost deals.
Read more
He also explains that Yalc.ai agents are graded by weekly market results across clients.
Claude Code is the environment Khadri uses to run his agents: the lemlist-Claap skill library is built to be wired into it, and he wired a Yalc.ai agent through it to analyse closed-lost deals.
Read more
A separate post compares his agency's work to a $200 Claude Max plan when explaining what a client is actually paying for. No commercial relationship with Anthropic is disclosed.
Claap appears alongside lemlist as the call-recording side of Khadri's free skill library, which reads booked calls to check whether the right buyers showed up.
Read more
He describes the tools as connected to his own workflow, without disclosing any commercial relationship.
Othmane Khadri's posts sit where AI agents meet everyday B2B sales work, and they read best in build order.
Read more
Start with the two workflow posts: the five open-source Claude skills for lemlist and Claap users, which score leads against an ideal customer profile, follow replies into booked calls, rank campaigns by good calls per hundred leads, and turn prospect language into opening lines; and the Yalc.ai agent that pulls closed-lost deals from Attio, clusters the objections behind them, and drafts a personalized reopening message.
Those two show the method.
The account of the six-figure upsell that nearly collapsed then explains the argument underneath it: when a buyer can run the same models, what a service sells is taste, private context, the last mile of execution, accountability, and a feedback loop graded by the market rather than by the model itself.
The 2022 origin post widens that into a claim about which work AI absorbs first, and the short content post applies the same standard to marketing decisions.
Two posts are outliers worth knowing about before following: an agent built to waste the time of automated commenters, and a plain recommendation of an accounting team to founders.
Also writes about wiring AI into company operations and teaches through numbered frameworks: one post maps seven stages of AI use at work, from asking a chatbot to agents that pass work to each other without a person in between, with a test attached to each stage.
Read more
Moura lays out a general ladder of autonomy for any team and draws on CrewAI's own growth team for examples. Khadri stays inside B2B sales and CRM and argues about what a paid service is worth once clients bring their own agents, which is a narrower and more commercial question than an adoption ladder.
Publishes step-by-step accounts of agents running a business process end to end, with the numbers attached: calls handled, web chats and texts answered, and revenue attributed to the AI rather than to a human on the phone.
Read more
Bedoya automates a one-person consulting business's whole funnel, from payment link through intake call to CRM update. Khadri's workflows run inside an agency's go-to-market stack, keep a person accountable for the final ask, and spend more time on which judgement a client should pay for.
Builds agents for a marketing job and writes the method as a short numbered playbook with a confrontational opener, such as telling readers to burn their SEO agency retainer, then test, track, and adjust on real data.
Read more
Astarabadi's examples are ecommerce and DTC: batches of AI ads run against a control on the same budget, and a Shopify store audited for AI search. Khadri's examples are B2B sales and CRM, and the argument is about judgement and accountability rather than production volume.
Othmane Khadri's LinkedIn profile recorded 18,939 followers on October 8, 2026.
Read more
The supplied posts show content aimed at B2B sales and GTM readers: outbound campaigns, CRM follow-up, lost-deal recovery, objection handling, and AI agents used in sales workflows.
That gives the buyer topical audience alignment for a lost-deal recovery or outbound objection-handling brief, but it does not establish audience composition.
No geography, age, gender, employer, occupation, seniority, reach, or engagement-quality data is attached.
Follower count is a scale signal, not proof that the audience contains the roles a B2B SaaS campaign needs.
For this brief, ask for recent LinkedIn audience demographics if available, plus reach and engagement on posts specifically about CRM reactivation, lost-deal recovery, or objection handling; where demographics are unavailable, request documented audience survey results or first-party evidence on job function and seniority.
Until then, the page can support a topical-match decision but not a measured professional-audience fit decision.
LinkedIn followers
18939
LinkedIn account record for @othmane-khadri-b48162236 · 2026-10-08
The reviewed posts document no sponsorship packages, paid placements, or media-kit terms for Othmane Khadri.
Read more
His commercial activity in these posts is service-based rather than publisher-style: he describes running the GTM agency Earleads and the GTM operating system Yalc.ai, and he distributes some agent tooling for free as open source.
That distinction matters for a buyer's plan.
There is no published rate card, audience or deliverable specification, or partnership inquiry route in the material reviewed, so a campaign would have to be scoped directly with Khadri rather than selected from an advertised format.
Nothing here supports a claim that he never accepts sponsorships; it only means no offer is documented in the posts reviewed.
His most partnership-relevant behavior is building and describing agent workflows that plug into other vendors' products, including a five-skill library for lemlist and Claap users and a lost-deal agent run through Attio and Claude Code.
Those are integration and consulting examples, not evidence of payment or a formal partner program.
Brief: a B2B SaaS company selling CRM, sales-engagement, or revenue-intelligence software wants a practitioner-led talk or workshop on lost-deal recovery and outbound objection handling.
It has a sales-leader audience and wants a build-along format rather than a keynote.
Why this is a good match.
Khadri's posts are a working demonstration of exactly this task, not commentary about it.
Their lost-deal post describes the full loop: pull every Closed Lost deal from Attio over 12 months, read the lost-reason field, the last three activities before the stage changed, and attached notes or transcripts, cluster objections against buyer attributes and the stage where deals died, score open deals sitting on the same losing pattern, and draft a reopening message.
The post quotes the message that worked — a specific guess at the reason the deal stalled, a request for 5-10 minutes, and a phone number — and says the prospect responded and the deal restarted.
That gives a workshop a real script and a real outcome to argue from.
Second, the objection-handling side is already packaged.
The six-figure upsell post answers a buyer who asked why they should pay a service when they could point their own agents at the same job, and it answers in five named parts: taste, context, execution, accountability, and an external feedback loop.
Each part is a teachable segment, and the client said yes before the call ended, so the frame has a documented close behind it.
Third, the outbound intelligence library is a ready-made demo.
The lemlist and Claap skills score leads against an ideal customer profile across recent campaigns, rank campaigns by good calls per 100 leads instead of reply rate, check whether a closed deal's context repeats across a campaign, draft a pre-call note, and turn prospect language from calls that progressed into opening lines.
Khadri describes the library as open source and free, which means a session can point the audience at something they can install rather than at a slide deck.
Audience alignment supports the brief.
The 18,939 LinkedIn followers and the recurring content on outbound campaigns, CRM follow-up, and deal recovery point at B2B sales and GTM readers.
Follower count is a scale signal only; the packet has no audience demographics, reach, or engagement data, so a sponsor should ask for LinkedIn audience demographics plus engagement on the lost-deal and objection-handling posts before committing budget.
The commercial caveat is real and worth stating plainly.
No rate card, media kit, sponsorship package, or inquiry route is documented in the reviewed posts.
Khadri's visible commercial activity is service-based — the agency Earleads and the Yalc.ai GTM operating system — plus free open-source tooling.
A campaign here has to be scoped directly rather than selected from an advertised format, and there is no evidence either way about whether Khadri accepts sponsorships.
When to choose a specialist
Fit: poor.
Read more
Brief: a vendor selling a fully automated outbound platform wants broad-reach sponsored posts guaranteeing a follower-count-driven audience, with the promise that its product replaces SDR headcount and sends without human review.
The buyer wants a straight product endorsement from a GTM voice, on a fixed fee per post, to a conservative enterprise audience.
Three things in the reviewed material push against this.
The product claim contradicts Khadri's stated position.
The upsell post argues that AI moved where the time goes rather than removing the work: the deliverable is maybe 20-30% of the job and the last 70% — described there as painful, expensive, and human — is the part the buyer actually feels.
It also warns that handing an agent full execution rights produces slop, and the accountability section says a human owns whatever sits between the automation and the outcome the client paid for.
A campaign promising an autonomous replacement of that human layer asks Khadri to argue the opposite of the argument that closed their own six-figure deal.
Specific tool behavior cuts the same way.
The pre-call insight skill reads a prospect's reply, finds a comparable past call, and drafts a short note — and the post is explicit that it never sends anything on its own.
The published pattern is draft-and-approve, not unattended sending.
Audience and tone are the wrong shape for the buyer.
The audience evidence is a follower count, not a verified enterprise-buyer composition, so a guaranteed-reach enterprise endorsement is not supported.
The most viral item in the set is the Naughty agent, which DMs people under false interest and keeps them in a question loop; it is a deliberate prank and it shows Khadri is willing to publish a confrontational anti-spam stunt.
That is a genuine creative asset for some briefs and a brand-safety problem for a conservative enterprise buyer who wants a clean endorsement.
Finally, the format does not exist yet.
Nothing in the reviewed material documents a paid post package or a rate card, so the brief as written assumes a commercial offer that has not been evidenced in the material reviewed.
That is not proof Khadri would decline; it means the brief cannot be priced or planned from what is available.
A worked campaign example
Campaign angle, built for the good-fit brief: a 45-minute working session titled around reopening closed-lost revenue with an agent you can audit. The format is demo plus build-along, not a keynote, because the material already contains steps and scripts an audience can follow on their own CRM.
Opening, 5 minutes.
Put the client question from Khadri's upsell post on screen — why pay a service when the buyer can point their own agents at the same work — and answer it with the five-part frame: taste, context, execution, accountability, and an external feedback loop graded by the market instead of by the agent itself.
This sets the session's argument that the human part of the job is the part being sold.
Workflow walkthrough, 15 minutes.
Step through the lost-deal loop as it is described: monthly pull of every Closed Lost record, the lost-reason field, the last three activities before the stage changed, attached notes or transcripts, objections clustered against buyer attributes and the death stage, open deals scored against the losing pattern, and a drafted reopening message for the team to review.
Show the message template Khadri published — the specific guess at what went wrong, the short time ask, the phone number — and explain why the guess is doing the work rather than the automation.
Build-along, 15 minutes.
Teams write their own answers to the five-part frame for their own offer, then write one reopening message for a real deal they lost, using the published template as the starting shape.
The point of the exercise is that the frame is portable: it makes the sales team state what it sells when the buyer can run the same models.
Outbound instrumentation, 10 minutes, optional if the audience uses lemlist or Claap.
Walk the five skills and what each reads and returns: ICP drift across recent campaigns, campaigns ranked by good calls per 100 leads instead of reply rate, a closed deal's context checked across a campaign, a pre-call note, and opening lines pulled from the words prospects actually used.
Note that the skills need a CRM plus recorded calls, which is the qualification question for any team considering the approach.
Close and measurement, 5 minutes.
Hand off a shortlist of next steps: connect the CRM, run the lost-deal review on one month of Closed Lost records, and compare reopens against a control set of deals left alone.
Measure reopened deals and booked calls from reopened accounts, plus how often the drafted message is edited before sending — that last number shows whether the team is adding the judgment the workflow assumes.
What to collect before booking.
The repositories, CRM records, call recordings, and transcripts the posts refer to are not in the reviewed material, so the tools are described as Khadri presents them and have not been independently inspected.
Collect the linked repositories and have someone run at least one skill against a real campaign before the session is sold as a working demo.
Also ask for audience demographics and engagement on the deal-recovery posts, since the packet has only a follower count.
Tone risk is low for this brief: the satirical Naughty agent stays out of the session.
Choose Moura when the session needs a shared decision framework for a leadership audience rather than a single-task script.
Read more
The seven-stage ladder — prompter, monitor, delegator, manager, conductor, ringmaster, voice-memo CEO — gives a sales or GTM leadership room a common vocabulary for judging where their team currently sits, and each stage ships with a test question, such as whether work happens before anyone asks for it.
The workflow-versus-agent post adds the sharpest single diagnostic in the comparator set: did you hand the system a list of steps or a goal?
That test takes one minute to apply to an accounts-payable-style process and produces a clear answer.
For a workshop whose real subject is deciding what to hand over, this material teaches the decision; Khadri's material demonstrates the outcome of one decision already made.
Moura's material is general and instructional.
It contains no lost-deal recovery loop, no reopening message, no objection-handling frame, and no documented outcome from a deal that was revived.
The examples come from a growth team's account-activity monitoring, support triage, and accounts payable, and the final stage is described as somewhere nobody runs a company yet.
A session built on Moura gives the audience a clean mental model and no field-tested script for the buyer's specific task.
Choose Bedoya when the audience is one-person or small professional-services businesses and the brief is about replacing routine labor around selling, not about winning back lost enterprise deals.
Read more
Her published walkthrough shows a full intake pipeline: payment link, agreement, AI voice-agent onboarding that collects documents and context, calendar unlock, call analysis, and CRM update.
The follow-up post reports the output of an AI-run sales motion spanning calls, web chat, text, and email — 19 buyers, $4,829 in revenue, $202 in AI and subscription cost — and frames the value as zero minutes on sales calls and zero commission.
For a session whose promise is headcount and time saved on low-ticket offers, that is a more direct story than Khadri's.
The numbers are self-published with no records in the reviewed material to check them against, and the offer sizes are far below a B2B SaaS deal: a $3,000 purchase, eighteen buyers at $99, one $47 add-on.
There is no CRM reactivation content, no objection handling, and no enterprise sales cycle.
The arguments also point opposite ways for the same audience question: Bedoya's posts sell removing the human from the call, while Khadri's closes by arguing the human last mile is the part the buyer actually feels.
A buyer who picks Bedoya is choosing that frame, not a neutral substitute.
Choose Astarabadi when the brief shifts to DTC ecommerce demand generation.
Read more
His posts cover creative testing velocity — pick one product, produce twenty AI ads, run them against the current control on the same budget, and move spend batch by batch — and an AI-search audit for Shopify stores that sorts Search Console queries, checks what ChatGPT-type assistants search for, and matches shopper questions to product or collection pages. For a session on fast creative iteration or AI-search visibility in an ecommerce catalog, this is closer to the buyer's daily work than anything in Khadri's material.
The buyer task is different, so it is a substitute only if the campaign definition changes.
There is no B2B sales cycle, no CRM, no closed-lost workflow, and no objection-handling framework.
The SEO post ends by asking readers to comment a keyword to get the walkthrough, which is a lead-capture pattern rather than a documented workshop or commercial package.
Under the current lost-deal recovery brief, this material does not answer the task; it is the alternative for a different brief.
This profile is based on seven LinkedIn posts published from Othmane Khadri's account, read in full.
The posts cover an open-source library of five Claude skills for lemlist and Claap users, a Yalc.ai agent wired through Claude Code to review closed-lost deals in Attio, a six-figure upsell that turned on a client asking why to pay a service when they could run their own agents, a satirical agent built to waste the time of automated commenters, the 2022 origin of the AI interest, a short post on content strategy, and a recommendation of the accounting firm Gradient.
Two limits shape what this page can claim.
First, none of the retained posts carries a publication date, so the account of Khadri's work is organised by subject and argument rather than by timeline; the follower count (18,939) was observed on 8 October 2026.
Second, the repositories, CRM records, call recordings, transcripts, and campaign data the posts refer to were not part of the reviewed material, so the agent workflows are described as Khadri presents them and have not been independently inspected or run.
The reviewed posts establish two businesses as Khadri's own, the agency Earleads and the GTM operating system Yalc.ai, and treat Claude, Claude Code, lemlist, Claap, and Attio as tools in those workflows.
No sponsorship, affiliate, client, employment, or investment relationship is disclosed in the material, and no rate card, media kit, or sponsorship package appears, so there is no evidence either way about whether Khadri accepts paid partnerships.
Comparisons with similar creators rest on two retained posts each from three other accounts.
One research gap remains open for a name search.
The reviewed posts do not establish whether Khadri has signature books, shows, or earlier public work, so this page should not be treated as a complete account of what Khadri is known for until those name-search results are checked against primary sources.
Recent engagement
Across their last 5 LinkedIn posts, the median post got 15 reactions and 2 comments. Checked 2026-11-02.
