João Moura is the founder of CrewAI, a company building software for creating and managing AI agents for business work. His posts explain how agents can handle changing goals, coordinate across workflows, and operate with controls such as cost limits and data policies.
João Moura explains agentic AI for operations leaders, turning agent goals, controls, and enterprise adoption into practical buying decisions.
Best for
B2B automation or AI platform seeking explainer content for operations and technical leaders.
Not for
Consumer mass-market or lifestyle campaigns needing high-volume short-form video and broad awareness.
Key takeaways
An agent is defined by being handed a goal, not a list of steps.
AI adoption is best described as a ladder of observable workplace behaviors.
Governance belongs in the execution runtime, not in code or at the network edge.
Check first: Confirm sponsorship offer, rate card, media kit, audience demographics, and prior video performance.
Main ideas
An agent is defined by being handed a goal, not a list of steps · AI adoption is best described as a ladder of observable workplace behaviors · Governance belongs in the execution runtime, not in code or at the network edge
João Moura is the founder of CrewAI, known for explaining how businesses can move from fixed AI workflows to agents that work toward goals, coordinate tasks and operate under controls.
His posts make abstract ideas concrete: one compares two ways to handle accounts payable, distinguishing a system that follows set steps from one given the goal of keeping accounts clean.
Another maps AI use across seven stages, from asking a chatbot for an answer to coordinating agents.
Moura also writes about the harder work of putting AI into businesses: fitting existing systems, meeting governance needs and earning buyer trust.
These examples make his work useful to readers trying to understand what agents can do—and where human judgment and operational controls still matter.
Moura’s strength is turning agentic AI into practical decisions rather than treating it as a technology trend.
He gives readers a test they can apply: did they give the system steps, or a goal?
He also makes room for exceptions, noting that fixed workflows can be cheaper and easier to trust when a process rarely changes.
In posts about enterprise buyers, he uses specific setbacks—a bank facing multiple vendors, or customers wanting internal integrations instead of a large public catalog—to explain why deployment, security and compatibility can matter more than technical breadth.
The result is a mix of accessible frameworks, product-level detail and candid corrections that helps readers judge both the promise and limits of AI agents.
Moura's posts return to one question from several angles: how much of a job do you hand to software, and what has to be in place before you can trust it with more.
Read more
They draw the line between a fixed workflow and an agent given an outcome, describe AI adoption as a ladder of observable workplace behaviors, and argue that governance has to sit in the runtime rather than in code or at the network edge.
A second thread is about selling and supporting this in large enterprises: buyers want fewer vendors even as the market rewards specialists, what they value in integrations is fit with systems they already run, and the hard work starts at production rather than in the prototype.
A third is personal: Moura keeps writing code as a source of ideas while staying out of critical tasks that would make them a blocker, and treats industry recognition as encouragement rather than proof.
An agent is defined by being handed a goal, not a list of steps
Moura draws the line between a workflow and an agent at whether someone wrote out the steps or handed over an outcome.
Read more
A system following specified steps does the same thing in the same order every day; a system given a goal works out its own path and keeps re-deciding as conditions change.
Moura is explicit that this is not an argument against workflows: when the path never changes, fixed steps are cheaper and easier to trust.
The reasoning is about where a person's judgment stops being needed, not about which approach is more advanced.
The same accounts-payable job run two ways.
In the first, software reads every invoice, matches it against a payment, flags anything more than 15 days overdue and drafts the follow-up, in that order every day.
In the second, it gets one sentence: keep payables clean, flag anything overdue or mismatched, and tell me what needs sign-off.
It pulls the ledger, works out what does not reconcile and decides what to chase.
The same distinction reappears in the maturity ladder, where delegating one defined job is an intermediate stage and handing over a goal marks the move from workflow to agent.
AI adoption is best described as a ladder of observable workplace behaviors
Moura's growth team maps seven stages of using AI at work, each defined by something you can check rather than by capability claims.
Read more
The stages run from asking a chatbot and copying the answer somewhere yourself, through something that runs before you ask, doing one job on its own, being handed a goal, agents passing work to each other, several systems connecting what a person would otherwise connect mentally, and finally one agent overseeing the rest.
Each stage comes with a test question.
Moura presents the last stage as a direction builders are heading, not a way companies operate today.
At the Monitor stage, an agent checks HubSpot, the website and the last 24 hours of deal activity each morning and drops five bullets into Slack before anyone asks, but a person still reads it and does the work. At the Ringmaster stage, a spike in support tickets and a shipping delay arrive as one decision to review rather than two alerts someone has to connect.
Governance belongs in the execution runtime, not in code or at the network edge
In launching CrewAI's Agent Control Plane, Moura argues that where controls live determines how usable they are.
Read more
Put guardrails in agent code and every policy update becomes a software release; put them in an external gateway and you get routing and API reuse, but a proxy cannot see inside the agent loop, cap how many reasoning iterations run, pause a run midway or restrict execution to business hours. Because what is being governed is the degree of autonomy delegated to a system, control belongs where that system runs.
The announced controls include visibility across deployed automations, cost control by model and agent, and thirteen built-in policies covering detection and masking of personal data, allowed models and tools, execution windows and frequency, content filters and cost limits, plus custom policies evaluated at runtime without a redeploy.
Enterprise buyers want fewer vendors, even when the market rewards specialists
Moura tells a story about selling CrewAI to a large bank, where each answer about observability, then memory, pointed to a different product.
Read more
Moura accepts that the industry is fragmenting into layers, each with companies trying to be the best in the world at one thing, but argues that this pattern looks different from the buyer's side of the table: nobody running a large enterprise wants five contracts, five security reviews or five vendors to chase when something breaks at two in the morning.
The resolution Moura describes is building the unglamorous parts anyway, while conceding CrewAI is not best-in-class at each of them.
The conclusion drawn is that architecture arguments get settled by procurement rather than by engineers.
By the third answer in the meeting, Moura says the buyer was thinking they would need five products to use one.
What enterprises value in integrations is fit with what they already run
Moura describes building access to thousands of public tool integrations and then learning from customers that they wanted a private registry of the tools their own teams had already built.
Read more
From outside a platform, more connectors reads as more capability; from inside an enterprise, what matters is what is already wired into your systems, with your authentication around it and your data rules on top.
Moura treats that as a mistake in perspective: thinking like someone building a platform instead of someone running an enterprise.
A related argument is that standards matter less for access to everything than for preserving the ability to change vendors later without losing the integration work.
Customers who had been given thousands of MCP servers said they actually had 20 internal MCPs and wanted to use those.
Model choice and workflow improvement should be measured in separate loops
Moura frames a recurring engineering question, whether to move crews to a new model, and argues the answer needs measurement rather than a dashboard someone squints at.
One benchmarks a model against a baseline on quality, cost, speed and tokens, so a frontier model can be used on one role and a cheaper open-weight model on another.
The other keeps models fixed and adjusts the prompt, tool definition or routing until a gate clears.
Moura flags the guardrail that matters: left unguarded, such a loop will overfit a small evaluation set or rewrite a configuration into a mess purely to turn a failing result into a passing one.
The described default is a person opening a dashboard, finding the failed step and pasting the finding back into the terminal so a coding agent can fix it, which Moura calls making the human the router.
The post also states that most CrewAI documentation access now comes from AI agents and that a fifth of CrewAI projects are created by coding agents. These are Moura's own figures and are not independently verified in the supplied material.
Prototypes are easy; production deployment is the hard part
Moura's position is that many companies can now build an AI prototype, and that the difficulty lies in what comes next: choosing the right business process, integrating it with existing systems, getting it through security and governance, putting it into production and operating it afterwards.
Read more
This is the gap partnerships are meant to close, and it is also why global expansion, in Moura's telling, means local people, partners and language rather than selling from far away. A related post looks at what remains after an approval, where status is inferred and work moves through email, spreadsheets and follow-ups that can go unnoticed, and argues that the coordination around a decision is where the delay lives.
The DXC Technology partnership is aimed at helping companies in Latin America move agentic AI from experimentation into operations across financial services, retail, manufacturing and consumer goods. In the credit case, the decision itself can be fast while the rest of the week disappears into a missing document, a title sent to the wrong address, or the person who could fix it being out that day.
The credit post is a webinar invitation, and Moura is joined by Arindam Choudhury of Capgemini to work through a live example and decide what an agent could own, what still needs a person, and what should not be automated at all. Those judgments are presented as shared with the Capgemini guest rather than as Moura's alone.
Keep building, stay out of the critical path, and don't mistake recognition for progress
Alongside the product arguments, Moura writes about how the company is run.
Read more
Moura still codes on weekends out of enjoyment and as a source of ideas, but stopped taking tasks after learning that working on anything critical makes them a blocker, so ideas go in as pull requests for the team to finish.
Improving the product is described as an ongoing effort, including an experiment that made all of CrewAI's agents run five times faster from a small fix that had been missed.
The stated aim is that every agent built with CrewAI should be as good by default as a coding agent Moura uses daily.
Industry recognition, meanwhile, is treated as validation of the category rather than of the company: a slide does not build a company.
A conference slide listing CrewAI next to Claude Code, Microsoft Copilot Studio, Agentforce, Devin and AutoGen is read as evidence the agent category has become real, with the caveat that the team, community, customers and weekly releases are what build the company.
The single clearest statement of the distinction a newcomer needs before anything else in this profile: what makes an agent different from an automated workflow.
Read more
Moura runs the same accounts-payable job two ways.
In the first, a system reads every invoice, matches it against a payment, flags anything more than 15 days overdue and drafts the follow-up, in the same order every day.
That is a workflow, because someone wrote out every step.
In the second, the system gets one sentence: keep payables clean, flag anything overdue or mismatched, and tell me what needs sign-off.
Nobody wrote the steps; it pulls the ledger, works out what does not reconcile and keeps deciding what to chase.
Moura calls handing over the goal the real line between an assistant and an agent, and adds that when the path never changes, fixed steps are cheaper and easier to trust.
For brands planning a campaign
A single-test format works here: ask operations leaders whether they gave a process steps or a goal, and have them apply it to one thing they automated this year. The accounts-payable example is concrete enough to carry a short video without added explanation.
The fullest framework in the supplied material for placing a team's current use of AI and seeing what the next step actually requires.
Read more
Moura's growth team maps seven stages, from The Prompter, where every task starts with a person opening a chat window and copying the answer somewhere, through the Monitor (something runs before you ask, but a person still reads it and does the work), the Delegator and the Manager (you hand it a goal rather than steps), the Conductor (agents hand work to each other), the Ringmaster (several systems connect what a person would otherwise connect in their head), up to the Voice Memo CEO, which Moura says nobody runs a company with yet. Each stage comes with a test question, such as whether a person is still passing work from one step to the next.
For brands planning a campaign
Ready-made series structure, one stage per episode, with the stage tests giving the audience something to check themselves against. The final stage is speculative, so a sponsor should keep the series grounded in the stages enterprises can reach today.
Explains why CrewAI builds deployment, metrics, traces, data handling and permissions instead of being best-in-class at one layer, told through a single sales meeting.
Read more
A large bank asked how it would do observability, then memory, and each answer pointed to a different vendor.
As Moura tells it, by the third answer the buyer was thinking they would have to buy five products to use one.
Nobody running a large enterprise wants five contracts or five security reviews, Moura argues; they want one number to call.
The lesson drawn is that architecture arguments get settled by procurement, not by engineers.
Moura is candid that CrewAI is not the best in the world at every one of those parts.
For brands planning a campaign
Strong material for a buyer-education segment on consolidation versus best-of-breed, and the one-vendor framing fits a platform sponsor better than a point-solution one.
A candid account of getting the enterprise value proposition wrong, and the clearest explanation of why standards matter to buyers.
Read more
CrewAI built access to thousands of public MCP servers, and customers said they wanted their own 20 internal MCPs instead: a private registry of what their teams had already built, with their own authentication and data rules.
From outside, more connectors reads as more capability; from inside, the ones that matter are whatever is already wired into your systems.
Moura says the mistake came from thinking like someone building a platform rather than someone running an enterprise, and argues that standardization is valuable mainly because it preserves the right to change vendors later.
For brands planning a campaign
The internal-registry detail is specific enough to anchor a technical discussion with platform engineers, and the correction is more persuasive than a claim of foresight would be.
The most technical selection in the set, and the clearest window into how Moura frames tooling gaps for developers.
Read more
Moura describes a default workflow in which a person opens a dashboard, finds the step that failed, and pastes the finding back into the terminal so a coding agent can fix it, making the human a router.
To answer whether crews should move to a new model, CrewAI shipped crewai eval, an experimental open-source command line tool with two separate loops: model selection, which benchmarks a new model against a baseline on quality, cost, speed and tokens, and self-improvement, which adjusts the prompt, tool definition or routing until a gate clears.
Moura flags the guardrail: left unguarded, a loop will overfit a small eval set or rewrite a configuration into a mess just to turn a red result green.
The post also states that most CrewAI documentation access now comes from AI agents and that a fifth of projects are created by coding agents, figures that are Moura's own and not independently verified here.
For brands planning a campaign
Aimed at a developer audience rather than operations leaders. Useful if a campaign needs technical depth or credibility with engineering teams; weaker as the opening item for a general business brief.
Shows how the governance argument for agent platforms is made, and what the shipped controls actually cover.
Read more
Moura argues that governance belongs in the execution runtime rather than in agent code or at the network perimeter.
Put guardrails in code, and every policy update becomes a software release; put them in an external gateway, and you get routing and API reuse, but a proxy cannot see inside the agent loop, cap reasoning iterations, pause a run mid-loop or enforce business-hours windows.
The announced controls include visibility across deployed automations, cost control by model and agent, and thirteen built-in policies covering data handling such as detection and masking of personal data, allowed models and tools, execution windows and frequency, content filters and cost limits, plus custom policies that evaluate at runtime.
This is a company launch post and reads as one.
For brands planning a campaign
The policy count and the runtime-versus-gateway argument give a governance-themed sponsored segment something specific to discuss. Treat the claims as CrewAI's own announcement rather than an independently tested result.
The clearest statement of the company's expansion strategy and of Moura's view that production deployment, not prototyping, is the hard part.
Read more
CrewAI is partnering with DXC Technology, which Moura describes as around 115,000 people across more than 70 countries, to help companies in Latin America move agentic AI from experimentation into operations in financial services, retail, manufacturing and consumer goods.
Moura, who describes himself as a Brazilian founder building a global company, argues that many companies can now build a prototype, while identifying the right process, integrating it with existing systems, getting it through security and governance, and operating it after launch is the harder work.
Global expansion, in this telling, means local people and partners rather than selling from far away.
For brands planning a campaign
Relevant if a campaign targets Latin American enterprise buyers or needs a market-entry and systems-integration angle. The partnership is a company announcement; the deployment difficulty behind it is the part a sponsor can build content around.
A short post that shows how Moura talks about the agent category's trajectory and where CrewAI sits in it.
Read more
A Gartner conference slide places CrewAI alongside Claude Code, Microsoft Copilot Studio, Agentforce, Devin and AutoGen, and, elsewhere on the same slide, ChatGPT, Claude, Gemini, NVIDIA, AWS, Google Cloud, Salesforce, Databricks, Palantir and IBM watsonx.
Moura reads it as validation that the agent category has become real and that CrewAI has earned a place in that conversation, while adding that a slide does not build a company.
The claim about the slide is Moura's own description of a third-party presentation.
For brands planning a campaign
A category-credibility angle a brand could reference when arguing that agent platforms have moved into mainstream enterprise buying. The evidence is a conference slide as described by Moura, not an independent ranking, so avoid presenting it as one.
A short, personal post that explains how the founder works and where the line sits between coding for interest and taking on tasks.
Read more
Moura still writes code on weekends, but stopped taking tasks after learning that working on anything critical makes him a blocker, so ideas go in as pull requests for the team to finish. The post also mentions an experiment that made all of CrewAI's agents run five times faster, with a fix Moura describes as something silly that had been missed, and states a goal that every agent built with CrewAI should be as good by default as the coding agent in Cursor, which Moura says is used daily.
For brands planning a campaign
Best used as supporting color about how the company is run rather than as a stand-alone campaign hook. It works if a sponsor wants a founder-perspective angle on building with AI rather than buying it.
Pair technical ambition with lessons from building a company3
Explain when an AI system is an agent, not just a workflow2
Make AI adoption levels understandable1
Connect autonomy with operational control1
Use evaluation to make model and workflow changes measurable1
Explain enterprise adoption through procurement and integration1
Treat customer feedback as a guide to useful integrations1
Show the work between an approval and a completed outcome1
Explain why enterprise deployment needs local context1
Explain when an AI system is an agent, not just a workflow
Moura distinguishes a fixed process, which follows steps someone specified, from an agent given an outcome that can work out changing steps.
Read more
His accounts-payable example makes the choice concrete: matching invoices to payments is a defined routine, while keeping payables clean requires deciding what to investigate and when to seek sign-off.
Uses paired examples, a concise test question, and a qualification: fixed steps can be cheaper and easier to trust when the path does not change.
For brands planning a campaign
A B2B AI or automation company could use this distinction in a short explainer series for operations leaders. Show one process that should remain a predictable workflow and one where adapting to changing conditions matters; the examples support a practical discussion rather than treating every automation as an agent.
Moura lays out seven stages, from asking a chatbot for an answer to agents coordinating work and a proposed system that oversees other agents.
Read more
Each stage is described through a workplace example, such as a daily sales summary or support-message triage. He presents the final stage as a direction builders are pursuing, not a way companies already operate.
A numbered ladder with a check question at each stage gives viewers a simple self-assessment.
The post says nobody runs a company in the proposed seventh-stage manner yet; it should not be presented as a proven deployment pattern.
For brands planning a campaign
This framework could anchor a video or post series that helps buyers locate their current practices and identify a sensible next step. Keep the stages tied to observable work and distinguish current examples from the aspirational endpoint.
In describing CrewAI’s Agent Control Plane, Moura argues that governance should operate where agents run.
Read more
He contrasts controls in code, which can make policy changes depend on software releases, and an external gateway, which he says cannot inspect or control every part of an agent’s execution. The post lists visibility, cost controls, and runtime policies among the product’s stated capabilities.
For brands planning a campaign
A platform company could commission an explainer about the practical questions that arise as teams delegate more work: what they need to see, limit, pause, or govern. Ground the discussion in concrete controls and clearly identify product claims as claims made in the post.
Use evaluation to make model and workflow changes measurable
Moura describes two distinct evaluation loops: comparing models on quality, cost, speed, and token use; and improving prompts, tool definitions, or routing while keeping models fixed.
Read more
He also warns that an unguarded improvement loop can overfit a small evaluation set or make a workflow worse simply to pass a test.
Starts with a recurring engineering decision, explains the old manual handoff, then separates the solution into two testable loops.
For brands planning a campaign
For an AI infrastructure or developer-tools brief, this supports a practical demonstration of how to compare a model change separately from a workflow change, and how to set safeguards before automating revisions.
Explain enterprise adoption through procurement and integration
Moura recounts a bank prospect asking for observability and memory, exposing the burden of assembling multiple products.
Read more
He uses the story to argue that enterprise buyers may value deployment, metrics, traces, permissions, and governance together, even when a company is not best-in-class in every layer.
A sales conversation becomes the setup for a broader argument, ending in a question about the buyer’s current stack.
For brands planning a campaign
A B2B technology brand could use this as a buyer-side case discussion about the gap between a compelling architecture and what it takes to deploy and support it in an enterprise. The specific procurement story gives the topic a recognizable decision point.
Treat customer feedback as a guide to useful integrations
Moura says his team built access to thousands of tool integrations, then learned that customers wanted a private registry for tools their own teams had already built.
Read more
He draws a distinction between broad public availability and fit with a company’s existing systems, authentication, and data rules. He also frames standards as a way to preserve the ability to change vendors later.
Uses a reversed expectation—customers wanted fewer, more relevant integrations—to make the lesson memorable.
For brands planning a campaign
An integration, API, or enterprise-platform company could use this example to discuss how to prioritize connectivity: ask what customers already use, what must remain governed, and how to avoid locking them into a single vendor.
Show the work between an approval and a completed outcome
A post about credit decisions points to the operational work that can remain after approval: missing documents, misdirected paperwork, and follow-up across people and systems.
Read more
Moura argues that status tracking and coordination can become bottlenecks as volume grows, while also proposing that a person should remain involved where appropriate.
For brands planning a campaign
A workflow or financial-technology brand could build a process-focused case study around where an agent might coordinate handoffs and where a human should still decide. The post supports examining the whole process, not only automating the initial decision.
Explain why enterprise deployment needs local context
In a post about a CrewAI partnership with DXC Technology, Moura says moving from prototypes to production involves choosing business processes, integrating with existing systems, security, governance, and ongoing operations.
Read more
He also argues that international expansion requires local teams, language, and partners who understand the market.
For brands planning a campaign
This suits a B2B services or enterprise-technology brief focused on implementation rather than product demonstrations alone. A campaign could examine what local delivery changes in a real deployment, while keeping partnership claims attributed to the post.
Pair technical ambition with lessons from building a company
Moura alternates product and market arguments with personal accounts of building CrewAI.
Read more
He describes learning not to take critical engineering tasks that could make him a blocker, and says customer feedback corrected his assumptions about integrations. Another post treats industry recognition as encouraging while noting that a slide does not build a company.
Uses first-person anecdotes and a clear contrast between expectation and what happened.
For brands planning a campaign
For founder-led content, these posts offer a more reflective format alongside technical explainers: show a mistaken assumption, what changed, and the operating lesson. This can make a campaign more specific than a stream of product announcements.
João Moura's visible brand relationships revolve around CrewAI, the company he founded and leads.
Read more
The reviewed posts also name DXC Technology as a global-expansion partner, preview a joint session with Capgemini, note that Gartner placed CrewAI on an analyst technology map, and record daily use of Cursor. No sponsored posts, affiliate arrangements, or investment relationships appear in the material reviewed.
CrewAI
Owned business
Moura describes CrewAI as the company he built and leads, and the posts are written from the position of its founder.
Read more
He announced general availability of the Agent Control Plane on CrewAI AMP, described the earlier Crew Studio launch, and shipped an experimental open-source evaluation CLI that benchmarks models against a customer's own baseline. The posts report internal CrewAI practices, such as an agent that checks deal activity each morning, which supports founder-level involvement rather than outside commentary.
Moura announced a partnership between CrewAI and DXC Technology to help companies across Latin America move agentic AI from pilots into production.
Read more
Moura cites DXC's scale, around 115,000 people across more than 70 countries, and frames the partnership around deployment, integration, governance, and post-launch operations in complex enterprises. The post presents a named, disclosed corporate partnership rather than a personal sponsorship.
Moura invited readers to a session on November 9 at 1 pm ET with Arindam Choudhury of Capgemini, working through a credit decision from approval to funding and deciding which steps an agent could own.
Read more
The post documents a joint public session; it does not state any broader commercial relationship between CrewAI and Capgemini, so the scope is limited to that announced collaboration.
Moura shared a photo from a Gartner conference where CrewAI appeared on a technology map alongside vendors such as Claude Code, Microsoft Copilot Studio, Agentforce, Devin, and AutoGen.
Read more
He treats the placement as validation that agent systems have become a recognized enterprise layer, while noting that a slide does not build a company. The relationship is analyst recognition, not a disclosed commercial engagement.
João Moura's work is easiest to explore as a ladder of questions about how much judgment to hand to software.
Read more
Begin with the accounts-payable contrast between a fixed workflow and an agent given an outcome, then read the seven-stage AI adoption ladder for a practical way to locate a team's current habits.
For the buying side, turn to the posts on enterprise deployment, integrations, procurement, and governance, where Moura argues that production support matters more than prototype demos.
A separate thread covers building a company: coding as a source of ideas, staying out of critical tasks, and treating recognition as encouragement rather than proof.
Readers who want adjacent teaching can compare Othmane Khadri on AI-assisted GTM experiments, Stephen G.
Pope on agent-versus-workflow design for automation builders, and Max Thilén on AI-native marketing operations and orchestration.
Both use hands-on examples to show how agents can change routine commercial work rather than just discuss AI in the abstract.
Read more
Khadri describes wiring an agent into Claude Code to review lost deals in a CRM, cluster objections, score open deals, and draft outreach, then building an open-source agent to handle AI-generated comments. The teaching method is similar: start with an operational annoyance, show the agent workflow, and end with a practical invitation.
Khadri's examples sit closer to GTM, sales operations, and social-platform mischief, with a playful tone and a focus on open-source experiments. Moura's material reaches further into agent platform strategy, runtime governance, model evaluation, enterprise procurement, and the operating lessons of running a company.
Both explain the distinction between a rigid workflow and an agent that can be handed a goal.
Read more
Pope uses an accountability coach and an outreach tool to show why step-by-step automation becomes brittle when every node has to fire perfectly, and argues for building one useful capability before adding complexity. That is close to Moura's accounts-payable example and his AI adoption ladder.
Pope's teaching is aimed at people building or maintaining automations, often around tools like CRM, Telegram, spreadsheets, and to-do lists. Moura writes more often from the perspective of an AI platform company addressing enterprise deployment, governance, and multi-agent orchestration.
Both treat agent workflows as practical operating systems for teams, not futuristic demos.
Read more
Thilén lays out how an AI-native marketing team can build self-improving loops, orchestrate sessions, choose models by task, make agents prove their work, and measure quality. Moura makes analogous arguments about evaluation loops, model selection, governance, and keeping humans in the right places.
Thilén's examples concentrate on marketing operations, content production, ad creative, and AI-native team design. Moura's examples concentrate on agent platform engineering, enterprise integration, procurement, and production deployment of agentic systems.
João Moura's evidence supports a placement decision at one platform: his own LinkedIn account shows 53,070 followers as of 8 October 2026, on the same account where the supplied posts were published.
Read more
That is a meaningful single-channel reach for a B2B software brief.
It does not describe who those followers are.
What the supplied material actually shows is content appeal, not measured audience composition.
The posts are written for people who run or buy agent systems inside companies: they use operations examples (accounts payable, support triage, credit decisions moving from approval to funding), enterprise buying language (procurement, security reviews, governance, vendor count), and direct questions back to builders and founders.
For your agentic-workflow brief, that pattern points at an operations and technical leadership readership — but it is an inference from the posts themselves, not an audience breakdown.
No platform demographic export and no audience survey were supplied, so age, geography, gender, employer, job title, and seniority remain unverified.
No video or other format performance data was supplied for this section, so it cannot say which topics, lengths, or formats performed better.
Before committing, ask Moura for recent results on comparable work: performance for explanations of the agent-versus-workflow line and the AI maturity ladder, plus any audience survey he holds.
If he can share platform demographics, use those for geography, age, and gender; for employer, occupation, and seniority, ask specifically whether he has audience survey results or another documented source.
LinkedIn followers
53,070
LinkedIn profile (joaomdmoura), same account that published the supplied posts · 2026-10-08
No creator sponsorship offer, media kit, rate card, or brand-deal inquiry route appears in the material reviewed for this section.
Read more
The partnership activity that is documented sits at the company level inside CrewAI's own business: a partnership with DXC Technology to help enterprises across Latin America move agentic AI from pilots into production, described as a company of around 115,000 people operating in more than 70 countries, and a joint live session on 9 November with Arindam Choudhury of Capgemini walking through a single credit decision from approval to funding.
Both are CrewAI programs, not placements on Moura's own account, and neither publishes a sponsorship specification, deliverable list, or application process.
For a buyer, that means there is no published inventory to brief against: no stated formats, pricing, disclosure terms, or scheduling commitments.
The practical reading is that a partnership involving Moura is more likely a business-development conversation tied to CrewAI's product and enterprise deployment work than a standard creator sponsorship.
The operational ground the posts cover — agent governance, cost controls, data handling, integration with systems an enterprise already runs — is the context any proposal would need to fit.
Whether a sponsorship route exists is not documented here; treat that as a research gap to close before committing budget.
The supplied brief - a B2B automation or AI platform that wants an explainer for operations leaders on agentic workflows - is a strong content fit.
Read more
João Moura's clearest material is exactly this kind of translation: the accounts-payable example separates a fixed workflow from an agent handed a goal, and the seven-stage AI ladder gives operations leaders a self-assessment with a test question at each level.
The posts use enterprise buying language such as procurement, security reviews, governance, and vendor count, and operations scenarios such as support triage, credit decisions, and accounts payable.
That matches an audience of operations and technical leaders.
The material can carry a sponsor through several formats: a short video using the steps-versus-goal test, a post series with one stage per episode, or a buyer-education segment on why deployment and governance matter after the demo.
The commercial format is the limit.
No creator sponsorship offer, rate card, or media kit was found in the reviewed material, so a buyer should treat any placement as a business-development conversation rather than a standard creator booking.
Audience demographics and prior video performance were also not supplied.
When to choose a specialist
A consumer mass-market brief - for example a direct-to-consumer app or retail brand looking for high-volume short-form video from a lifestyle influencer to drive broad awareness - is a poor fit on the supplied evidence.
Read more
The content is written for people who run or buy agent systems inside companies: accounts payable, support triage, procurement, security reviews, observability, and evaluation harnesses.
There is no supplied lifestyle, entertainment, or consumer-product material.
The only measured audience figure is 53,070 LinkedIn followers on one business platform as of 8 October 2026.
That figure does not show age, geography, gender, employer, job title, seniority, or whether the audience buys consumer products.
No video performance data, platform demographic export, or audience survey was included, so a buyer could not predict reach or conversion for a consumer campaign.
The reviewed material also contains no published sponsorship formats, pricing, or scheduling commitments.
A consumer brief could still be tested, but it would be an unproven use of this creator's known strengths.
A worked campaign example
Brief: a B2B AI automation platform wants a short video or post series explaining agentic workflows to operations leaders, including the steps-versus-goal distinction and the AI maturity ladder. Angle: 'Steps or a goal?' - a three-part series built from material already in the reviewed LinkedIn posts.
Episode 1, 60 to 90 seconds: open with the question 'Did you hand your AI a list of steps, or a goal?' Use the accounts-payable example.
On screen, show two columns.
Column one: a system reads every invoice, matches it to a payment, flags anything over 15 days overdue, and drafts the follow-up, in the same order every day.
Column two: the system gets one sentence - keep payables clean, flag anything overdue or mismatched, and tell me what needs sign-off.
The first is a workflow; the second is an agent because nobody wrote out the steps.
Close with the test: did you give it steps, or a goal?
Episode 2, post series: run the seven-stage AI ladder one stage per post or short video, with each stage's test question.
Start with the Prompter and the Monitor because most operations teams will recognize themselves there.
Move through the Delegator and the Manager, where the goal-setting line appears.
For an enterprise sponsor, the strongest stopping point is the Conductor and Ringmaster, where agents hand work to each other and connect alerts a person would otherwise connect in their head.
Label the final Voice Memo CEO stage as aspirational because the post itself says nobody runs a company that way yet.
Episode 3, buyer education: use the Agent Control Plane argument to explain what a company needs once it delegates more work - visibility across deployed automations, cost control by model and agent, runtime policies, and the difference between controls in code, a gateway at the perimeter, and controls inside the execution loop. Treat the product details as CrewAI's own launch claims, not independent test results.
Call to action: ask viewers which stage their team is on, and offer a one-page self-assessment that uses the post's test questions.
The sponsor can attach a practical next step, such as a workshop on moving one process from fixed steps to a delegated goal.
A reviewer should check that each episode names a specific operation, uses one test question, and keeps product claims attributed to the source.
Missing evidence to close before booking: recent video performance, audience demographics, and whether a paid creator sponsorship route exists at all.
Pope if the campaign should teach teams how to start building a useful AI tool without overengineering it.
In a post about an accountability coach, he shows how a rigid chain of steps can become brittle, then explains the agent alternative using memory, tools and a system prompt.
Another post recommends starting with one painful task and adding complexity only after it works.
Those examples make him a closer fit than João Moura for a hands-on beginner series about workflow design and implementation.
The supplied examples focus on building and simplifying individual tools, not on João Moura’s broader operations framing: the steps-versus-goal distinction, a staged AI maturity ladder, and enterprise concerns such as governance and deployment. A sponsor seeking a sequence that helps operations leaders assess organizational readiness may find João’s material more directly suited to that brief.
Choose Othmane Khadri if the same campaign should center on practical sales and go-to-market workflows.
Read more
One post describes using an agent to review lost CRM deals, group objections and draft follow-ups; another describes an agent that responds to unwanted AI-generated comments. These concrete sales and outreach examples could make the campaign more relevant to revenue teams evaluating agentic workflows in their own work.
The supplied examples are narrower and more sales-oriented than João Moura’s operations-leader material.
They do not offer the same broad explainer structure as João’s seven-stage AI ladder or his accounts-payable example contrasting fixed steps with a delegated goal.
The reviewed sample also does not establish whether Othmane offers paid creator partnerships.
Choose Max Thilén if the buyer task is specifically to explain how a marketing team can work with AI.
Read more
His post lays out operating practices including shared business context, connecting agents to marketing systems, measuring quality, setting safety boundaries and keeping human checks. It also gives concrete examples, such as building a targeted account list from conference attendees and finding usable clips in a podcast archive.
That material is aimed at marketing operations rather than the broader operations audience in João Moura’s brief.
It offers a detailed team playbook, but not the same staged explanation of agent maturity or direct workflow-versus-agent test.
João is the more natural fit if the campaign needs to explain those concepts across business functions.
This profile draws on ten LinkedIn posts supplied from João Moura's account.
The posts cover agent design and model evaluation, AI adoption stages, a governance product launch, a technology partnership, integration strategy, credit operations, and company building.
The posts carry no publication dates in the reviewed material; the account snapshot is dated 8 October 2026.
Statements about company metrics, customer conversations, a conference slide and internal experiments are Moura's own and were not independently verified.
Comparisons with three other creators rest on their posts dated May to September 2026.
The reviewed material does not establish a signature book, show, or earlier career history, so any such work would need to be checked against primary sources before it is described.
Recent engagement
Across their last 8 LinkedIn posts, the median post got 121 reactions and 12 comments. Checked 2026-11-02.
