From Cargo to Cursor: How Modern Professionals Are Using AI Workflows to 10X Productivity

The inaugural Rifff session; where real AI adoption happens

In a candid 30-minute group discussion among six professionals (product marketers to growth engineers), a clear picture emerged: the AI tools that matter aren't the hyped ones on LinkedIn. They're the practical, under-the-radar applications transforming teams today.

This was Rifff's first session; an experiment in bringing together thoughtful professionals at similar career levels to discuss what's working. No performances. No hype. Just real talk about AI workflows, productivity tools, and the reality of adoption in 2025.

The sleeper tools nobody's talking about

Cargo; LinkedIn outreach on steroids

Jared Waxman, founder of the Go-to-Market Engineer, shared a workflow that caught everyone's attention. One of his course instructors built an automation in Cargo that goes beyond typical LinkedIn outreach.

The workflow works like this:

Identify people talking about specific topics on LinkedIn
Find people reacting to or commenting on those posts
Filter those commenters through ICP (Ideal Customer Profile) matching
Send AI-generated but human-sounding direct messages
Route responses to specific sales reps based on engagement patterns
Feed responding conversations into Slack channels by department

"It was way more complicated and sophisticated than anything I'd built," Jared admitted. The key insight? Context is everything. Cargo, compared to Clay or Zapier, excels at building workflows that feel human because they maintain rich context throughout the automation.

Whisper; the AI dictation tool changing how people input context

One of the most surprising "sleeper tools" mentioned wasn't even new—it was Whisper, OpenAI's transcription model.

The insight that stood out: Typing limits your context, but speaking frees it. One participant explained that while they struggled to input complex context by typing, speaking naturally into Whisper translated their thoughts into high-quality prompts. It's like talking to a coworker, but with AI.

This matters because the quality of your input directly determines the quality of AI output. Better context = better results.

Custom GPTs for ethnography; understanding communities at scale

Jonathan, a product marketer also using AI to understand developer communities, revealed how he's cracking Reddit—a notoriously hostile space for marketers.

His approach: Digital ethnography using custom GPTs. Before AI, understanding Reddit subreddits took months of manual research. Now? He can analyze 30 posts and extract community insights in under a day.

The challenge: Reddit communities will roast outsiders. Success requires understanding written and unwritten rules. His custom GPT performs ethnographic analysis—studying the culture, language, and norms before any outreach happens.

This is how smart GTM teams are thinking about communities in 2025: not as distribution channels, but as cultures to understand deeply.

The Cursor revolution; from developer tool to marketer's best friend

The most compelling revelation came from Harsh Sharma, a product marketer who rebuilt his entire workflow around Cursor and Claude.

"Everything I do; newsletters, content, competitive intelligence; runs through Cursor," he explained. This includes:

Building competitive intelligence CLIs that scrape competitor data
Creating LinkedIn post analyzers that study competitors' successful posts
Writing newsletters powered by AI-guided research
Building open-source tools shared on GitHub

The insight: Modern professionals aren't choosing between ChatGPT, Claude, and Cursor based on arbitrary preferences. They're choosing based on workflow fit. Cursor works best when you want to see and edit code. Claude is easier for beginners. Both often use the same underlying models, but the interface and integration matter enormously.

The real power of Cursor for marketers? It becomes a second brain. One person can now do what previously required a small team of researchers and writers.

Beyond hype; the real role of AI agents

Harsh also walked through Anthropic's new Skills feature; an under-the-radar release that could reshape how teams use AI.

Skills vs. Custom GPTs; why this matters

Traditional approaches dump all context into one chat:

Brand guidelines
Tone of voice documents
Previous case studies
Brand strategy

The problem? Context windows are finite. LLMs lose focus when overwhelmed with information, degrading output quality.

Skills are different. They're reusable instructions bundled with specific documents. Instead of stuffing everything into one prompt, you create:

A "Brand Voice Skill" with tone guidelines
A "Competitive Intelligence Skill" with market research
A "Content Writer Skill" with case studies
An "SEO Skill" with keyword research

The agent accesses these only when needed, maintaining focus and producing better results.

Harsh demonstrated this with a marketing brief example. When agents had skills with proper context management, their output went from generic to genuinely valuable; personalized, localized, and on-brand.

This is the future of AI adoption: Not complex N8N workflows or dozens of integrations, but intelligently structured context that agents retrieve when needed.

Workflows that work

The consensus around workflow tools (N8N, Zapier, Cargo, Air Ops, Clay) was clear; they're all powerful, but simplicity wins.

One marketer noted: "I've never used N8N even though I have access. It looks too complex. I can do the same things faster with Cursor and Python scripts running cron jobs."

The real lesson: Choose your tool based on how your team thinks, not on feature counts. Someone who thinks visually prefers Clay's table interface. Someone who prefers workflow builders uses Cargo. Someone comfortable with code uses Cursor or Python.

All can accomplish nearly anything with APIs and webhooks. The difference is ease and team comfort.

The job security question nobody's comfortable asking

The conversation turned serious when someone asked: "How do you convince your team to embrace AI when everyone's secretly worried it'll replace them?"

The responses were honest:

Perspective 1 (Acceptance): "AI will probably replace some roles in 10-20 years. But learning it now means you won't be among the first laid off. That's your best job security."

Perspective 2 (Evolution): "Your job isn't getting replaced. Your job is evolving. The question is whether you evolve with it or get left behind."

Perspective 3 (Corporate Reality): Angela, working at Okta (a security company with 18,000+ employees), shared the enterprise perspective: "Getting approval to use Claude takes forever. It's slow. But we're pushing it because productivity gains are real, and yes, there's concern about what automation means for headcount."

The real issue: AI adoption in large companies isn't about capability gaps. It's about organizational friction and fear.

What nobody tweets about; the reality of AI content

There was refreshing cynicism about LinkedIn hype culture.

One comment: "All this LinkedIn content about 'I built 87 agents and fired my marketing team' is bullshit. People are just chasing engagement."

The truth? Using AI effectively requires:

Understanding your domain deeply
Knowing what good output looks like
Being able to refine and iterate
Having realistic expectations about what AI can do

No tool creates value in a vacuum. Human judgment remains irreplaceable.

The participants; what they actually do

Jared Waxman: Founder of Go-to-Market Engineer. Previously worked at startups. Now trains GTM teams on AI-driven workflows using tools like Cargo.
Harsh Sharma: Product marketer turned "AI engineer." Uses Claude, Cursor, and custom AI applications. Shares actionable AI insights on LinkedIn and X.
Jonathan: Product marketer studying developer communities. Uses custom GPTs and workflow tools to understand Reddit at scale.
Angela: Team lead at Okta. Uses ChatGPT for brainstorming, content generation, and strategic planning. Navigating enterprise AI adoption.
Mara: Product marketer. Focused on practical, actionable AI use cases rather than theoretical discussions.
Drew: Founder of Rifff. Organizing these sessions to cut through the hype.

Key takeaways; what this means for your work

1. Context engineering is the new skill. The difference between mediocre and excellent AI output isn't the model; it's the context you provide. Structure your documents, clarify your instructions, and use tools like Skills to manage attention.

2. Sleeper tools matter more than hype tools. Cargo, Whisper, custom GPTs, and Cursor are powerful because they solve specific problems elegantly. Pick tools based on your workflow, not your feed.

3. Simplicity beats complexity. You don't need N8N with 47 integrations. Cursor, Python, and strategic Zapier flows often get you 90% of the way there with 10% of the complexity.

4. Learn it or get left behind. Whether you're worried about job security or excited about productivity gains, upskilling in AI is no longer optional. It's a career survival tactic.

5. LinkedIn content is theater. Filter out the hype. Focus on people doing real work in your domain. Seek out actionable, honest takes over engagement-bait threads.

6. Corporate adoption is slower than you think. Even at large, tech-forward companies, AI tool adoption faces friction; security reviews, approval processes, risk aversion. This creates opportunity for smaller, more nimble teams.

7. The job isn't being replaced; it's evolving. People worried about AI replacing their jobs are asking the wrong question. The real question is; Will you evolve faster than the technology?

What Rifff is about

These sessions exist because the hype around AI drowns out honest conversations. Rifff brings together thoughtful professionals at similar career levels to discuss what's working.

No stage. No performance. Just people comparing notes on:

What tools save time
What's overhyped garbage
How to convince your team to adopt AI responsibly
What this means for careers and companies

If you want to join future Rifff sessions and have these conversations with peers who are doing real work, [sign up here].

What sleeper AI tools are you using that never get mentioned online? What's your honest take on the hype? The next Rifff session is coming, and we'd love to hear your thoughts.