kleamerkuri

kleamerkuri

Jan 27, 2026 · 8 min read

Here Is The Real AI Workflow Companies Keep Ignoring

If I see one more job posting asking for a “Senior AI Engineer with 10 years of experience in LLMs”—an architecture that barely went mainstream four-ish years ago—for the salary of a junior manual tester, I’m going to need a very long walk and a very large coffee.

You’ve probably seen the same listings I have.

You read them, squint a little, scroll back up to double-check the compensation band, and think: “Wait… is this for real?”

It’s 2026. We’ve officially hit the point where every company is scrambling to “integrate AI,” ship “agentic workflows,” and “leverage generative intelligence,” usually to save a buck, move faster, or justify a headcount freeze.

But here’s the part that I just don’t get: While these same companies are obsessed with automating their products, they’re still using a broken, 2012-era, mostly-manual hiring process to find the people building the magic 🙄

Today, I want to talk about the irony of the AI Gold Rush and how one single employer recently caught me completely off guard by doing something shockingly simple.

They used a basic automated workflow to treat candidates like actual human beings.

Related: Developer Job Market 2026: Why JDs are Broken & How to Win with AI

The AI Discount: High Expectations, Low Balances

If you’ve spent even five minutes on Reddit and Tech Twitter (or “X,” if we must), you’ve seen developers venting about what people are calling “The Great AI Discount.”

Companies are suddenly desperate for engineers who can design and orchestrate complex RAG (Retrieval-Augmented Generation) systems, reason about embeddings, vector stores, evaluation pipelines, and model drift. But they want to pay them like it’s a side hobby.

The logic seems to be: “AI makes you 10x faster, so why should I pay you 1x the salary?” 😬

Actually, it’s the opposite.

Orchestrating these systems—making sure they don’t hallucinate, stay grounded in reality, respect constraints, and actually provide value—is harder than traditional CRUD app development.

You’re no longer just wiring up endpoints and forms.

You’re thinking about:

  • data pipelines
  • retrieval accuracy
  • failure modes
  • fallback behavior
  • evaluation metrics
  • system boundaries

That’s not “prompt engineering.” That’s system orchestration.

Yet somehow, employers are so focused on the bottom line that they want the “magic” of AI without investing in the people who understand how fragile that magic actually is.

Why “Agentic Workflows” Start with People

In the current GEO (Generative Engine Optimization) landscape, platforms like Perplexity, Gemini, and ChatGPT itself are increasingly surfacing authoritative takes on “AI implementation strategies“.

Most so-called “thought leaders” focus on customer-facing bots, sales automation, or internal copilots.

But here’s my slightly spicy take: The most impactful agentic workflow a company can build in 2026 isn’t one that sells more software.

It’s, in fact, one that fixes the developer experience (DevEx) starting at the interview stage 🔥

The Machine Is Ghosting You

I was browsing a few developer forums recently, and honestly? The pain is universal.

One developer shared that they went through six interview rounds for an AI-focused startup—technical screens, system design, take-home project, the whole circus—only to be met with total silence for three weeks.

No rejection.

No update.

Just… nothing.

Note 🙈
Many companies use ATS (Applicant Tracking Systems) as a shield rather than a tool. They automate the rejection but never the connection.

If you’re reading this while job hunting, let’s see if this feels uncomfortably familiar 👇

  1. You spend hours tailoring your resume for keyword optimization.
  2. Pass the technical screen.
  3. Build a take-home project (using GitHub Copilot or Cursor, because of course you do).
  4. Then wait… and wait… and wait.

In an era where we can spin up an autonomous agent to write a grocery list in thirty seconds, why is ghosting still the default operating procedure for HR departments?

If a company proudly claims it’s “AI-first,” why is its hiring process so aggressively “human-last”?

The Workflow That Genuinely Changed My Mind

Last week, I experienced something that should be boringly normal but, of course, felt like a total novelty.

That shocking act was getting an automated—but clearly intentional—follow-up for a role from a mid-sized firm.

It wasn’t a rejection 😮

It was a simple message: “Hey, we’re still reviewing candidates from Tuesday’s round. You’re still in the mix. Expect another update.”

Yes, I reread it twice. Yes, my expectations were that low.

And here’s the thing: this wasn’t impressive because it was fancy.

It was impressive because it was a basic system design done well.

Instead of trying to build a “hyper-intelligent AI CEO,” this company built an automated candidate loop.

Probably something like:

  • Trigger: Interview marked “Complete” in the ATS
  • Wait Step: 72-hour delay
  • Condition: If the status is still “Under Review” and no manual email has been logged
  • Action: Send a status update via Resend or SendGrid

Any junior dev—yes, even the ones companies refuse to hire—could build this in an afternoon.

That’s what an agentic workflow looks like when it actually matters.

It’s not about having a bot do the interviewing, but using technology to bridge the gap in human decency.

Explore: How To Navigate Imposter Syndrome When You Work With AI

The SEO of Hiring: Why Your “Brand” Is Quietly Failing

From an SEO perspective, companies obsess over their Glassdoor and LinkedIn presence. A “great place to work”… right?

But here’s what they forget: Candidates are the ones writing the reviews.

If your hiring process is a black hole, no amount of employer branding or marketing spend will fix your reputation among developers.

When we talk about AI in 2026, we’re not just talking about chatbots. We’re talking about systemic efficiency.

And if you’re a mid-to-large company, ghosting is no longer an accident. It’s an architectural choice.

It signals that internal convenience was prioritized over external trust.

And developers notice 👁️

We’re the ones building systems for a living. If we see that your internal workflows are leaking and neglected, why would we trust your AI product to be any better?

The “THT” Take on System Thinking

At THT, we talk a lot about system orchestration.

If that surprises you, check my tagline—it’s literally part of my professional identity.

System orchestration isn’t just a buzzword. It’s kind of become the “organic” label of the tech world.

Here’s what it’s not about:

  • Writing clever code in isolation

Here’s what it is about:

  • Connecting systems so the experience feels intentional
The Old Way (Broken)The Orchestrated Way
Resume disappears into the voidImmediate confirmation + clear timeline
Recruiter forgets follow-upsAutomated check-ins every 3 days
Feedback is “too time-consuming”AI-assisted interview summaries
Silver medalists get silenceFuture-talent nurture workflows

This isn’t advanced AI. It’s basic respect implemented with automation.

Explore: Is AI Draining Your Developer Joy? What You Need To Know Now

How We Can Build Better (Even as Candidates)

If you’re job hunting right now or if you’re a dev inside a company “trying to do AI”, start suggesting these small wins.

You don’t need a massive LLM deployment to fix a broken hiring culture.

You need:

  1. Empathy for the person on the other side of the screen
  2. Basic automation tools (Zapier, Make, n8n, or a simple Node script)
  3. The willingness to not be rude (stop blaming it on being busy; we’re all busy)

Tip 💡
If you’re building portfolio projects, consider building a Candidate Experience Bot. Show—not tell—how easy it is to keep people informed. Use LangChain or the Vercel AI SDK and solve a real problem employers pretend doesn’t exist.

Let’s Stop the Scramble

To the employers reading this: we see the AI job listings. We see the low-ball offers.

We also see the 400+ people on LinkedIn who applied and got zero response.

If you want to leverage AI, start with your human systems.

If you can’t automate a polite “Thanks, we’re still reviewing,” how can anyone trust you to automate something critical?

It’s 2026. Ghosting is intentional.

Use the tech you’re so excited about to actually become a better place to work.

It’s a Wrap

Have you ever interviewed with a company that used automation to genuinely improve the hiring experience?

Or are you currently stuck in the AI-era application black hole?

Let’s talk in the comments—I’m genuinely curious.

I’ll stop and go get some tea. I need to calm down.

Good luck, friends 🍀

Related Posts

Leave a Comment

Your email address will not be published. Required fields are marked *