Founder & Lead ML Engineer @ AI Flow | Ex-Google & Bloomberg | Building AI Agents that scale | 10 years in AI & ML | Building AI Flow, an AI & ML focused consultancy that designs, builds, and scales end-to-end AI products for startups and mid-market companies.


> What we do at AI Flow

- AI consulting - understand how to make most of your data and benefit from the Agentic AI wave.
- AI Strategy & Roadmapping - cut through hype, map the fastest path to ROI
- Custom ML & Gen-AI systems - from data pipelines to fine-tuned LLMs and Agents
- Data platforms - ETL, ingestion pipelines, document understanding end to end solutions
- Embedded AI Teams - senior engineers who plug in and deliver from day 1

> Proof over promises

- 45% pipeline cost & time drop for hyper-realistic video at Metaphysic
- 65% faster ML iterations for Google internal ML teams
- Millions saved automating invoice processing for a top-5 Oil & Gas firm
- Built tools used by 200K daily users, with more than 200 request/second at Bloomberg.
- 85% RAG & LLM pipeline cost drop through data cleaning pipelines for law companies.

> Why clients trust us

“I trusted you to deliver a good result, and that’s exactly what happened.”
- Josua Vieten, CEO @ ExoMatter

“Deep technical expertise across the stack. Mihai understands business context and proposes scalable solutions.”
- John Hack, CPO @ Interflexion

“Mihai makes it easy to get projects done. He is pragmatic, ambitious, and takes initiative.”
Paul Craciunoiu, Founder @ Uplift


> Let’s talk about building the AI future

Open to B2B, remote projects. If you need a sharp, senior team to ship real AI, book a 30-min call: antonmih.ai/meet

Mihai Anton

You wouldn't hire a mechanical engineer who only used AI-generated blueprints.

You wouldn't trust a doctor who only used ChatGPT for diagnosis.

Yet many companies hire "AI engineers" who've never debugged a production system.

Tools change. Fundamentals don't.

undefined undefined

Comment

Watch it here: https://www.youtube.com/watch?v=CLuv-X2Fm_s

Machine Learning by Stanford University on Coursera

Cambridge English Advanced English

Founder & Lead ML Engineer @ AI Flow | Ex-Google & Bloomberg | Building AI systems that scale | 10 years in AI & ML

Founder & Lead ML Engineer @ AI Flow | Ex-Google & Bloomberg | Building AI systems that scale | 10 years in AI & ML | Building AI Flow, an AI & ML focused consultancy that designs, builds, and scales end-to-end AI products for startups and mid-market companies.


> What we do at AI Flow

- AI consulting - understand how to make most of your data and benefit from the AI wave.
- AI Strategy & Roadmapping - cut through hype, map the fastest path to ROI
- Custom ML & Gen-AI systems - from data pipelines to fine-tuned LLMs
- Data platforms - ETL, ingestion pipelines, document understanding end to end solutions
- Embedded AI Teams - senior engineers who plug in and deliver from day 1

> Proof over promises

- 45% pipeline cost & time drop for hyper-realistic video at Metaphysic
- 65% faster ML iterations for Google internal ML teams
- Millions saved automating invoice processing for a top-5 Oil & Gas firm
- Built tools used by 200K daily users, with more than 200 request/second at Bloomberg.
- 85% RAG & LLM pipeline cost drop through data cleaning pipelines for law companies.

> Why clients trust us

“I trusted you to deliver a good result, and that’s exactly what happened.”
- Josua Vieten, CEO @ ExoMatter

“Deep technical expertise across the stack. Mihai understands business context and proposes scalable solutions.”
- John Hack, CPO @ Interflexion

“Mihai makes it easy to get projects done. He is pragmatic, ambitious, and takes initiative.”
Paul Craciunoiu, Founder @ Uplift


> Let’s talk about building the AI future

Open to B2B, remote projects. If you need a sharp, senior team to ship real AI, book a 30-min call: antonmih.ai/meet

AI Flow | We build AI powered products | AI, Machine Learning, and Data managed services

AI Flow

Senior Software Engineer | AI consultant | Working in AI before it was cool

Resume

Created a social network for frequent travelers, where they can post experiences, see what others explore, and follow their steps. Take a look to see how it works!

IdeaList

MihaiAnton has 14 repositories available. Follow their code on GitHub.

MihaiAnton - GitHub

Founder & Lead ML Engineer @ AI Flow | Ex-Google & Bloomberg | Building Agentic AI & ML systems that scale | 10 years in AI & ML

AI Flow #2 is up.

This week: how we helped a client cut spending from $10K/month to $2K/month, without touching the model.

After analyzing the situation, we found the problem was the input.

They were feeding entire HTML pages to the LLM (cookie banners, navigation menus, scripts, SEO content). The model was drowning in noise.

The fix was simpler than they expected. I break down what we did and the results in this week's edition.

Every time a new model drops, I get the same wave of messages from clients: should we rebuild? Should we switch? Are we falling behind?

This week it's GPT-5.4. Next month it'll be something else.

The honest answer is always the same: it depends on what you're optimizing for, and most teams haven't written that down anywhere.

- If cost is the constraint and the new model delivers similar quality for less, run the eval and test it.
- If accuracy drives revenue or directly reduces support volume, measure whether the delta is real enough to justify the migration work.
- If you're upgrading because it's newer and that feels like the right move, that's the one reason I'd push back on.

The things that determine whether your system works didn't change this week. Your data quality didn't change. Your evaluation criteria didn't change. The problem your users need solved didn't change.

A better model sitting on top of a broken pipeline is still a broken pipeline. Test it against the metrics you already track. If they move, ship it. If they don't, look at other areas that actually move them, like the data sources feeding the system.

AI Flow #3 is up.

This week: a client came in ready to invest in AI. Budget approved, timeline set.
We told them: not yet.

After analyzing their setup, we found data in 6 different systems. No tagging, no structure, half of it contradicting the other half.

Building on that foundation would have guaranteed a broken system. So, we made a different plan instead.

"How fast can we see ROI from AI?"

Depends on what you're measuring.


- Time savings? Weeks. We automated proposal generation for a construction company. From 4 hours per proposal to 20 minutes, they felt the ROI immediately.

- Revenue increase? Months. You need time to test, iterate, and prove the feature changes behavior.

- Cost reduction? Somewhere in between.


Most companies expect all three at once, immediately.

My advice: focus on time savings first, as it’s fastest to prove.

"Should we rebuild for the new model?"

The problem isn't the question itself. It's what comes next, as most teams usually ask, "Is it better?" instead of "Is it better for what we need?"

The answer depends on what you're optimizing for. Here are 3 examples.

Should we use the latest model?

2026 isn't about chasing what's new. 

It's about building systems that work regardless of which model is trending.

AI Flow #4 is up. This week: how fast can you actually see ROI from AI?

I’m sharing one case we had with AI Flow: We helped a construction company cut proposal time from 4 hours to 20 minutes; they felt the ROI in week one.

But that's time savings. Revenue increase? Cost reduction? Each has a different timeline.

The problem is that most companies expect all three immediately, and that's why projects stall before showing value.

“Can we automate this?” is one of the most common questions I get.

Sure, we can. But that’s not the first thing you should ask.

Before you jump into workflow scripts or n8n hacks, zoom out:

- Does this process actually need to exist?
- How much time or cost will automation really save?
- Do you have clarity on the pain points you’re trying to remove?

Most teams skip that step. That’s how you end up automating messy processes, which only automate more mess.

"We need AI to analyze our customer feedback."

This request came from a head of product at a 200-person company.

What they really had:

- Feedback in 7 different places
- Support tickets in Zendesk
- Feature requests in Productboard
- Sales calls in Gong
- User interviews in Dovetail
- Social mentions everywhere
- NPS surveys in their own tool
- Churn interviews in random Google Docs

Now, they wanted AI to magically connect all of it.

And here's what we told them: AI can't fix a structural problem.

First step: pick one source, the one with the highest signal (for them, it was support tickets because every issue was already tagged and categorized).

We built an AI analysis for that one source, and it worked. From that, they got insights they could act on.

Now they're adding the others. One at a time.

If you have data everywhere, don't start by connecting everything. Start by getting value from one source first.

Client last month: "We just need you to make our chatbot smarter."

Turns out their chatbot was giving wrong answers 40% of the time.

We dug deeper, and we found out that the problem wasn't the AI, but their knowledge base:

- Outdated pricing in 12 different PDFs
- Product specs scattered across Slack threads
- Support answers that contradicted each other

No model can fix messy data. It just makes confident mistakes faster.

What we did next was to spend a couple of weeks organizing their data before touching the AI. Now the chatbot works.

If your data lives in 10 different places with no single source of truth, start there.

A founder told me they spent 6 months building the perfect AI architecture, while never talking to a single user.

They launched last month, and it turns out users needed something quite different. 

Perfect systems built for the wrong problem are just expensive mistakes. I recommend prioritizing talking to users as much as possible to ensure you build what they really need.

A16z recently led a $30M round into Lio, a startup automating enterprise procurement with AI agents.

Sourcing, contract negotiation, spend management. All of it.

I'm not surprised by the funding. Procurement is one of the most automation-ready functions in a mid-size company, and one of the most overlooked.

When I talk to operations leaders about where AI could save them time, they usually mention customer-facing workflows first. Support, sales, marketing. Procurement comes up later, almost as an afterthought.

But when we actually map where decisions are slow and errors are expensive, procurement keeps appearing: requests sitting in inboxes waiting for approval, vendor comparisons done manually in spreadsheets, contracts reviewed line by line every single time.

The pattern underneath all of it is the same: high volume, repetitive structure, clear rules for what good looks like. That's exactly where automation compounds.

The question worth asking isn't whether procurement can be automated. It's which part of your procurement workflow is costing you the most time right now, and whether the process is clean enough to hand to an agent.

If the answer to the second part is no, that's where to start.

Most enterprise clients are not ready to build AI. This would be the wrong first move, anyway.

Very often, they are excited about building AI. Budget approved, timeline set, excited to start building.

Then I look at their setup. Data scattered across 6 systems with no structure. records contradicting each other, no single source of truth.

In such cases, AI won’t be a positive addition. It will only automate the chaos.

So I get back to them with a different plan. Spend two months organizing the data first, establish structure, then build AI that actually works. 

This approach might feel slow when you're excited to start. But it’s faster than building something that breaks and spending the next 6 months trying to fix it.

If everyone you talk to says yes immediately, you're probably not getting honest advice.

AI Flow #5 is up.

This week, I'm exploring the gap between demo and production, using a client situation as an example.

They had a recommendation engine already working in production. Users were getting suggestions, and the feature was live. But it performed at maybe 70% of where it could be.

After digging into the case, I noticed some problems in how it was built. It was rigid, with hard-coded paths, no observability, and every new requirement broke previous flows.

"We built a demo that works great. Now what?"

Now you find what breaks in production. 
What you should prioritize:

- Add monitoring
- Add fallback logic
- Test with real users, not perfect test cases.

Keep in mind that the gap between demo and production is where most projects fail.

AI Flow #6 is up.

This week, I'm looking at procurement, one of the most automation-ready functions in most businesses, and the one that rarely makes it onto the AI priority list.

When we map where decisions are slow and errors are expensive across the companies I work with, procurement keeps appearing. Vendor comparisons rebuilt from scratch every time. Purchase requests waiting three days for a two-minute approval. Contracts reviewed line by line for suppliers the company has worked with for years.

Nobody designed it this way. It just accumulated.

This edition covers why this happens, what to look for in your own workflows, and where to start if the process isn't clean enough to automate yet.

Pavle Lucic

Senior UI/UX Designer @ Toptal | Prototyping, Branding, UI Development

I am a Senior UI/UX Designer at Toptal, a network of the world's top 3% of software engineering, design, and finance talent. With over 10 years of experience in the design industry, I have a proven track record of creating visually engaging and user-friendly interfaces for in-car navigation systems, SaaS apps, and marketing websites.

My background in business and economics gives me a unique perspective on how to design UX and UI that are not only aesthetically pleasing, but also functional, efficient, and user-friendly. I am passionate about solving complex problems, collaborating with diverse teams, and learning new skills. I am also Adwords Search Certified, which demonstrates my ability to optimize the online presence and performance of my clients. My goal is to help tech companies increase their revenue by improving the user experience of their products.

Senior UI/UX Desiger

2 tools that will replace entire product teams in 2026

(designers can do the development now)

1/ Claude → Your thinking layer 
↳ Industry analysis in 15 minutes 
↳ Competitive breakdowns in 20 minutes
↳ User journeys mapped instantly
↳ Feature priorities without PMs 
↳ Strategy without strategists

2/ Figma → Your design AND development layer 
↳ Systems, not screens 
↳ Components that scale 
↳ Figma to code pipelines 
↳ AI-assisted building 
↳ Ship without developers

Old model: 

PM + Strategist + Designer + Developer 4 people. 
4 weeks. 4 handoffs.

New model: 

Designer + Claude + Figma 1 person. 
2 days. 0 handoffs.

Strategy. Design. Development. One designer.

The handoff era is over.

Repost this ♻️ if you're building this way.

P.S. Are you still handing off or already shipping?

Senior UI/UX Designer

Senior UX Designer

UI / UX designer & UI Developer

Visual Designer

UI/UX Designer

Senior UI/UX Product Designer | UX Strategy | Figma

English

Serbian

Every designer should SAVE this Accessibility Checklist: 

1. Images include alt text
2. Buttons have clear labels
3. Body text is 16px or larger
4. Text contrast is at least 4.5:1
5. Click targets are minimum 44px
6. Forms show clear error messages
7. Icons are never used without labels
8. Navigation works with keyboard only
9. Animations can be reduced or disabled
10. Focus states are visible for keyboard users

Follow this checklist, and you'r already ahead of 90% of products online.

Accessible design isn’t complicated.
It’s just intentional.

Save this post for your next UI review.

(Repost for other designers ♻️)

P.S. Which one do you check last before shipping a design?

Here's a proven way to stand out as a UX designer in 2026.

(Without adding another certification to your wall)

Stop focusing on "designing screens". 
Focus on "designing outcomes".

When you present work to stakeholders?

Your job is not to show how pretty it looks. 
It's to show why it matters.

This is the main "metric" of everything you design"

1. User pain points → Connect them to cost
2. Redesigns → Show the numbers behind them
3. They need a solution → Be their strategic partner

Stop hiding behind "I'm just the designer".

Share the "why" → Not just the "what"
This builds irrevocable trust.

Don't just push pixels. 
Push thinking. 

Make sure your seat at the table stays. 

When restructuring comes? 
(And it will) You'll still be there.

Pretty? No. → "Impactful". Always.

(Repost this for your network ♻️)

P.S. Do you notice some designers hiding behind Figma files? Avoiding the business conversation entirely?

Most recruiters ask designers:

“Do you have experience in our specific industry?”

I think the better question is:

Can they understand the audience?
Can they define the real problem?
Can they bring a fresh point of view?

They’re hired to:

Think deeper.
Observe better.
Design smarter.

Industry knowledge is useful. Sure.

But if that’s your no. 1 hiring filter...

Chances are: you’ll get familiar ideas, safer choices, and the same patterns your competitors already use.

Sometimes, the best designer for the job is the one who is not stuck inside the industry bubble.

Because fresh eyes catch what insiders stop noticing.
Great design is not about knowing the category best.

It’s about understanding people best.

P.S. What matters more: industry experience or design thinking?

Every time I share this advice, some designers get mad. 

“Users do judge products visually”
“The UI matters more than ever now”
“Brand perception starts with visuals”

My advice: Stop using UI and UX like they mean the same thing.

Chances are: Your interface matters a lot.
Chances are: Your user experience matters even more.

(Fun fact: A product can look amazing and still frustrate users)

So...

If you want better product discussions
If you want smarter design decisions
If you want to solve the right problems

Separate UI from UX. (Thank me later)

P.S. Don’t get mad. ✌️

UX vs UX

Junior designers add design elements. 

But senior designers remove.

One rule that fixed 80% of my designs:

Take your finished screen.
Start deleting things. One by one.

In practice:

That extra icon? Remove.
That extra border? Remove.
That extra shadow? Remove.

Keep going until something breaks. Stop there.

Everything you deleted before that? Never needed.

Took me years to trust subtraction over addition. 
Now it's the first thing I do on every project.

💾 Save this for your next design review.

P.S. What element do you always overuse?

Unpopular opinion: Figma is the last tool you should open.

Not the first.

Here's the process I use for every project:

Phase 1 → Discovery (1 week)
Research. Competitors. Customer Journey.
↳ You can't solve what you don't understand.

Phase 2 → Prototyping (2 weeks)
Architecture. Flows. Wireframes. Testing.
↳ You can't design what you haven't structured.

Phase 3 → Design (5 weeks)
UI Concepts. Final UI. UI Kit. Handoff.
↳ Now you design. Not before.

8 weeks total. 

"But that's too slow."
No. 

Redesigning a product 3 times because you skipped research is slow. 

Arguing with devs because there's no handoff doc is slow. 
Rebuilding flows because nobody tested the wireframe is slow.

Process isn't slow.

Skipping process is slow.

(Save + Repost for others ♻️)

P.S. Do you follow a process or wing it?

For years, they told us:

Learn computers. Avoid manual work.

Now?

AI is writing emails.
Designing logos.
Making presentations.
Even coding.

Meanwhile…

The plumber is somewhere chilling like:
“Good luck teaching ChatGPT to fix your bathroom.” 🙂

Funny times, friends.

The “safe” jobs suddenly don’t feel that safe.
And the jobs people looked down on?
They’re looking very… employed.

P.S. Which non-computer job is winning the AI era? 😂

Being great at design is not enough.

You can be very skilled...and still lose the project.

What clients remember is not just:

1. Your portfolio
2. Your systems
3. Your creative taste

They also remember:

1. How you handled feedback
2. How easy it was to build with you
3. How safe it felt to disagree with you

Design does not happen in isolation. It happens through conversations.

Through trust.
Through people.
Through uncertainty.

The best designers don’t just push pixels. They move people.

P.S. People don’t just remember the final design. They remember what it felt like to get there. 😊

P.P.S.  Design is not just visual. It’s human ♻️

"Good designers have a natural eye for color."

No they don't.

They have a system.

And most of them use the same one:
60-30-10.

60% of the screen is one neutral 
30% is one supporting tone 
10% is one accent

It's simple a ratio.

You could pick ugly colors and still make a clean UI with this rule.
You could pick beautiful colors and still make a mess without it.

The "eye" people talk about? It's pattern recognition.

And pattern recognition comes from systems.
Not talent.

Start with the system.

♻️ Repost or Save if you're tired of the "natural talent" myth.

P.S. Talent or system what matters more? 👇

I've been studying dark UX patterns for a while now.

Here's everything I see on one Temu product card:

"Deal" badge → urgency without a deadline
"17K+ sold" → social proof replacing analysis
1,180 reviews → volume creating false authority
Fake "original" price → anchoring your perception
Cart icon right next to the price → reducing friction
Product image with dimensions → implying precision
Everything above the fold → no scrolling needed to buy
Percentage shown in orange → framing savings over cost

8 design decisions.

But...0 of them are about helping you choose.
All of them are about helping you click.

Feel free to repost this ♻️ 

But my ask is:
Next time you shop online, count the tricks first. Then decide.

P.S. Which one surprised you (1-8)? Anything I missed? 👇

My non-negotiable rules for writing button copy.

(7 micro-copy rules every junior designer needs to know)

Lead with the outcome e.g. "Get Your Free Guide"
"Sign Up" = vague → "Start for Free" = clear value
Keep it to 2-5 words → Longer than that? Rewrite it.
Never use "Submit" → Nobody wants to "submit" anything
Never write what the system does → Write what the user gets
Kill the # 1 objection inside the button e.g. "Try It, No Credit Card"

Read it out loud → If it sounds like a robot wrote it, it's wrong.

Bonus: 
Never use "Submit" on a button because you ran out of ideas. 

If you can't describe the value in 5 words, you don't understand the feature well enough.

Bad CTAs describe the process. 
Good CTAs describe the outcome = That's the whole game.

(Save + Repost for others ♻️)

P.S. Do you follow UX copy rules?

2 years ago I was handing off Figma files.

Today I push code to GitHub.

Same person. 
Same skills. 

One new tool. AI.

I didn't hire a developer.
I didn't go back to school. 
I didn't learn computer science. 

I just started "building."

Backend? AI. 
Databases? AI. 
The scary stuff? AI.

Design, UX, how it feels?

That's still me.
Always will be.

Turns out, a designer who builds the "whole thing"?
Pretty powerful.

♻️ Repost if you know a designer who needs to hear this.

P.S. Which idea have you been sitting on? Tell me below. 👇

Here's something I never post about.

(it starts with a DM I almost ignored)

Last week someone messaged me on Linkedin.

"Hey, do you do branding?"

My first instinct: no.
My second instinct: why not?
My third instinct: LinkedIn says I'm a product designer.

I replied anyway. "What do you need?"

She said: 
"Logo for pilates studio".

I quoted the project. She said yes.
I sketched directions in Procreate on the couch.
I opened Adobe Illustrator. Not Figma. Illustrator.

My iPad on my lap. Coffee getting cold.

She wrote back 4 words.
"Don't change a thing."

And I just sat there.

Because I almost said no.
Because I convinced myself that product designers don't do this kind of work.

I thought:

That it wasn't my lane. 
That I needed to be a "brand designer" first.

That night I scrolled through my files.

Logos I made and never showed anyone. 
Brand identities sitting in folders named "personal" and "side work."

Years of work that never made it to my portfolio.

Because I thought it didn't count.
It wasn't "product design" so it wasn't real work.

That's the dumbest thing I've ever believed.

It's design. 
It counts.

P.S. The coffee was cold. The logo was good. Fair trade.

Most people think AI builders must be engineers.

But the next generation may be designers.

AI can write code.
But AI cannot design the product.

Someone must decide:

• What problem to solve
• How the flow works
• How the system behaves

That is product thinking.

Designers already do this.

P.S. Agree?

Top negotiators use this hack. It's called "echoing."

(here's how designers can use it on every feedback call)

So how it works?
Repeat the client's own words back to them.

That's why it's called the "Echo" method.

People trust their own words more than yours.
So make sure you use that.

Here's how it looks:

Client says: "I want it clean and minimal." 
Then later: "Add more icons, a banner, and a pattern." 
You say: "You wanted clean and minimal. Which of these fits that?"

Client says: "The homepage should feel premium." 
Then sends a reference that looks like a discount flyer. 
You say: "You said premium. This reference feels promotional. Let's find a better match."

Client says: "I love version 2." 
Two days later: "Actually, let's go back to version 1." 
You say: "You approved version 2 last week. What changed?"

No arguing. 

No "I told you so." 
Just their words "Echoing".

What if the client hasn't said anything specific?

Then quote the brief. 
The approved direction. 
The project goals defined.

Why this works: people respect their own words more than yours.

Use playback → fewer revisions, faster approvals.
Start quoting your clients back to them 😇

P.S. Repost this ( Save ⭐ this and Repost ♻️ )

Quick UX test:

Where is “Cancel subscription”?

A. Upgrade plan
B. Personalization
C. Settings

Drop your answer below.

P.S. Bonus points if you explain the UX logic behind your choice.

Dark patterns are the biggest scam in tech.

And nobody's talking about it.

Today I tried to buy an CapCut app subscription.
Price on screen: €59.00/month.
Price at checkout: €299.99.

That's 30x markup.

This isn't a bug.
It's a feature.

It's called "bait and switch"one of the most common dark patterns in UX design.

Here's why it keeps working:

1. The purchase screen looks trustworthy
2. The real price only shows at the very last step
3. Most users tap "confirm" without reading

I teach designers to reduce friction.
But some companies use that skill to reduce your attention.

This wasn't some sketchy app.

It was in the App Store.
Processed through Apple.

If you're building products:

- Transparent pricing isn't a "nice to have"
- Transparent pricing is the bare minimum

Dark patterns are a tax on user trust.

And trust, once broken, doesn't come back.

♻️ Share this with someone who builds digital products.

CapCut

Building a portfolio shouldn't take 6 months.

Here's how to create yours ultra-fast.

1. In Google Docs, write 1 case study
2. Use UX Pilot AI to explore design directions in minutes
3. Export to Figma or Framer, drop in your real projects, and make it yours

Always remember:

Write your case studies before you design.
Focus each one on one problem and one outcome.
Skip the long backstory ( get straight to the project ).
End every case study with results or lessons learned.

Never do this:

Obsess over fonts and colors before you have content.
Cram more than 3–5 projects on your homepage.
Show projects without explaining your process.
Drop in random mockups with no context.
Design more than you write.

Bonus:

If a hiring manager can't find your thinking in 30 seconds… It means your portfolio failed.

Try UX Pilot for free: https://lnkd.in/d3jFWw7x

Hope someone lands an interview this week. Good luck! ❤️

P.S. Repost ♻️ for designers in your network.

P.P.S. What would you add to the list?

No one will taught you this at any design workshop.

Your clients used to find you like this:

1. Search Contra ( or other freelance platforms )
2. Check your portfolio
3. Like your vibe
4. Read your bio
5. Send a message

Every step there was a real person deciding with their gut.

Do they like you.
Will they hire you.

But, software is now doing that process. 

Browsing profiles. 
Comparing freelancers. 
Choosing who to hire. 

Without a human in the loop.

Now:
AI agents entered the freelance market. 

Which means...

1. A new buyer category that evaluates without messaging you first
2. Profiles being scanned by software, not just humans
3. Less "let me think about it" - more instant decisions
4. More freelancers who have no idea this is happening
5. A new competitor category that delivers work 24/7

The point I want to make is:

Most freelancers don't fall behind because of skill. 
They fall behind because of awareness. 

I really hope YOU won't. 

(Repost further for others ♻️)

P.S. I created a playbook on how to attract these AI agents. 

P.P.S. Comment “AGENT” and I’ll send it your way. 

( Make sure we’re connected first ❤️ )

taplio

The Battle

Mihai Anton

Pavle Lucic

Followers

Posts

Engagements

Likes

Performance ScoreMihai Anton's LinkedIn Performance

Performance ScorePavle Lucic's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Mihai Anton

Pavle Lucic

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Mihai Anton's LinkedIn Performance

Performance ScorePavle Lucic's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreMihai Anton's LinkedIn Performance