Adrian Guerra

Adaminds

On a mission to help agency owners grow 💜⚡️

🚀 Book a discovery call → 📆 https://calendly.com/adaminds/discovery-call

Founder

Swisscom

Business Analyst

Capgemini

Senior Data Analyst

Cebedia

Koch Supply & Trading

Data Scientist

AI & Business Intelligence Consultant

Marketing Scientist helping agency founders grow 💜⚡️

DM me GROW for a FREE Linkedin outreach strategy you can plug into your business today 🎁

Adamax is a full-service software development and technology consulting company, specializing in data science and analytics solutions. Our services include product development, digital transformation and IT operations for businesses.

Voila AI

Voila AI is a Swiss-based one-stop-shop helping organizations navigate the AI revolution.

Concretely, we help you:
- Identify and prioritize AI opportunities tailored to your organization’s specific needs, to start realizing ROI in weeks, not months.
- Build and deploy AI solutions and support adoption to optimize impact and maximize ROI.
- Train your teams on cutting-edge AI tools (e.g., ChatGPT, Copilot) for immediate productivity gains of up to 40%.

Interested in bringing AI to your organization?
Reach out directly on LinkedIn or via email at contact@voilalabs.ai

Co-founder & Managing Director

Boston Consulting Group (BCG)

I was focusing on digital, tech and AI topics across industries.

Associate

University of Cambridge

Language Technology Lab
Topic: Multilingual zero-shot learning for vision-and-language models.

Visiting Student - Master Thesis

Databricks

Storage & IO | Databricks Runtime
Project: Early detection of corruptions in Parquet files

Software Engineer Intern

Ecole polytechnique fédérale de Lausanne

• Continued my work at LSIR Lab started as a Summer Intern. I was engaged part time (10-12h a week) in parallel of my studies.
• Prototyped an extension of the platform (using Streamlit) to index faces in videos and track them.

Master Research Scholar

• Developed a quality pipeline to improve faces extraction for a European Red Cross project ("Trace the Face" initiative). Implemented a solution decreasing the number of indexed faces by 30%.
• Worked on both the backend (using Django and PostgreSQL database) and frontend (Vue.js, HTML, CSS) of the application. I improved the interface used for images exploration.

Research Intern

🔹 𝗔𝗿𝗲 𝘆𝗼𝘂 𝘄𝗼𝗻𝗱𝗲𝗿𝗶𝗻𝗴 𝗵𝗼𝘄 𝗔𝗜 𝗰𝗮𝗻 𝗵𝗲𝗹𝗽 𝘆𝗼𝘂𝗿 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗯𝗲𝗰𝗼𝗺𝗲 𝗺𝗼𝗿𝗲 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁, 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗲, 𝗮𝗻𝗱 𝗽𝗿𝗼𝗳𝗶𝘁𝗮𝗯𝗹𝗲?
🔹 𝗗𝗼 𝘆𝗼𝘂 𝘄𝗮𝗻𝘁 𝘆𝗼𝘂𝗿 𝘁𝗲𝗮𝗺𝘀 𝘁𝗼 𝗯𝗼𝗼𝘀𝘁 𝘁𝗵𝗲𝗶𝗿 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝘄𝗶𝘁𝗵 𝗔𝗜 𝘁𝗼𝗼𝗹𝘀, 𝗯𝘂𝘁 𝗮𝗿𝗲𝗻’𝘁 𝘀𝘂𝗿𝗲 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗮𝗿𝘁?

Many organizations struggle to keep pace with the rapid advancements in AI, 𝗺𝗶𝘀𝘀𝗶𝗻𝗴 𝗼𝘂𝘁 𝗼𝗻 𝗰𝗼𝗻𝗰𝗿𝗲𝘁𝗲 𝗼𝗽𝗽𝗼𝗿𝘁𝘂𝗻𝗶𝘁𝗶𝗲𝘀 𝘁𝗼 𝗱𝗿𝗶𝘃𝗲 𝗿𝗲𝗮𝗹 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗿𝗲𝘀𝘂𝗹𝘁𝘀.

With extensive experience in AI and digital transformation from EPFL and as a Digital Consultant at BCG, I co-founded Voila AI, a consultancy dedicated to 𝗵𝗲𝗹𝗽𝗶𝗻𝗴 𝗼𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻𝘀 𝗻𝗮𝘃𝗶𝗴𝗮𝘁𝗲 𝘁𝗵𝗲 𝗔𝗜 𝗿𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻 – translating complexity into real impact and 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝗴𝗮𝗶𝗻𝘀 𝗼𝗳 𝘂𝗽 𝘁𝗼 𝟰𝟬% 🚀.

🔹 How I can help:

✅ 𝗔𝗜 𝗔𝘂𝗱𝗶𝘁 & 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆:
	• Quickly identify high-impact AI opportunities tailored to your business
	• Practical roadmap to see ROI within weeks

✅ 𝗔𝗜 𝗔𝗰𝗮𝗱𝗲𝗺𝘆:
	• Equip your teams with skills to boost productivity thanks to efficient usage of AI tools (e.g., ChatGPT, Copilot) 
	• Broaden their perspective on AI so they can confidently lead initiatives within their departments

✅ 𝗔𝗜 𝗦𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝘀 𝗕𝘂𝗶𝗹𝗱 & 𝗗𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁:
	• Custom-built AI solutions matching your exact business needs and constraints
	• End-to-end support ensuring successful deployment and adoption

🔹 𝗪𝗵𝘆 𝗰𝗹𝗶𝗲𝗻𝘁𝘀 𝗰𝗵𝗼𝗼𝘀𝗲 𝘁𝗼 𝘄𝗼𝗿𝗸 𝘄𝗶𝘁𝗵 𝘂𝘀:
	• No bullshit: clear, tangible recommendations that deliver real impact
	• Proven experience delivering AI projects in complex, global corporate environments
	• Strong emphasis on rapid implementation and measurable results

📅 𝗥𝗲𝗮𝗱𝘆 𝘁𝗼 𝘂𝗻𝗹𝗼𝗰𝗸 𝘆𝗼𝘂𝗿 𝗼𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻’𝘀 𝗔𝗜 𝗽𝗼𝘁𝗲𝗻𝘁𝗶𝗮𝗹?
Book a discovery call: contact@voilalabs.ai
Or send me a direct message here on LinkedIn – let’s get started!

Maxence Jouve

n8n is everywhere on LinkedIn right now. Everyone is building "no-code" agents.
But (most of) your employees still cannot use them.

I was discussing with an IT leader recently.
He told me he is fan of n8n.

I then asked him who can use it internally?
“For our typical non-tech users, it is true that it’s gonna be hard to onboard them”

What you should avoid when doing AI "internally”:
- Your IT team (maybe 1-3 people) builds an agent in n8n
- It works very well on paper
- Then 10 employees need to use it.
- A small tweak like changing a prompt, or adding a data source —> a new ticket to IT is open

Don’t get me wrong, I also believe it is a very powerful tool if you know how to use it.

However, the question you should ask yourself:
Can the business users iterate on the use case themselves?

Tool choice matters here.
If you mostly have non-tech users, or if your team is already very tech savvy, the tool choice should be different.

“AI agents” or workflows need to improve through use and that requires fast feedback loops. When only developers can modify agents, those loops die.

This separates AI that get adopted from AI that goes to the trash.

On which side of the picture are you?

AI for operations | SMEs & investment firms | ex-BCG | EPFL

"Do we need a data lake before we can do AI?"
This belief is costing SMEs 12-18 months of waiting.

This question came up last week in a call with a SME:
"Is there a basic entry point for IT infrastructure? Some companies have data all over the place, nothing connected. Others have structured data lakes."

The question shows a misconception (I hear this in half my discovery calls):
SMEs believe AI requires enterprise-grade data architecture first.

This used to be true for “traditional” AI use cases…
…but this is no longer the case for many “GenAI” use cases.

It creates sequential thinking. 
"First we clean our data." 
"Then we can do AI."

For SMEs without engineering teams, this means waiting indefinitely.

Modern AI agents can connect to existing systems already in place: SharePoint, CRM, ERP, ticketing system, etc.

The AI layer then handle data retrieval (data messiness included).

We saw a client with 100,000+ documents in SharePoint (unstructured), notes from sales in CRM, a ticketing system with 5k+ tickets this year…

…and AI agents are working fine today, even in this mess.

They act on top of your existing infrastructure and retrieve data “semantically”, without needing the rigid structure humans require.

The blocker is not always technical. 
It is the belief that you need VERY clean foundations to begin with.

What's your version of "we're not ready yet"?

Are we in an AI bubble?
J.P. Morgan published the math: $650B in annual revenue. Into perpetuity. For a 10% return.

J.P. Morgan published the math AI vendors / and startups don't want you to know.

To deliver just a 10% return on AI infrastructure investments through 2030, the industry needs to generate $650 billion in annual revenue "into perpetuity."

That's 0.6% of current world GDP.
Nearly Switzerland's GDP ($936B in 2024).

J.P. Morgan puts this into perspective, it is equivalent to:
- $34.72/month from every iPhone user
- $180 from every Netflix subscriber monthly

However, while the comparison is interesting, the real “payers” won’t be consumers.
They will be companies and public entities.

Is it really that crazy to expect AI to generate $650B annually by 2030?
I personally think the numbers are more achievable than they seem.

"Our biggest fear would be a repeat of the telecom and fiber buildout experience, where the revenue curve failed to materialize at a pace that justified continued investment."
Is AI comparable to telecom and fiber?

Do you think we are in a AI bubble?

Your best employee just ignored 80 customer emails.
On purpose.

A few weeks ago, a CEO of a Swiss SME told me his best employees could NOT keep up with the volume of customer tickets.

The issue: legacy ERP system has become so complex that his staff don't know which screen processes a return. Or how to log a sale correctly.

So the employees simply avoid it.
They also have to do in-store service, admin work, and data entry.

Then, the phone rings… and nobody picks up.
Because picking up means another query they can’t answer, or another system they can mess up.

It is hard to hire people who are both great with customers AND great dealing with your complex internal IT systems.

Most companies try to solve this by hiring more people or better training.
The system complexity is the blocker, not the human capacity.

This is exactly where an “AI agent” make sense (in this case, for customer support).

The AI agent routes queries, drafts responses (with human-in-the-loop for validation), updates multiple systems simultaneously, and formats data correctly…

Your team focuses on what moves the needle: judgment and customer relationships.

In most cases, AI DOES NOT replace humans, but it does augments them.

Every week, new AI platforms or tools launch. Companies are waiting for the dust to settle. 
They will end up 18 months behind. This is what you can do:

Last week, a CTO asked: "How do we avoid investing in a platform that disappears in two years?" This is 2008 cloud adoption all over again.

The companies that waited for AWS vs. Azure to "shake out" lost 24 months of learning. 
The ones that picked imperfectly and started building won.

Not because they chose better. Because they learned faster.
AI platforms will change regardless of your timeline.

The true “durable” asset is not the platform itself.
It's your team's adopting AI, experimenting with it and starting to extract value from AI.

To choose a platform / tool, we typically evaluate three things:

1. Vendor reputation.
Who is building this? We check their reputation in developer communities, Reddit, GitHub, Twitter. 

If it is an unknown team? Probably a red flag.

2. Constraints. 
Regulatory, hosting, requirements. This eliminates 70% of platforms immediately.

3. Prior experience. 
We stick with platforms we tested. 
New ones? We test for 30 days internally to build our own view.

At the end of the day, your learning speed will beat the perfect initial choice.

While you are evaluating vendors, some competitors are 12 months ahead. 
They know what works, what breaks, how teams adopt.

This will compound quickly.

Are you building that muscle already, or waiting for the landscape to settle?

McKinsey found that companies successfully scaling AI are 3x more likely to have this one “factor”...
...and SMEs have it by default.

This factor is “Leadership ownership”. 

McKinsey just published their 2025 AI report (full report below).

In enterprises, this means a VP three layers down pitching a steering committee for six months.

In an SME with 50 people, the CEO can walk to the support engineer's desk, sees unresolved tickets adding up, and decides to deploy an AI agent. The same day.

For the same outcome, timelines may be “different”.

Large enterprises: 18 months and six committees. 
SMEs: 6 weeks.

This is where SMEs have an advantage over larger orgs.

This new generation of AI technology (”GenAI”), allows for rapid deployment and iteration.

You no longer need huge resources, but curiosity to experiment, and direct ownership.

Are you using proximity as an advantage, or are you copying “large” enterprise playbooks for your SME?

The state of AI in 2025 - McKinsey

Every AI vendor wants to sell you a platform…
…without knowing what’s really broken in your operations.

This happens on nearly every discovery call.
Companies jump to "what tool?" before they understand "what are their challenges"

The fix: an audit phase.
“Consulting” work that happens before you build anything or write code.

We follow three steps:

1. Get executive buy-in 
Find a department head who can champion the work and articulate what success looks like.

→ Typically AI adoption dies when it stays fully within the IT department.

2. Survey all employees 
Key questions we ask:
- "List your weekly tasks and time spent"
- "If you had a magic wand: what would you fix?"
- "Which tools (e.g., website, database, software) do you currently use?"

→ This finds what's broken and where time disappears.

3. Interview 6-10 operators (30 min each)
We record everything. Ask them to walk through exact steps (while sharing screen), where they lose time, etc.

→ Surveys give you the pain. Interviews show you the WHY.

By the time you build, you know exactly what to do and how to measure it.
No “strategic” paralysis. No tool FOMO.

This is our Express AI Audit done in 3-4 days.

At the end, you get an opportunity list (with ROI per use case), and a prioritised roadmap.

15 years ago, three guys in London made a big bet. Two weeks ago, it paid off.
Here is the story behind Google’s AI lab.

Gemini 3, Google’s latest model, was released last week and beats most other LLMs (across most benchmarks).

Google is back on top. But it did not happen overnight.
It is the result of a strategy that looked (very) patient.

Here is DeepMind’s timeline:

- 2010: Demis Hassabis (chess prodigy who coded Theme Park at 17), Shane Legg, and Mustafa Suleyman start DeepMind with a mission to "solve intelligence."

- 2014: Google acquires DeepMind.
Not for immediate product revenue (it didn't come for almost a decade). For talent and methodology.
They gave the lab massive compute but let it maintain independence.

- 2016: AlphaGo defeats Go champion Lee Sedol.
The match was broadcast live across Asia. Millions watched an AI do what the research community thought was a decade away.

2020: AlphaFold solves the 50-year protein-folding problem (it then won the Nobel Prize in 2024).
Hassabis tweets in '21: "This is a day I've dreamed of my whole life, this is the reason DeepMind was founded, to build AI and use it to accomplish breakthroughs in science."

- 2023: Google merges its Brain team with DeepMind. They launch Gemini. But OpenAI is still far ahead in terms of “mindshare”.

- November 2025: Gemini 3 drops.

DeepMind did not go for quick wins. Actually, they started before AI as we know it today was a thing.

Rather, they built fundamental capabilities through constrained domains first: games, proteins, world models etc. before going for “general intelligence”.

What AI platform is your company using *officially* (I insist here)?
(If half your colleagues are using ChatGPT without approval, it doesn't count)

I am curious to see if this matches what I am seeing with SMEs in terms of adoption.

What AI platform is your company officially using?

The author can see how you vote. Learn more

taplio

The Battle

Adrian Guerra

Maxence Jouve

Followers

Posts

Engagements

Likes

Performance ScoreMaxence Jouve's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Adrian Guerra

Maxence Jouve

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);}Maxence Jouve's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreMaxence Jouve's LinkedIn Performance