Case study
Building an Enterprise AI Upskilling Pipeline for Nebius Academy in LATAM
How DK Global helped Nebius Academy validate a new market, find positioning that worked — and grow SQL conversion from 7% to 73%
7–8 min read
28
SQLs
Qualified opportunities from outreach
7% → 73%
SQL conversion
From early to late cohort
5
Enterprise deals
In active negotiation
9
Months
Market testing and pipeline development
51
Meetings
Enterprise conversations across LATAM
7
LATAM countries
Engaged through active outreach
Client context
Nebius Academy — the driver of AI Adoption for B2B of Nebius Group — helps organizations embed AI into how they actually work, not just send employees through a course catalog. The platform combines Nebius Group’s technical infrastructure with specialized learning programs built on knowledge, experience and success
The Brief
The product was strong. The market didn’t know it existed — and there was no data to show who was actually ready to buy.
When Nebius Academy entered the Spanish-speaking LATAM market, regional activity consisted of events and occasional personal introductions, with an unstructured pipeline, an unproven database, and limited visibility into how enterprise decisions around corporate learning were made or who made them.
DK Global was brought in to test demand, identify positioning that resonated, and build a Social Selling engine capable of generating qualified enterprise conversations at scale.
Starting Point
No established pipeline
No proven database
No visibility into decision-making
No clear signal of who was ready to buy
Three things made this harder than a typical market entry
01/
Saturated EdTech market
The EdTech space in LATAM is saturated. Generic platforms, off-the-shelf corporate training, and e-learning providers have already created a wall of skepticism. Cutting through with a product that operates fundamentally differently — in cold outreach, with no brand recognition — was the first problem
02/
No way to pre-qualify intent
There was no way to pre-qualify intent. No historical data, no industry benchmarks, no signal to distinguish organizations actively investing in AI transformation from those casually curious about it. That distinction was critical — and only became visible through real conversations
03/
Decision-making split across roles
The buying decision didn’t sit in one place. In some organizations, AI upskilling is driven by technical leadership — CTO, Head of AI, Engineering Directors. In others, it lives in HR and L&D — CHRO, CLO, Learning Directors. Both tracks required simultaneous engagement, with different framing for each
Core insight
The assignment wasn’t to generate leads. It was to build a system that could find the right organizations, reach the right people, and figure out what actually resonated — before the pipeline could be built at all
Who We Were Talking To
Finding where AI adoption was already an active priority
The program focused on mid-market and enterprise companies across Spanish-speaking LATAM where AI adoption had already become part of the strategic agenda — not a future consideration, but an active organizational priority.
Audience Snapshot
Mid-market & enterprise
Companies with active AI adoption agenda
250–5,000+ employees
Company size
Spanish-speaking LATAM
Regional focus
Target industries:
Fintech
IT Services
Software Development
Telecom
E-commerce
HealthTech
EdTech
Cybersecurity
SaaS
B2B Tech Platforms
Wholesale & Retail
Selection logic
The selection wasn’t arbitrary — but it wasn’t proven either. These were sectors where AI upskilling appeared strategically relevant based on initial assumptions. Which of them would actually convert was one of the questions the project was designed to answer
Two parallel tracks
Within each target account, outreach ran across two parallel tracks:
Technical leadership:
CTO, Head of AI/ML/Data, Engineering Leaders
Learning & development:
CHRO, CLO, L&D Directors
The rationale was the same as in any enterprise sale: a single contact rarely moves a decision forward
Not a Single Playbook– an Iterative System
The strategy was built around one principle: treat outreach as a learning process, not a delivery mechanism.
Rather than fixing the ICP at the start and executing against it, the team updated targeting criteria continuously — based on response patterns, objections, and what emerged from live meetings. Each conversation was a data point, not just a pipeline event.
The system ran across seven stages:
Stage 1
Audience & Database
Initial ICP and priority countries were defined based on product assumptions and industry relevance. The database was refined as new patterns emerged — targeting became more accurate over time rather than fixed from the start
Stage 2
Messaging & Positioning
Three approaches were tested: networking-first, direct value proposition, and signal-based personalization. Every shift was tracked against market response. What changed the trajectory most is covered in section 5
Stage 3
Outreach funnel & profiles
Outreach ran through the LinkedIn profiles of the regional commercial manager and two sales managers, each repositioned to reflect Nebius Academy’s offering. Sequences were adapted with each positioning iteration
Stage 4
Signal-based personalization
Observable company-level triggers — AI hiring activity, news of digital transformation initiatives, tech stack expansion — were used to prioritize outreach and shape messaging. Generic targeting was replaced with intent signals wherever possible
Stage 5
Content & thought leadership
Content was published on a regular basis through the outreach profiles, positioning Nebius Academy within broader AI adoption conversations — ecosystem observations, implementation challenges, partnership development across LATAM — rather than course promotion
To strengthen engagement and relatability, professional topics were intentionally mixed with personal stories and local experiences. It helped humanise the communications and build stronger connections with the audience
Beyond publishing, the team actively engaged with the target audience through comments. On behalf of the regional team, targeted comments were left on posts of ICP decision-makers, contributing to ongoing discussions around AI adoption and digital transformation
This approach helped increase visibility within the right audience, establish early familiarity before direct outreach, and position Nebius Academy as an active participant in the ecosystem rather than an external vendor. Content was also distributed through relevant LATAM IT communities, building credibility and attracting inbound interest from target decision-makers
Stage 6
Dialogue & qualification
Every meeting was analyzed for patterns: company context, decision-making structure, internal constraints. Qualification criteria were updated based on what real conversations revealed, not initial assumptions
Stage 7
Continuous optimization
All stages were monitored in real time. The objective at each iteration was the same: improve the accuracy of identifying high-probability prospects before the call, not after
Positioning as the Turning Point
One finding shaped the project more than any other.
When Nebius Academy was introduced as corporate training or a course platform, conversations stalled quickly. The most common response: “we already have everything we need.” In a saturated EdTech market, that framing closed doors before any real dialogue could begin.
When the positioning shifted — from training provider to AI adoption platform, one that helps organizations embed AI into how they actually operate — the nature of responses changed. The offering became harder to dismiss and easier to discuss at a strategic level.
This shift had a direct impact on SQL conversion — and it informed every subsequent iteration of messaging, targeting, and dialogue development across the project.
The shift
Training provider → AI adoption platform
The result
SQL conversion: 7% → 73%
Getting "we already have everything we need" too often?
Turn that into a meeting
Turn that into a meeting
What Worked. What Didn’t
Worked
Signal-based personalization
Signal-based personalization opened doors — including senior conversations at companies like Banco Santander and Raízen
Worked
Framing through business outcomes
“Your team will be able to implement AI solutions independently” landed differently than “complete a machine learning course.” The more specific the outcome, the more seriously the conversation was taken
Worked
Flexible entry points
In some contexts a direct meeting request was effective. In others, a softer qualification dialogue produced stronger engagement. The approach was adjusted to context rather than applied uniformly
Didn’t Work
Networking-first outreach
attracted contacts who were curious about AI but had no active initiative, no allocated budget, and no internal mandate to act
Didn’t Work
Vendor-style messaging
did not cut through in an oversaturated education market. Standard sales-led copy was filtered out before the conversation could begin
Didn’t Work
Mid-level targeting proved ineffective
Managers at that seniority rarely held the internal influence required to move a corporate learning initiative forward
After Nine Months
The conversion growth reflected increasing qualification accuracy — not volume or luck.
As real conversations accumulated, the ability to identify high-probability prospects before the meeting improved substantially.
Pipeline results:
~6,000
Prospects reached
600
Leads
51
Meetings
28
SQLs
5
Deals
SQL conversion rate
7% → 25% → 73%
Each stage of the project produced better signal, which produced better targeting, which produced better conversations
Market validation
7
LATAM countries engaged through active outreach
6
industries explored for AI upskilling demand
3
positioning hypotheses tested against live market response
1
core positioning validated through enterprise conversations
Over 50% of meetings progressed to SQL status — well above typical B2B qualification benchmarks — indicating that outreach was consistently reaching stakeholders with both the intent and the organizational influence to act
Impact Beyond the Numbers
Nine months of outreach conversations produced something harder to quantify than meetings or SQLs: a clear picture of who actually buys AI upskilling in LATAM, how those decisions are structured, and which industries are moving fastest on AI transformation.
Critically, the project revealed how to distinguish organizations actively exploring AI transformation from those casually curious about it — a distinction that isn’t visible from the outside and only emerges through real conversations at scale.
Nebius Academy moved from fragmented, event-dependent activity to a systematic market presence: a defined ICP, a functioning pipeline, and a validated foundation for regional scale.
In Their Words
Before this project, our activity in LATAM was largely event-driven and fragmented.
What changed is not just the number of conversations, but their quality. We now have a clearer understanding of who our buyers are, how they make decisions, and how to position our offering in the context of AI adoption.
The improvement in SQL conversion reflects this shift.
Aleksei Timofeev
From Our Side
The core difficulty was that AI upskilling is a space where companies often don’t feel an immediate pain point. We had to make the gap between their current capabilities and where they could be tangible and specific — not abstract.
We weren’t selling courses. We were helping organizations see what stands between them and real AI implementation — and what to do about it.
Through continuous testing and live market conversations, we found positioning that resonated. That’s what drove the SQL growth.
Daniel Kocherga
Co-Founder, DK Global
·
LinkedIn
LinkedIn
Return to main page
Return to main page
Talk to founders
© 2024-2026 all rights reserved
