Case 03 · Nebius Academy Nebius Group · NASDAQ: NBIS

Zero to product-market fit. With a $2M pipeline to prove it.

At Nebius Academy, the education arm of NASDAQ-listed Nebius Group, we built an AI-based GTM engine with one job: find product-market fit for a B2B product in a market nobody had mapped. The engine synced outbound with sales: every call analyzed, and what buyers said on calls became the targeting filters for outbound. It surfaced the three segments that convert, product-market fit in under six months, and $2M in qualified pipeline over the year. Our co-founder Pavel Papin ran Nebius Academy's commercial side as CCO.

0 → $2Mqualified pipeline in 12 months
<6 monthsfrom cold start to product-market fit
3 segmentsthat convert, found by the loop
~2h/dayback per rep, as a bonus
01 The client

An AI-based engine built to find the market.

Nebius Academy did not need more outreach activity: it needed to know who its buyer is. So the engine was designed as a market-search instrument first and an outreach machine second, and the whole go-to-market ran on what it learned.

Nebius Academy Nebius Group · NASDAQ: NBIS

The education arm of Nebius Group, a $50B AI infrastructure company listed on NASDAQ. Nebius Academy sells AI and machine-learning training to engineers and enterprise teams. Pavel ran its commercial side as CCO.

RolePavel Papin, Chief Commercial Officer, carrying the revenue number
ChallengeNew B2B product, unmapped market, no playbook to inherit
MotionHigh-touch B2B sales, call-heavy pipeline
BuiltOne intelligence loop between the outbound engine and the sales engine
TimeframeProduct-market fit in under 6 months; pipeline counted over 12
StatusEngine still running; Pavel keeps sharpening it on an ongoing external engagement
02 The problem

A new product in an unmapped market. Scaling outreach would have burned it.

Nobody knew which segments would buy. The classic path is hiring ahead of knowledge and spending a year guessing.
Outbound and sales ran as two departments: outbound optimized for meetings, sales learned the market on calls, and nothing flowed back.
The most valuable market data the company owned, objections, pricing reactions, who actually converts, died inside call recordings.
Meanwhile the reps' selling time was eaten by prep, notes, follow-ups and decks.

The expensive mistake is not slow outbound. It is scaling outreach before you know who converts.

03 What was built

One loop between the outbound engine and the sales engine.

The sync is the whole case. Outbound stopped being a parallel department and became the market-search instrument of sales.

S1

The outbound-to-sales sync

Everything the sales engine learned on calls flowed back into the outbound engine as concrete targeting filters: which segments to pursue, which to drop, which buying triggers signal a real buyer, which opening angle earns a meeting. Filters stopped being opinions and became call evidence.

S2

Call intelligence as market listening

Every sales call transcribed and analyzed: objections, competitor mentions, pricing reactions, the language that converts. Not for QA. For market intelligence: each call is a data point about where the market actually is.

S3

Segment validation on call evidence

ICP hypotheses were tested and killed against what happened in real conversations, until three converting segments emerged: known buyers, known pains, repeatable messaging. That is what product-market fit means in practice.

S4

The workflow layer, as a bonus

Around the loop, the friction killers: pre-meeting briefs 45 to 5 minutes, follow-ups 30 to 3, per-account decks 90 to 2, ~2 hours a day back per rep and ~10 hours a week for the sales lead. Useful. But the point was never the minutes: the point was the pipeline.

04 Results

Market knowledge first. Revenue followed.

MetricBeforeAfter
Qualified pipelineStarting from zero$2M built over 12 months
Product-market fitUnmapped marketFound in under 6 months
Converting segmentsHypotheses and opinions3, validated on call evidence
Outbound targetingBest-guess listsFilters driven by what buyers say on calls
Call knowledgeDied in recordingsFeeds targeting, messaging and coaching
Rep selling timeEaten by process~2 hours a day returned

Figures from the engine's run inside Nebius Academy, where our co-founder Pavel Papin was CCO. We walk through the artifacts and workflows behind them on the audit call.

"When we brought AI into the sales team, the call analytics and the prep artifacts especially, I got a few hours back every single day. Now I spend that time where it actually moves the number: closing deals."

Cyril Olkhovik · Principal Account Executive, Nebius Academy
05 Installed, not rented

The engine outlived the job.

The CCO left. The engine stayed.Pavel moved on from the seat; the team runs the engine daily without its builder in the room. That is the definition of installed.
Proven inside a public company.Built and adopted within a NASDAQ-listed group, under real security, brand and process constraints.
Still compounding.Pavel continues to sharpen the system on an ongoing external engagement with Nebius.
Now it is the blueprint.The outbound-to-sales loop from this engine is the same loop Spice GTM installs: call evidence steering targeting, messaging and coaching.
Honest framing. Nebius is not a Spice GTM client. This engine was built inside Nebius Academy while our co-founder Pavel Papin was its CCO, and it is shown here as the origin of the Spice GTM method. Details and artifacts are walked through on the audit call.

Want this loop in your house?

One free audit call. You leave with a written teardown of your GTM and the exact spec of your engine, priced.

Book the free audit
All cases Refty · Creator-data platform · Nebius Academy