CASE STUDY
Sales by day.
Data cleanup by night.
Something had to break.
How ATLANT 3D automated data ops – and freed 60% of RevOps time for strategy.
WHAT OUR CUSTOMER SAYS
Hear from Mikel Mangold, Team Lead BD & Channel Partnerships, and Yacine Cherraoui, Revenue Operations expert at ATLANT 3D:
THE SITUATION
Right structure. Wrong data.
ATLANT 3D makes atomic-scale nanofabrication systems. Complex technology. Long sales cycles. Academia as a core market.
Their HubSpot was chaos. Different people using different tags. Properties that meant different things to different team members. Contacts mixed with investors. One person contacted prospects who turned out to be board members.
HubSpot was the sales CRM. Monday handled the detailed sales notes – and parts of project and customer management. The pipeline, and the chaos, lived in HubSpot.
When Mikel joined as Team Lead BD, he rebuilt the framework. Standardized lifecycle stages. Clear processes.
The structure was right. The data was still wrong.
"I just quickly realized I do not know where each of these contacts stands. I have no clue. From a pure CRM perspective, you didn't know what to do with any contact."
Mikel Mangold, Team Lead BD & Channel Partnerships
WHAT WAS BROKEN
Four consequences. All compounding.
Reports you couldn't trust
Different numbers in different views. No way to know which was right. Leadership meetings stuck on "is this data accurate?" instead of "what should we do?"
Systems that couldn't connect
CRM to email. CRM to outreach tools. Every integration broke because the underlying data was inconsistent. Double-checking everything. Manually. Forever.
Decisions too slow
Dashboards showing last month. Not this week. By the time you saw a problem, it had already cost you.
No way to prioritize
The data that tells you who matters – and how to reach them – wasn't there. No signals to rank people. Nothing to tailor outreach around. Every contact looked the same.
The Root Cause
Team members typed data instead of importing it properly. Typos. Wrong companies. Incomplete fields. Each person created their own properties. Each import added duplicates. The mess grew faster than anyone could clean it.
"If you have humans to remember something, then maybe your data already has an issue. No humans can know thousands of data accurately. A machine can."
Yacine Cherraoui, Revenue Operations
WHY INTERNAL FIXES FAILED
Night shifts couldn't save it.
Mikel ran sales by day. Cleaned data until 2 AM.
Still not enough. The monthly report needed accurate numbers. The CRM had chaos. So he cleaned. Every month. For hours.
They kept telling themselves. "You're not focusing on customers. You're not focusing on real conversations."
He was right. He couldn't do both.
HubSpot's Data Hub? $10K/year. For a startup burning runway, that's a hire they couldn't make.
Better data entry rules? Mikel mandated Surfe for contact imports. Required website domains instead of typed company names. It helped – but historical data stayed wrong. And new mess kept arriving.
Doing it themselves? Proper data cleanup needs adults in the room – people who've done it before. So does building AI agents whose output you can trust.
ATLANT 3D tried building their own agents. The output wasn't good enough to rely on.
"I was sometimes until 2:00 AM in the morning cleaning the data. So I have the right data to show in my monthly report the next day. It started to be too much."
Mikel Mangold
What They Needed
- Automation that runs without babysitting
- Enrichment that makes every record actionable
- Someone who could guide them through the AI learning curve
THE SOLUTION
They came for AI outreach. They got a data foundation.
ATLANT 3D found Sellestial looking for one thing: AI-powered outreach. Personalized messages at scale.
We started building it. Then hit the wall.
Personalization without clean data is sophisticated spam.
Wrong job titles meant wrong angles. Missing company data meant generic pitches. Duplicates meant the same person got conflicting messages. The AI was fast. The data made it useless.
So we pivoted. Data first. Outreach second.
"I didn't even realize I could use AI to clean my data. No one was offering that. When we started the AI outreach, we realized how important it is to have the right data."
Mikel Mangold
Data Foundation
Fixed company associations. Merged duplicates. Standardized fields. Created the clean base everything else builds on.
Contact Enrichment
Filled missing job titles, LinkedIn profiles, seniority signals. From "contact exists" to "contact is scoreable."
Lead Scoring
Built accurate scoring based on enriched data. Who fits the ICP? Who's worth senior attention? Scoring that reflects reality, not guesswork.
Funding Enrichment
Custom-built for academia sales. Which research institutions have active grants? Which PIs have budget authority? Where's the money to buy?
Data Entry Enforcement
Workflows that block bad data at the source. Prevention, not just cleanup.
THE RESULTS
From 2 AM cleanups to 2026 planning.
This isn't from one cleanup sprint. It's from data ops that never stop running.
"We are right now setting goals and objectives for 2026. We can leverage the data – who has funding, who's already in conversation – and redirect efforts toward them. Without the agents, it would have been random. Gut feeling."
Mikel Mangold
IN PRACTICE
Two capabilities. Two examples.
What does automated data ops look like for a deep-tech company? Two illustrations from ATLANT 3D's system.
10× outreach velocity – without sacrificing quality.
ATLANT 3D tested two approaches to outreach. One consultant worked manually – careful research, hand-crafted emails, phone calls. Another used AI-powered sequences built on enriched data.
The Comparison
The math was clear. Volume with good-enough personalization beats perfection at low volume.
But this only works if the underlying data is clean. Send 600 AI-personalized emails with wrong job titles? You get 600 embarrassments.
"Yes, the outreach to 10 people was better. But we contacted 600 people. We got 20 more meetings. It doesn't matter quality or not quality – we got the results."
Mikel Mangold
Knowing who can actually buy.
ATLANT 3D sells to academia – universities, research institutes, government labs. Interest is easy to find. Budget is not.
Funding cycles. Grant timelines. Institutional procurement. A researcher might love your technology. Their lab might have no money for 18 months.
Without knowing who has active grants, every lead looks the same.
What We Built
Funding enrichment agent. For each academic contact, the system searches for active grants, research funding status, institutional budget indicators. Writes directly to HubSpot.
The Result:
Lead scoring now includes "can they buy" – not just "do they fit." The 2026 planning conversation changed completely.
"We can now target based on current funnel progress combined with if they have money or not. And redirect our efforts toward them instead of randomly contacting other people. That is a clear outcome of the work we've done."
Mikel Mangold
WHY IT WORKED
Platform + guidance through new territory.
AI tools are everywhere. Knowing how to use them is rare.
"We hit challenges not because Sellestial was a bad tool. It's a new era – new technology stack, new way of working. We didn't have the background. We had to upgrade ourselves."
Mikel Mangold
ATLANT 3D didn't just need software. They needed a partner who could guide them through the learning curve – prompt engineering, model selection, CRM integration patterns. Territory nobody had mapped for them before.
"There are a lot of tools, but there are not a lot of Klemens. We are in a transition – AI tools are booming everywhere. We need experts that guide companies on how to use them. This is probably your strongest value proposition."
Yacine Cherraoui, Revenue Operations
Data first, then automation
The AI outreach pivot proved it. Clean data enables everything. Skip it and you're automating garbage.
Built for deep-tech, not SaaS
Academia sales don't follow SaaS playbooks. Funding cycles matter. Grant timelines matter. Sellestial built custom agents for how ATLANT 3D actually sells – not generic templates.
Audit before action
Understanding what matters before building anything. Not just fixing what's broken – fixing what actually hurts the business.
"You always try to decide and audit what matters before you do anything. Making sure we are focusing on one core objective. When we have this, we can go 100% in that direction."
Yacine Cherraoui
THE BIGGER PICTURE
Technology companies failing at commercial operations.
"A lot of startups struggle not because of the technology or science – they are brilliant in these fields. They struggle because of their commercial operations.
Most businesses are making 100% of their decisions based on only 20% of data.
Having a functional CRM isn't nice-to-have anymore. It's survival."
Yacine Cherraoui, Revenue Operations, ATLANT 3D
ADVICE
"I don't work for a company that doesn't use this."
"Most people believe it's not a priority or they'll figure it out later. That's a bias created by lack of knowledge.
If you could save one or two employees per year and have three times more customers – who would say no?
I am a technologist. I believe technology empowers us. You have technologies out there that can empower your team, your predictions, your visibility. It makes you so much more efficient.
I will never go back to a company that doesn't use it."
Mikel Mangold, Team Lead BD & Channel Partnerships, ATLANT 3D
Is your startup drowning in data fires?
Start with an audit. See what's broken. What it costs you. How to fix it.
Start with Scoping2-3 week deep dive · $2,500 · Credited toward implementation