Back to customersAt a glance
- Every sales conversation lands as structured account intelligence
- Forecast prep is no longer a manual reconstruction project
- Pipeline leadership trusts without spot-checking it
- Deals with six or more stakeholders tracked in structure, not free-text notes
About Tangos AI
Tangos AI builds an autonomous investigation engine for risk, compliance, and financial crime, taking escalated alerts through hypothesis, evidence, reasoning, and resolution to produce a single examiner-ready case file with every fact traced to its source. Founded in 2025 by Eyal Azoulay, the company is headquartered in Israel with a Washington, D.C. team of former U.S. Treasury OFAC officials, and raised a $20M seed round in July 2026. It sells into banking, fintech, payments, marketplaces, and government agencies — buyers for whom, as the company puts it, almost right is not good enough.
The challenge
Tangos never had a real CRM. Pipeline lived in Notion databases the team designed themselves — deals, contacts, next steps, forecast, all maintained by hand. It worked while the company was small enough that everyone had been in every room.
“It’s something we built and maintained, and the data was only as current as the last person who remembered to open it.”
Nir Stanger, VP Business & Operations, Tangos AI
Two things broke it at the same time. The first was deal shape. Tangos sells into regulated financial institutions, which means long cycles, six or more stakeholders, and compliance, procurement, and security review all running in parallel. That doesn’t survive in free-text notes. The second was headcount. Once deals were running, “ask me where that stands” stopped being a system.
What the team needed was every conversation coming out the other side as structured account intelligence, and a pipeline legible to someone who hadn’t been in the room — without adding an hour of admin to every meeting.
The turning point
The shortlist was the standard one: HubSpot and Salesforce. Nir’s read on both was pragmatic rather than dismissive. “Both are credible products, and either one solves ‘we don’t have a CRM.’” What neither solves is the thing underneath it.
“A CRM is only ever as good as the data someone remembered to enter.”
Nir Stanger, VP Business & Operations, Tangos AI
That is the distinction between a system of record and a system of intelligence. A horizontal CRM gives you a place to put the data; it still depends on someone opening it. Dreamhub captures calls, emails, and meetings automatically and structures them against a B2B software revenue ontology — so multi-stakeholder, compliance-gated deals land as qualification data and stakeholder maps rather than paragraphs someone typed at 9pm.
The rollout
“Easiest part of the project, and I’ll be honest that our bar was low.” With no legacy CRM in place, there was no schema to map and no stale records to clean up. Tangos stood Dreamhub up and started working out of it.
How Tangos uses Dreamhub
The core of it is that account intelligence now assembles itself. After a discovery call, Dreamhub captured the prospect’s technology stack and a pain point that hadn’t made it into Nir’s own notes. As he put it: “That’s the difference between capturing what was actually said and capturing what somebody had time to write down.”
For deals that run six or more stakeholders through compliance, procurement, and security in parallel, that difference compounds. Stakeholder positions, next steps, and qualification state stay current across a cycle long enough that nobody’s memory is a reliable backup.
Reporting is the other half. Forecast cycles and board conversations used to mean reconstructing pipeline state by hand, going meeting by meeting and filling gaps from memory. Now the pipeline view is standing, and it’s legible to people who weren’t in the room.
The results
Reporting stopped being a project. The pre-forecast reconstruction pass is gone, and the pipeline stays current without anyone maintaining it. But the change Nir points to is trust, not time.
“Now it’s just there, and more importantly I trust it without spot-checking it. A pipeline view you don’t trust is worse than not having one, because you end up making decisions off it anyway.”
Nir Stanger, VP Business & Operations, Tangos AI
That closes the arc. The original problem was never that Notion was the wrong tool — it’s that any hand-maintained pipeline is only as good as the last person who remembered to update it, which means every number gets mentally discounted before it’s used. Automatic capture removes the discount.
Asked what he’d tell another VP of Sales, RevOps leader, or founder weighing the same decision, Nir’s answer moves past the admin math: “If your team spends more time logging meetings than learning from them, that’s the signal. And it’s not really about the admin hours, it’s about what gets lost.”
Like Tangos, B2B software teams are running their revenue motion on a CRM that captures the intelligence instead of asking for it. Book a demo to see what your sales motion looks like on a CRM built for it.