The challenge
A company that had never needed territory infrastructure, until now
Mistral is an enterprise-focused frontier AI lab that has raised billions in funding, with customers around the world across a variety of industries, including finance, manufacturing, public sector and defense.
"Our Salesforce instance, up until maybe two or three months ago, was relatively nascent. We were trying to intentionally keep it as lightweight as possible. We just turned it into a database, a very expensive one."
Joel Davidson
Revenue Operations
But the sales team was about to rapidly scale across four regional theaters. That meant territories, rules of engagement, more granular targets and reporting, which required accurate hierarchy data and account master data across geographies, verticals, and enterprise size.
Full hierarchies mapped with operating parent logic, not legal entity structure, that go deep into multi-entity accounts
The enterprise account base ($1B+ in revenue) was largely flat, with minimal built parent-child relationships. There was also no concept of domestic parents, which matters because Mistral sells to enterprise accounts in every region and needs to know which parent accounts or subsidiaries a rep should own.
The RevOps team at Mistral is at the cutting edge, and they wanted to deploy agents to take actions across their CRM. Unfortunately, data accuracy in the key fields used by their agents was not sufficient for them to be confident in success.
"There were too many duplicates."
Joel Davidson
Revenue Operations
Thousands of duplicates identified and resolved across the CRM
Every inbound lead auto-created a new account to house it. On top of the flat hierarchy and duplicate mess, the existing firmographic enrichment gave the team numbers with no supporting evidence.
"For our existing enrichment, we just get a number and there's no background context."
Joel Davidson
Revenue Operations
That left RevOps defending data quality with no backup. Reps debated segment boundaries because nobody could point to a source.
As a leading AI company, Mistral could have tried to build its own infrastructure. However, their RevOps team decided that outsourcing data infrastructure to a trusted vendor was the more scalable option, with mass efficiency, network, and support advantages.
The decision
Kernel ran a proof of concept against Mistral's account data
"The POC was incredibly impressive with a rapid turnaround and presented in a way that we could easily audit, interrogate and understand."
Joel Davidson
Revenue Operations
Mistral was evaluating multiple vendors in parallel. After the Kernel POC, Mistral chose Kernel for the full suite of account enrichment and entity resolution, hierarchies, deduplication, AI firmographics, and custom verticals. Mistral had their first real territory build upcoming, and hierarchies were the most important part to get right fast.
Legacy hierarchy data providers returned legal entity structures that didn't match how Mistral sells.
HG Insights and Traction Complete were unable to solve hierarchies
"The reason we wanted to move so quickly is: do it once, do it right."
Joel Davidson
Revenue Operations
How Kernel works inside Mistral today
Kernel runs identity resolution across Mistral's full CRM. Each account is matched to the Kernel entity database via KERN ID, and Salesforce is populated with hierarchy relationships, firmographics, custom verticals, and cleaning action recommendations.
The implementation started with deduplication and hierarchy association across 35,000 accounts. To maintain control, Mistral created the required CRM fields and managed CRM-side configuration. The Mistral RevOps team uses the Kernel customer application to review and execute merges and associations, with safeguard rules protecting accounts with open opportunities or active owners.
Kernel's inbound enrichment API enriches new accounts as they're created in Salesforce. The custom vertical taxonomy (22 industries with 90+ sub-verticals) is maintained in the Kernel app and applied automatically.
Custom 22-industry vertical taxonomy preserved and extended with sub-verticals + nuanced segmentation
The account data reasoning fields have changed the dynamic between RevOps and reps: instead of RevOps having to defend every data point, reps can now see the sourced rationale for each figure directly in Salesforce.
The agents that the Mistral team has built now run on Kernel data. With Kernel's data consistency as the source of truth, Mistral can ensure their agents act predictably, yielding extremely promising results so far.
