Navan uses Kernel to elevate its CRM with travel-spend intelligence for more than 300k accounts

+300 Reps

Series G

Travel & Expense Management

300k

Accounts enriched with travel spend intelligence

300k

Accounts prioritized based on ICP criteria

HQ: Palo Alto, California

Founded: 2015

Website: navan.com

Navan runs on Kernel

Navan is known for pushing the frontier of what is possible both with our products and ways of working. To stay on the frontier, we reviewed hundreds of AI solutions for internal use cases. From the first conversation, it was clear that Kernel was built enterprise-first, able to handle the scale and complexity of large-scale deployments.

Chris Price

Chief Information Officer

Overview & Challenge

Navan is the all-in-one travel and expense platform empowering employees to achieve business goals while managing expenses effectively.

On the travel side, Navan targets companies whose teams travel often - firms that benefit most from faster, simpler booking. However, travel budgets are usually private. Headcount can be used as a proxy, but it’s a blunt instrument at best.

For a start, Navan’s headcount data came from LinkedIn Sales Insights (sunset in early 2025), but headcount data from LinkedIn is inaccurate and generic: undercounting in blue-collar industries and DACH markets, overcounting for gig-work platforms.

The real issue with headcount is that higher headcount doesn’t always equal higher spend. A ten-person boutique consultancy crossing the Atlantic every month can outspend a thousand-strong manufacturer whose crews rarely leave town.

Going by headcount alone leads to false positive errors and false negative errors, and the only real fallback is for sales reps to look manually for relevant travel spend clues or simply to reach out based on instincts and hope they’re right. Reps are forced to spend time researching every account and then disqualifying an account out later in the research process. This doesn’t scale well across a sizeable CRM.

Navan partnered with Kernel to deliver precise travel spend intelligence to over 300,000 accounts. Kernel also replaced LinkedIn Sales Insights, offering nuanced headcount data based on real evidence from primary sources rather than relying exclusively on LinkedIn. Sales reps now see detailed insights of all their accounts inside their CRM. RevOps is equipped with the data needed to run the CRM efficiently by prioritizing ICP-fit accounts.

Before Kernel

After Kernel

Scope & Use Case:

Estimating travel spend:

Kernel built an agentic model to assess company travel spends by combining external and internal signals, interviewing Navan reps, and validating model outputs against past data.

On the external side, Kernel deployed over 10 custom agents that identify insights that correspond to higher travel spend. One agent finds explicit travel requirements mentioned in job posts. Another checks if the company has hired a dedicated travel manager. A third scans all employees at the company, determining which are more likely to travel based on their title, seniority, and work description. Others look for signs like regular company retreats or a wide spread of offices across the U.S. and abroad.

On the internal side, Kernel searched the CRM for any notes, lost-opportunity reasons, or emails where prospects talked about a company’s travel budget - valuable insights hidden in plain sight. This means RevOps has all the insights of every email exchange where confirmed travel spend was revealed.

Discovery sessions with seasoned Navan reps uncovered unique clues about drivers of travel spend, which informed the initial base model. Training the custom model on previously closed won account data anchored estimates to reality, allowing Kernel to iteratively refine the model.

Kernel refreshes this travel spend estimate regularly, providing Navan with the latest context, making sales calls highly relevant for potential customers. Reps have a direct line to Kernel via a CRM feedback field, which is regularly ingested and used for model updates and refinements.

Reliable headcount firmographics

LinkedIn retired Sales Insights at the end of 2024. Kernel took its place as a drop-in replacement. This included feature parity with the original requirements, including headcount segmented by geographies and departments, allowing RevOps to continue their work using existing business rules.

However, LinkedIn's headcount data is often misleading. On LinkedIn, headcount is reported in two ways: as the number of associated LinkedIn profiles of the account, and as a self-reported headcount range that often goes stale.

The number of associated profiles assumes that all the company’s employees are on LinkedIn, which is wrong for whole sectors - manufacturing, for one - and for markets like Germany, Austria, and Switzerland (DACH), where local players dominate over LinkedIn. It also folds in alumni, investors, freelancers, ambassadors, and other outsiders - look to Upwork or Airbnb as examples of inflated headcount numbers. On top of that, LinkedIn’s fuzzy matching often trips up, grouping “fabric traders” in India under companies called “Fabric”.

Kernel resolves LinkedIn’s shortcomings by providing its own headcount figure. We start with primary sources: We check the firm’s website, annual reports, Wikipedia, but also LinkedIn. Because many companies keep headcount private, LinkedIn often offers the only clue. When that happens, we cross-check the number of LinkedIn profiles against the firm’s self-reported range, then adjust for its growth rate, industry, and country. The result is a headcount figure you can trust, backed by transparent sources and clear reasoning.

Navan’s RevOps team can now assign territories across more than 300,000 CRM accounts using a reliable headcount figure verified from primary data. Pipeline generation becomes more predictable and scalable with accurately mapped territories.

Case Study →

’’High-conversion target account lists are integral to our go-to-market strategy. Kernel provides us with insights we previously could only get from reps manually researching thousands of accounts.’’

Vadim Zakiyan

VP of RevOps

Before Kernel

After Kernel

CRM Hygiene

Over the years, Navan’s travel and expense platform has grown rapidly, already serving more than 10,000 customers. Like any company experiencing this kind of hypergrowth, the volume of accounts in the CRM has dramatically increased. Fields like website, name, and LinkedIn URL clashed internally, making it difficult to accurately map accounts. This left Kernel with a definitional question when reviewing each record: Which company is this, really? Without a universal understanding of this definition, it makes enrichment impossible.

Traditional enrichment vendors and credit bureaus anchor on one field as the “source of truth”, typically the name or the website. In Navan’s CRM, those fields often clash with each other, and neither always matches what a sales rep would recognize.

Kernel began by fixing the basics. If an account lacked a website, we inferred it from the company name, postal address, or the email domains of its contacts. If it had a website, we checked to see if it loaded or redirected elsewhere.

To fix the ‘account identity’, we approached the account with the question: “How would a rep see this account?”. Consider Pepsi after it bought Frito-Lay. The account name might show “Pepsi”, while the website and LinkedIn link point to Frito-Lay. To decide, we looked at other data points for the account: past opportunities (reps tend to name those correctly), the email domains of contacts, and the office address. Contacts with @pepsico.com and an address in Purchase, New York, meant Pepsi; @fritolay.com and Plano, Texas, meant Frito-Lay.

Clean data is the prerequisite for AI-powered insights, including accurate headcount and complex estimates like travel spend. Get the foundations wrong, and you enrich the wrong account.

Summary:

Navan’s category-defining travel-and-expense platform already serves more than 10,000 customers, but its RevOps team was struggling with a messy CRM to find the next 10,000.

Kernel worked with Navan to clean every record, enrich accounts with accurate headcount data, and, most importantly, ranked all companies by their travel spend. This puts Navan’s GTM team in the driver’s seat, allowing them to slice their CRM into meaningful territories and frontload the accounts that are most likely to benefit from Navan’s offering. Navan’s RevOps team can now prioritize all accounts in the CRM based on travel spend intelligence rather than being limited by generic enrichment fields.

Case Study →

’With Kernel, we have been able to focus all our go-to-market efforts on the companies who actually match our target buyer definitions.”

Markus Goth

Director of RevOps

Flatpay uses Kernel agents to research small merchants at massive scale

100 Reps

Point of Sale

Series B

Read Story →

Remote relies on Kernel custom data to build pipeline

50 Reps

HR Tech

Series C

Read Story →

Ada got RevOps superpowers with Kernel’s custom data tooling

7 Reps

Customer support

Series C

Read Story →

Hear from our other customers

Case Study →

“Kernel let me create stellar territories for 100+ reps in 1 week, including finding 15,000 brand-new accounts.”

Sara Jordão

Revenue Operations Manager, Remote

Case Study →

“Kernel is critical to our sales strategy by ensuring everyone in the revenue function is aligned and focused on the right accounts of the right size.”

Greg Johnson

Chief Sales Officer, Array

Case Study →

"With Kernel, we can be confident that our target market is well-defined and fully researched, allowing us to prioritise accounts by need, resulting in higher sales velocity."

Alun

Head of Revenue Operations

Navan uses Kernel to elevate its CRM with travel-spend intelligence for more than 300k accounts

Navan runs on Kernel

Overview & Challenge

Scope & Use Case:

Estimating travel spend:

Reliable headcount firmographics

’’High-conversion target account lists are integral to our go-to-market strategy. Kernel provides us with insights we previously could only get from reps manually researching thousands of accounts.’’

CRM Hygiene

Clean data is the prerequisite for AI-powered insights, including accurate headcount and complex estimates like travel spend. Get the foundations wrong, and you enrich the wrong account.

Summary:

’With Kernel, we have been able to focus all our go-to-market efforts on the companies who actually match our target buyer definitions.”

Read more customer stories

Flatpay uses Kernel agents to research small merchants at massive scale

Remote relies on Kernel custom data to build pipeline

Ada got RevOps superpowers with Kernel’s custom data tooling

Hear from our other customers

“Kernel let me create stellar territories for 100+ reps in 1 week, including finding 15,000 brand-new accounts.”

“Kernel is critical to our sales strategy by ensuring everyone in the revenue function is aligned and focused on the right accounts of the right size.”

"With Kernel, we can be confident that our target market is well-defined and fully researched, allowing us to prioritise accounts by need, resulting in higher sales velocity."