Shiplog raises $1M to build AI customer intelligence for B2B SaaS

Paris-based Shiplog has raised about $1 million in a pre-seed round to build an AI-powered customer lifecycle and expansion platform for B2B SaaS companies.

The round is backed by Kima Ventures and Project Europe, alongside Purple, No Label Ventures, 100IN and Station F Fund. Shiplog is building agentic customer intelligence to enable personalisation at scale. I spoke to Khushi Mehta, co-founder and CEO, to find out more. 

Shiplog’s AI agent, Ada, evaluates every customer individually and decides the next best action across the full lifecycle, in real time.

From marketing bottleneck to AI-native customer intelligence

The idea for Ada grew out of Mehta's experience leading GTM in e-commerce companies, where she saw a widening gap between what product teams could build and what marketing teams could deliver.

"Personalisation has become the holy grail of modern marketing," she said.

"Engineering teams were shipping features at a pace we'd never seen before, especially with AI, but marketing couldn't keep up."

Despite increasingly sophisticated products, companies were still relying on broad customer segments and generic campaigns to drive conversion.

"With AI and usage-based pricing, every customer uses a product differently," said Mehta.  

"But we're still putting people into the same buckets, even though they expect the kind of personalised experiences they get elsewhere."

She believes the disconnect has turned marketing into a bottleneck rather than a growth engine. "There's an infrastructure gap between what engineering can build and what marketing can deliver," she said.

"That's what inspired us to build Ada. We think the next decade of software will be defined by hyper-personalisation at the individual customer level."

Co-founders Khusi Mehta and Mehdi Gribaa met during the Entrepreneurs First process and now build out of Station F. 

Mehta founded Shiplog at 23 after graduating from ESCP, having previously worked for 6 years in go-to-market, most recently leading marketing for agentic commerce infrastructure. Gribaa founded the company at 25 after graduating from École des Mines with a master's in computer engineering, having previously shipped AI products across finance, healthcare and retail.

From customer segments to a segment of one

Mehta argues that today's customer data platforms (CDPs), customer success software, personalisation engines and marketing automation tools were built for a world that no longer exists.  

"The SaaS model is changing," said Mehta.

"Many incumbents have architectures that were fundamentally designed before AI. Now they're trying to layer AI on top."

Mehta believes that the next generation will be AI-native from the ground up. Instead of layering AI onto existing software, the next generation of companies will be built around AI agents that perform the majority of work autonomously.

As AI compresses software development and pricing collapses from per-seat subscriptions into usage-based models, customers have stopped behaving like predictable cohorts and started behaving like individuals. Yet most companies still read them through segments drawn up weeks or months earlier, and act through tools that either display a dashboard or fire a rule written for the average. 

Instead of dropping customers into fixed buckets or surfacing a health score for a human to interpret, Ada continuously updates each customer's profile as behaviour changes and personalises everything downstream: marketing campaigns, product interfaces, onboarding flows, recommendations, support, lifecycle communications, and the judgement calls that customer success and account teams make by hand today. 

Every interaction adapts to the person, not the average. Shiplog calls the result a live "segment of one," though the company is quick to say segmentation is only one of the things it replaces. Mehta illustrates the problem with a familiar fintech scenario. Traditionally, customers who abandon a Know Your Customer (KYC) verification are grouped into a single segment and sent the same automated reminder a few days later.

"Let's say someone starts KYC but never completes it," she said.

"Traditionally, you'd treat everyone who dropped out as one segment and send the same reminder after three days."

Ada instead combines signals such as previous customer support interactions, emails, phone calls, location and product usage to build a richer picture of each customer.

"You're no longer just looking at someone who abandoned KYC," said Mehta.

"Our AI agent remembers every interaction they've had with your business and uses that context to personalise the next step."

Rather than sending a generic reminder, Ada can tailor messages based on an individual's circumstances and behaviour.

"You're using the customer's entire history and context instead of treating them as part of a generic marketing segment," she said

. "That's how you increase product adoption through personalisation."

Mehta believes the same approach can extend beyond communications to the product itself.

"That could mean dynamic paywalls, personalised onboarding or other product interactions that adapt to each individual customer."

For companies using a product-led growth model, Ada enriches every new signup with publicly available information and data collected during onboarding to build a richer customer profile.

"We try to understand as much as possible about each person," said Mehta.

"That includes professional profiles and the roles they play within an organisation."

By combining enrichment with product usage data, Ada identifies workspace administrators, billing owners, decision-makers and other users, allowing it to distinguish between buyers and end users.

"We can see who's approving billing, who's completing onboarding and who's actually making purchasing decisions," she said.

"That lets us personalise communications for each person's role, rather than treating everyone the same."

An AI decision layer for the customer stack

Shiplog sits above the existing stack, plugging into Salesforce, HubSpot, Snowflake, Shopify, Stripe and other customer platforms. Ada connects to them, builds a live customer profile, and decides what happens next.

Because every revenue team pulls from the same live context, marketing, customer success and account management stop working from separate snapshots of the same account. Every recommendation and autonomous action is logged with supporting evidence, and organisations decide which actions require a human in the loop. Ada is designed to be fully transparent. 

Users can inspect every piece of information the AI has used, including where it came from and when it was sourced. Whenever the platform presents a recommendation, the supporting sources are attached so users can verify everything themselves.

“Nothing is sent to customers without human approval, " explained Mehta.

“There's a review interface where teams can edit every AI-generated message before it's sent. Shiplog also built a conversational interface. Users can simply ask the platform, "Show me customers that are good candidates for expansion this week," and Ada analyses the available information before returning recommendations.

Building a defensible AI platform

While Shiplog is broadly focused on SaaS companies, it has gained the most traction in the fintech and cybersecurity sectors, specifically businesses with very large customer bases where you simply can't dedicate customer success managers to every account.

Mehta shared:

“Most companies naturally prioritise their highest-value customers—the accounts paying significant amounts of money. Those customers receive dedicated account managers and customer success teams.

But there are thousands of other customers who might not generate as much revenue individually, yet they're still incredibly important to the business. That's where AI agents become valuable. They allow companies to provide personalised engagement to customers who otherwise wouldn't receive that level of attention. That's why fintech and cybersecurity have been such a good fit.”

In terms of defensibility, Shiplog is focused on a specific vertical within software, which allows it to accumulate industry-specific knowledge over time. Mehta asserts:

“We understand which messages convert, where customers typically drop off in a fintech onboarding journey, how those funnels can be improved, and what customer behaviour looks like across multiple companies.”

Over time, Shiplog builds a detailed understanding of individual customers — their behaviour, preferences and interactions across different products. 

“Looking five years ahead, every tool a customer uses could contribute to a persistent layer of personalisation. That long-term customer memory becomes incredibly valuable.”

Ahead of its public launch, Shiplog ran two pilot programmes to validate the platform. In the first, it analysed more than four million events across around 1,000 customers, building a complete 360-degree view of every customer while proving the platform's ability to operate at scale.

Personalisation stops being a feature and becomes the default operating system for customer software, where every interface, recommendation, message and workflow adapts continuously to the person on the other side of the screen.