Tome's founders raise $47M for Lightfield's agent-first CRM
Keith Peiris and Henri Liriani raised the Series A after replacing their viral presentation product with a system of record built for software agents.
By RuntimeWire Staff · Published
Primary source: PR Newswire
Why it matters
Peiris and Liriani shut down Tome after it reached 25 million users and are now betting that a CRM holding customer context can become the system software agents use to act on a company's behalf.
Lightfield's founders, Keith Peiris and Henri Liriani, have raised a $47 million Series A for the former Tome team's CRM built for software agents.
The San Francisco-based CRM developer announced the financing on September 9. The available announcement does not establish the lead investor or participating funds. Lightfield did not attach a valuation to the announcement.
Tome had raised about $81 million before the pivot, including a $43 million Series B led by Lightspeed Venture Partners in February 2023 at a reported $300 million valuation. Greylock, Coatue and Audacious Ventures participated in that earlier round, according to Contrary Research's history of the company.
Calling the new financing a Series A gives Lightfield a clean stage marker for a new product. It does not erase the capital, team or operating history accumulated under Tome. The announcement does not say whether the $47 million is entirely incremental to the previously reported $81 million, leaving Lightfield's cumulative capitalization unclear.
The founders traded reach for context
Peiris and Liriani founded the business in 2020 after working at Meta. Peiris studied nanotechnology engineering at the University of Waterloo and later held product roles across Facebook, Instagram, Glossier and Citizen. At Instagram, he worked on messaging and camera products, including Instagram Direct. Liriani previously worked at Meta before co-founding Tome with Peiris, according to Contrary Research's history of Lightfield.
The pair first applied that experience to Tome, which generated presentations from prompts and reportedly reached roughly 25 million users. Peiris said Tome was shut down because it was not mission-critical and the founders lacked enough understanding of their customers and companies, according to Contrary Research's history of Lightfield. They then started Lightfield as an AI-native CRM focused on customer context.
Sales conversations, account history, meeting transcripts and commitments were spread across inboxes, calendars, call recorders and conventional CRM fields. That problem became the basis for Lightfield, which the founders launched publicly in November 2025 after about a year in stealth.
In their founding essay, Peiris and Liriani argued that a CRM should preserve the chronology and reasoning behind customer relationships instead of storing conclusions entered by sales representatives. They wrote that "automation is a consequence of understanding - not the starting point."
The pivot amounts to a bet that proprietary business context will prove more valuable than another general-purpose AI interface. Tome helped users create an artifact. Lightfield wants to own the record agents consult before they research an account, draft an email, update a deal or decide which prospect deserves attention.
Lightfield wants the database and the agents
Lightfield says its CRM connects to email, calendars, meetings and other business systems, then creates and updates records for contacts, accounts and deals. Lightfield describes the resulting history as a "world model" of a business, preserving who said what, when it happened and how a relationship changed.
The platform's developer documentation lists APIs, a model context protocol connector, Python, TypeScript and Go SDKs, and a command-line interface. Lightfield also provides a code-execution environment for agents and customer-built automations.
The architecture puts Lightfield in competition with two groups at once. Salesforce and HubSpot already control customer records inside established businesses and are adding agents to those databases. Newer CRM developers including Attio and Clarify are pursuing cleaner, more automated alternatives for teams willing to reconsider the incumbent systems.
Lightfield's answer is to capture underlying interactions automatically and treat conventional fields as generated views. That approach could reduce the administrative work that has long made CRM deployments unpopular with sales representatives. It also asks customers to place sensitive communications and commercial history inside a young platform whose value increases with the breadth of access it receives.
The migration path remains unfinished. Lightfield's integration documentation says direct CRM integrations are not yet available. Customers can run Lightfield alongside an existing CRM, synchronize selected information through webhooks or build connections using the API. That limitation matters when Lightfield is selling itself as the foundational record for agent-run businesses: established teams rarely discard Salesforce or HubSpot before a replacement can reliably exchange data with them.
Five thousand signups, with the harder metrics still ahead
Lightfield says more than 5,000 companies have signed up since the November 2025 launch. The figure describes signups. Lightfield has not supplied corresponding figures for active customers, paying accounts, retention or annual recurring revenue.
The distinction matters because Lightfield's pricing has moved beyond a lightweight experiment. Its published plans begin at $89 per user each month for a Starter tier capped at three users. Pro costs $999 per workspace each month, while Growth costs $1,999 and includes more seats and credits for agent work. Enterprise contracts use custom pricing.
The Series A gives Peiris and Liriani time to convert curiosity into recurring usage. The financing backs Lightfield's team, which has already demonstrated an ability to attract millions of users and the willingness to shut down a popular product after concluding that it was not mission-critical.
Lightfield now faces a different test. CRM buyers care about data integrity, permissions, migrations and predictable execution. Viral distribution will not settle those questions. The founders have chosen a narrower market where successful products can become deeply embedded, and where replacing the incumbent requires earning access to the information that runs a customer's revenue operation.