Tofu

Tofu is an AI personalization platform built for B2B marketing teams that run account-based campaigns. It connects to your CRM, ABM list, and content library, then generates personalized email, landing page, and ad copy variants for each named account based on what your team already knows about them.

So the use case is the part of ABM that breaks at scale. A team can hand-write a custom email to ten target accounts. At two hundred, the work falls apart. Tofu sits in that gap. It produces account-tailored variants that read like the SDR or PMM wrote each one, without the team actually doing it for every contact.

The personalization engine pulls from:

• CRM fields (industry, stage, ARR band, tech stack)
• Account intelligence (recent funding, headcount, growth signals)
• Past engagement (pages viewed, content downloaded, emails opened)
• Public sources (job postings, news, executive moves)
• Your existing content library so messaging stays on-brand

Then the output layer renders that data into campaign assets:

• Cold email and sequence copy customized to the account's situation
• Landing pages where the headline, copy, and proof points shift per visitor
• Ad headlines and primary text variants per industry or persona
• Direct mail and outbound video scripts for high-touch tiers

Common workflows by role:

• Demand gen teams run named-account campaigns where each tier gets a different level of personalization depth
• ABM specialists scale one-to-few campaigns into one-to-many without losing voice
• Field marketers personalize event-driven outreach for regional pushes
• Product marketers test which messaging angles land per industry or persona

Plus the orchestration side ties into existing tooling rather than replacing it. Variants push into the marketing automation platform, sequencing tool, or ad system the team already uses. Tofu generates the copy and the targeting logic. The existing stack still ships the campaign.

A few specifics worth knowing:

• Connects to Salesforce, HubSpot, Marketo, and Outreach as the core CRM and engagement layer
• Reads from 6sense, Demandbase, or Clearbit data if those signals already exist in the stack
• Pulls content from a connected library so generated copy uses approved messaging
• Reporting ties campaign performance back to which personalization angles performed

Now, the fit narrows fast outside B2B with named-account motions. For PLG or self-serve businesses where pipeline comes from broad acquisition rather than targeted accounts, the personalization engine doesn't have enough to chew on. The signal density only pays off when there's a meaningful account list to personalize against.

Paid subscription with enterprise pricing, scaled by accounts personalized, integrations, and team seats.

Best for B2B marketing teams running ABM or named-account campaigns at the scale where hand-personalization stops being feasible. Not ideal for PLG or broad self-serve motions where the buyer isn't on a target list to begin with.