Skip to main content New Conversion Agents are live Read more
Conversion Box

Enterprise marketing control, built for Box.

Conversion is the only only marketing automation platform that can unify your data across BigQuery and Salesforce, run agents directly on top of it, and build the custom enterprise controls you need.

3 to 1

Marketo, Iterable, and Segment consolidated into one platform

5 days to hours

From a list request to a live, QA'd campaign

7.5M+

End-user records handled natively, with room up to 20M

Custom Build

FDE team builds custom enterprise requirements

One data layer

Every source, unified. Then agents on top.

Salesforce, BigQuery, and Segment converge into one record per contact, enriched in real time. Once your data is unified and governed in Conversion, agents can act on all of it, instead of waiting days for a list to be pulled and handed off.

Salesforce

CRM

BigQuery

Warehouse

Segment

Segment

CDP

AI Agents

Enrichment

Unified Contact Record
live
SC

Sarah Chen

VP of Marketing · Acme Corp

Lead Score

0

email [email protected] Salesforce
lifecycle_stage Opportunity Salesforce
annual_revenue $4.2M BigQuery
plan_tier Enterprise BigQuery
active_seats 47 BigQuery
feature_adoption 89% Segment
last_active 2h ago Segment
intent_segment High-intent buyer AI Agent
Built for Enterprise

Data, seamlessly integrated and rigorously controlled

Your team controls exactly what Conversion can read and write, at the source and again on the way in. Nothing writes back to your warehouse unless you ask it to.

Your data

BigQuery Read only

A service account you own scopes which tables and fields Conversion can read from.

Salesforce Read / write

An integration user sets the objects and fields Conversion can read and write back.

In Conversion

Global sync settings

Decide which objects can be read or written at all. Admin permissions gate every change.

Field mappings and sync modes

Per field, choose prefer Salesforce, prefer Conversion, or two-way. Most fields stay read-only by default.

Conversion

Unified, governed data ready for agents

SOC 2 Type II

DPA in place

No training on your data or metadata

GDPR data purge on request

RBAC to the object and field level

Full sync and audit logs

Sandbox environment

99.99% uptime SLA

Built for Box

We build the Box integrations no one else will

Our forward-deployed team builds native connections into your own systems as part of the deployment. Two we will commit to for Box.

Box + CONVERSION

A brand-guideline agent over Box Hub

Connect your Box Hubs over MCP so an agent checks every asset against your brand guidelines and PMM tonality, and flags anything off-brand before it goes out.

Box Hub
Guideline agent
On-brand
YOUR MODELS, IN PLATFORM

Run Box's own AI models inside Conversion

Swap our default LLMs for Box's internal model deployments, so every AI action runs inside the boundaries your security team has already approved.

Default Conversion LLMs
Box internal models

No other vendor on your shortlist will build this for you.

How we compare

Conversion vs Inflection vs HubSpot

The honest read on the three platforms in your evaluation, against what Box actually needs.

Capability Conversion Inflection HubSpot
Data foundation
Native BigQuery warehouse sync Limited New, limited
Consolidates Marketo, Iterable, and Segment Partial
Accounts, opportunities, and custom objects CRM-synced Standard objects
Real-time bidirectional Salesforce sync Batch
AI and automation
Agents that build full campaigns end to end Lifecycle only
Recurring, self-optimizing agents Partial
Bring your own AI models
Artifact-first human approval on every output Partial n/a
Enterprise and IT
RBAC to the object and field level Role-based Role-based
No training on your data by default Varies Opt-out
Custom integrations built for you (Box Hub, your models)
Partnership
Forward-deployed engineers on the migration Via partners
Pricing model Marketable contacts Seats + contacts Tiered + contacts
Proven at scale

Your scale is well within range

Box's 7.5M end-user records are handled natively, no separate database tier required. Conversion Enterprise users are managing up to 30M contacts across SFDC & warehouse.

Box: 7.5M Conversion range: up to 20M
Plaid

We're consolidating Plaid's GTM stack onto Conversion under high-stakes security control.

Productboard

Enterprise, running a large migration onto Conversion with our forward-deployed team.

Forward-deployed

An engineer who becomes part of your team

Conversion ships with a San Francisco-based forward-deployed engineering team. They are technical and marketing-fluent, available in Slack, and they run the migration with you so the cutover is never the bottleneck.

Day 0

You grant access

We connect to Salesforce and BigQuery through scoped, controlled credentials you own.

Days 3 to 5

A large share is already rebuilt

Our forward-deployed engineers stand up your data model and migrate a big chunk of campaigns within the first week.

90 to 180 days

Full migration, live

Run Conversion alongside Marketo until you are ready to cut over, with nothing left stranded.

See it running on Box's data

Book a working session with our team. We will tailor the demo to your stack and show you exactly how the migration would run.

FAQ

Frequently asked questions

We query your warehouse in near real time and do not charge for those queries. Contacts and standard objects from Salesforce and the warehouse are stored in Conversion so we can unify and act on them. You control what is synced and what is written back, down to the field level.

Pricing is based on marketable contacts. If a contact sits in the platform but never gets marketed to, you are not billed for it. AI usage is unlimited, so the number is predictable in advance based on your marketable database size.

Yes. BigQuery access runs through a service account you control with table-level scoping and read-only options. Salesforce uses an integration user plus global object-level settings and per-field sync modes, so most fields can be set to prefer Salesforce and never be overwritten.

No. By default we never train on your data, and that includes metadata and workflow inferences. We have data-retention policies and a CDR in place across our LLM providers, and you can opt out of any model use entirely.

Any BigQuery compute cost comes from Google, not from us, and in practice it is nominal next to the cost of the software it replaces. We are happy to walk through expected consumption with your team during evaluation.

Let's prove it on Box's data

Run a pilot alongside your current stack, take one workflow out of Marketo, and see the value before the contract is up. Our team handles the heavy lifting.