Data 360
Data 360 vs Twilio Segment: the CDP choice when the consumer is an agent
Both now claim to be the best data layer for AI agents, and they arrive at it from opposite ends. Segment collects events and builds profiles from scratch; Data 360 reads data where it already lives and grounds the agent in the same platform. Here is where each one wins, the pricing models that decide the bill, and the question that settles it.
Two customer data platforms now open their pitch with the same line: we are the best foundation for your AI agents. Both are telling the truth. They just describe different buildings.
Twilio Segment says it because it has collected clean, real-time behavioural events for a decade and can feed a profile the instant a user clicks. Salesforce Data 360 says it because the agent you are grounding probably runs on Salesforce already, and the data is one query away, with no copy and no integration.
If you are choosing between them in 2026, start from that difference. Segment is an event-collection platform that builds a profile from scratch out of what you instrument. Data 360 is a Salesforce-native lakehouse that reads data where it already sits and resolves it into a profile the platform can act on.
Pick Segment when your richest signal lives in product and web events across many non-Salesforce apps and you want developer control over the pipeline. Pick Data 360 when your customer record already lives in Salesforce or a cloud warehouse and the thing consuming it is Agentforce.
The Adobe Real-Time CDP comparison is a different one. Data 360 against Adobe pits two suite CDPs built for the same marketer against each other. Segment is the developer’s CDP, and the architectural gap is wider.
Two architectures, drawn from opposite ends
Segment was founded in 2011 and bought by Twilio in 2020. Its core idea has not changed: you install a tracking library, you emit track, identify, and page events from every surface a customer touches, and Segment turns that stream into profiles and routes it to wherever you want it. The product splits into four parts that you license roughly in this order:
- Connections, the pipeline. Hundreds of prebuilt sources and destinations (Twilio puts it around 450) that collect events and fan them out to analytics, ad, and messaging tools.
- Unify, the identity layer. Identity resolution, Profile Explorer, the Profiles API, and Profiles Sync, which is what turns raw events into a single customer profile.
- Engage, the activation layer. Audiences and traits you build on top of Unify and push to channels.
- Protocols, data governance and schema enforcement, sold separately.
Data 360 starts from the opposite assumption: the data already exists, and copying it is the problem. Its defining architectural choice is zero copy. Records in Snowflake, Databricks, BigQuery, or Amazon Redshift can be queried where they live, mapped into Data 360’s model, and used for segmentation or grounding without a nightly sync moving them anywhere.
Where Segment’s instinct is to collect and centralise, Data 360’s instinct is to federate and leave data in place. That single difference cascades into almost everything else.
Identity resolution: build a profile vs reconcile one
Both platforms unify profiles, and both will tell you identity is their hard-won strength. The mechanics differ.
Segment’s Unify joins identifiers it sees in the event stream: a userId you set at login, the anonymousId in the browser before login, an email captured in a form. Identity is assembled forward, from the events as they arrive, which is why Segment is strong the moment a customer is active and weaker on a customer who has been dormant since before you instrumented anything.
Data 360 resolves identity with an identity resolution ruleset: match rules that decide which records describe the same person, and reconciliation rules that decide which value wins when they disagree. Reconciliation is the part Segment has no real equivalent for. When three systems hold three spellings of a name, you tell Data 360 to take the most recently updated value, the most frequent, or the one from your highest-priority source.
Event joining is a different job. Reconciliation is the job you have when your truth is spread across a CRM, a billing system, and a warehouse rather than arriving as a clean stream.
So the identity question is really a data-shape question. If your customer is best described by what they do, Segment resolves it well. If your customer is best described by records scattered across back-office systems, Data 360’s reconciliation is the tool built for that mess.
The pricing models are not comparable line for line
Teams get surprised here, because the two meters measure different things.
Segment prices on monthly tracked users (MTUs). The Team plan lists at $120/month for 10,000 MTUs, with overage tiered down as volume rises, and the full CDP tier that adds Unify and Engage moves to custom pricing.
The trap is the anonymous visitor. Segment counts a userId, then adds any anonymousId that never got joined to one during the month, so a high-traffic B2C site with a lot of logged-out browsing can generate far more MTUs than it has customers. MTU pricing rewards B2B, where most traffic is known, and punishes anonymous B2C scale.
Data 360 prices on credits for the work done, not the people counted. Operations meter as rows processed divided by a million, times an operation rate: a query costs on the order of 2 credits per million rows, while identity resolution is far heavier at roughly 100,000 credits per million.
The detail that changes the maths is the free allocation. Data 360 credits come bundled with the core Salesforce editions, and the credit model and included allocations mean a Salesforce customer often starts with a large balance already paid for, drawn from the same pool as Agentforce. A standalone Segment contract is net-new spend. A Data 360 build frequently runs inside credits you already own.
Neither meter is cheaper in the abstract. MTU pricing is predictable and punishes anonymous volume; credit pricing is variable and punishes heavy, badly scheduled processing. You can only compare them against your own traffic and your own credit consumption discipline.
The warehouse question, where they have converged
The interesting part of 2026 is that both platforms learned the same lesson about the data warehouse, and met in the middle from opposite sides.
Segment added Reverse ETL and Linked Audiences. Reverse ETL pulls data back out of Snowflake, BigQuery, Redshift, or Databricks with SQL or dbt models and pushes it to business tools on schedules as frequent as every 15 minutes. Linked Audiences lets you build segments directly against warehouse tables, blending historical warehouse data with real-time behavioural signals. Segment, in other words, stopped insisting that everything flow through its own store and learned to treat the warehouse as a source of truth.
Data 360 started there. Zero copy against the big four warehouses is the native path, not a bolt-on, and the data model federates rather than ingests. The practical result is that both can now ground an agent on warehouse data.
The difference is how many hops it takes. With Segment the warehouse data rides a reverse-ETL schedule to a destination and then to the agent. With Data 360 the agent queries the resolved profile in the same platform it runs in.
When the consumer is an agent, proximity wins
Both vendors are circling the same shift. A CDP used to feed dashboards, ad platforms, and email tools. In 2026 it increasingly feeds an autonomous agent that reads a profile and acts on it, and that changes the scoring.
Twilio positions Segment as a top data platform for AI agents, API-first with real-time profiles and a large source catalogue. For an agent you build outside Salesforce, on your own stack, Segment is strong there, and its Profiles API is a clean, fast way to hand any application a unified profile.
But if the agent is Agentforce, Data 360 has an advantage Segment cannot copy: the agent lives in the same platform as the data. No integration to maintain between the profile store and the reasoning engine, no second system to secure, no copy to keep fresh. Data 360 feeds Agentforce through data graphs and retrievers that the agent reads natively, and the same sharing rules and Trust Layer govern both.
That proximity is the whole argument behind why a serious Agentforce build usually wants Data 360 underneath it. Grounding quality and governance stop being two separate projects when the data and the agent share a platform.
An agent can only reason about a customer it can resolve. The question is where you want that resolution to happen, and how far it has to travel to reach the thing acting on it.
How to choose
Strip away the marketing and the decision comes down to three questions about your own situation.
Start with where your richest customer signal lives. If it is product and web events across many non-Salesforce applications, Segment was built to capture exactly that, and its source catalogue and developer ergonomics are ahead. If it is records spread across a CRM, a billing system, and a cloud warehouse, Data 360’s reconciliation and zero copy are built for that shape.
Then ask what consumes the profile. A bespoke application or an agent on your own stack is a clean fit for Segment’s API-first model. Agentforce, Einstein, or Salesforce reporting removes an entire integration from the picture with Data 360.
Last, weigh what you already pay for. A Segment contract is new money. Data 360 credits often arrive bundled with editions you already bought. That does not make Data 360 free, because heavy processing still burns credits, but it changes the starting line.
The answer most Salesforce-centred teams reach is this. If the data and the agent both live in Salesforce, the integration you avoid with Data 360 is worth more than the source breadth you give up versus Segment. If your centre of gravity is a product analytics stack and a warehouse that Salesforce is only a spoke of, Segment’s event-first model earns its place, and reverse ETL now lets it reach the warehouse you already trust.
Choose the platform that sits closest to both your truth and the thing acting on it. Every hop between them is latency, cost, and one more system to keep in sync.
Understanding the basics
Is Salesforce Data 360 a customer data platform like Segment?
Both resolve data into unified customer profiles and activate them, so they compete for the same budget, but they are built differently. Segment is an event-collection CDP: you instrument your apps, it captures behavioural events and builds profiles from that stream, and it routes data to hundreds of destinations. Data 360 is a Salesforce-native lakehouse that reads data where it already lives, including zero-copy access to Snowflake, Databricks, BigQuery, and Redshift, resolves it with match and reconciliation rules, and grounds Salesforce AI directly. They overlap on outcome and differ on architecture.
Which is better for grounding an AI agent?
It depends on where the agent runs. For an agent built on your own stack outside Salesforce, Segment’s API-first real-time profiles and large source catalogue are a strong, clean fit. For Agentforce, Data 360 wins on proximity: the agent and the data share a platform, so there is no integration to build between the profile store and the reasoning engine, and the same sharing rules and Trust Layer govern both.
How do the pricing models differ?
Segment prices on monthly tracked users (MTUs), counting known users plus unjoined anonymous visitors, which makes it predictable but expensive for high-anonymous-traffic B2C sites. Data 360 prices on credits for work done, metered as rows processed times an operation rate, with credits often bundled into Salesforce editions you already own. The two meters measure different things, so you can only compare them against your own traffic volume and processing patterns, not line for line.
Can Segment read from a data warehouse the way Data 360 does?
Partly, and it is newer. Segment added Reverse ETL, which pulls warehouse data out with SQL or dbt and pushes it to tools on schedules as frequent as every 15 minutes, and Linked Audiences, which builds segments directly against warehouse tables. Data 360’s zero copy queries warehouse data in place as its native path rather than on a schedule. Both can now ground on warehouse data; Data 360 does it with fewer hops when the consumer is a Salesforce agent.
Weighing a CDP decision, or trying to get a data foundation trustworthy enough to ground an agent without a second platform in the middle? Talk to us. Getting the data layer right, so an agent is grounded rather than guessing, is exactly the work we do.
