Agentic marketing platforms, surveyed honestly
An agentic marketing platform is one where AI agents execute work within permissions rather than assist a human operator. Every vendor now claims the phrase, so this survey sorts by the only question that separates them: what is the agent actually allowed to do, and on whose data?
We are not neutral: the build-your-own row at the bottom is the one we practice, on Kissmetrics as the data layer, and we say so where it matters. Everything else here is written the way we would want a rival to write about us.
How to judge any of them
- 1. What the agent may touch
- Summarise, recommend, or act? Ask for the list of actions the agent can take without a human building something first. This one answer sorts the market.
- 2. Whose data it reasons over
- Suite agents are excellent on their own cloud's data and blind outside it. Your stack's shape decides whether that is a feature or a cage.
- 3. The verify stage
- Does the platform show what the agent did and what it changed in the numbers? A platform that cannot answer 'what did your agent earn last month' is selling activity.
- 4. Exit cost
- Agent configurations, prompts and learned behaviour rarely export. Assume lock-in at the agent layer is total, and weigh it accordingly.
The landscape
Three groups, and the honest clause on every row: when that platform is the right call.
| Platform | Group | The right call when |
|---|---|---|
| Salesforce (Agentforce / Marketing Cloud Next) | Full suite | Your CRM and marketing data already live in Salesforce and procurement prefers one throat to choke; the agents are strongest where the data never leaves the cloud. |
| Adobe (Experience Platform AI) | Full suite | Enterprise content and journey operations on an existing Adobe contract, with a team to run it. |
| HubSpot (Breeze) | Full suite | Mid-market, HubSpot-centric, and you want approachable agents inside the tool marketing already lives in rather than best-of-breed depth. |
| Braze | Lifecycle messaging | Cross-channel lifecycle messaging is the job and you want agents optimising sends, journeys and experiments inside it. |
| Optimove | Lifecycle messaging | CRM-marketing-led retention businesses (gaming, retail) wanting agentic campaign selection on top of a customer model. |
| Netcore | Lifecycle messaging | Volume messaging economics matter most and the agent layer rides on deliverability infrastructure you are already buying. |
| LangGraph / CrewAI / agent SDKs + data layer | Build your own | You want agents across the whole stack, not inside one vendor: frameworks orchestrate, model APIs reason, and an agent-operable analytics layer supplies the ground truth. Cheapest to start, most honest about failures, yours to maintain. |
No pricing column, deliberately: every suite and lifecycle platform here quotes rather than lists, and a survey that invented their numbers would be wrong within a quarter. The questions that make a quote comparable are in the section above.
Where we sit, stated plainly
We run the build-your-own pattern: orchestration in code, models by API, and Kissmetrics as the layer every agent reads and operates: person-level events, revenue attribution, and a chat, API and CLI that expose the same actions to agents as to people. The build logs on the blog are that pattern running in public, configurations included. If a suite above fits your gravity better, take the suite; the criteria do not care what we practice.
Questions
- What is an agentic marketing platform?
- Software where AI agents execute marketing work end to end within permissions: reading the data, deciding what needs attention, and acting (adjusting campaigns, drafting content, updating audiences) rather than only assisting a human operator. The honest test is what the agent is allowed to do without a person building something first.
- Which agentic marketing platform is best?
- It depends on where your gravity already is. Deep in one enterprise cloud, the suite agent (Salesforce, Adobe) that lives on your data wins by default. Lifecycle-messaging-centric teams get the most from Braze, Optimove or Netcore's agents. Teams that want agents across their whole stack rather than inside one vendor's walls end up building on frameworks plus an agent-operable data layer, which is cheaper and more flexible at the cost of assembly.
- How much do agentic marketing platforms cost?
- The suites and lifecycle platforms are quote-based enterprise contracts; none publishes agentic pricing separately, and quotes vary widely by data volume and modules. The build-your-own path is the only one with public arithmetic: framework (open source), orchestration runtime, model API usage, and the data layer. Treat any specific suite price you read as an anecdote.
- Can I build an agentic marketing stack instead of buying one?
- Yes, and below enterprise scale it is increasingly the rational default: an orchestration framework, model APIs, and an analytics layer the agents can operate through an API. The trade is assembly work and ownership of failures in exchange for stack-wide agents, portability, and costs that scale with usage instead of with a contract.
Related: Agent orchestration · Agentic loops · MCP