By George Deglin, Co-founder & CEO
For more than a decade, customer engagement platforms have offered the same deal: we hand you a powerful dashboard, and you supply the people, the time, and the expertise to run it. The more capable the platform got, the more of all three you needed. Sophistication meant headcount. Growth meant more budget. UIs got more buttons and more screens.
That deal is ending. Dashboards are dying.
We're now building toward Autonomous Lifecycle Marketing (ALM). The premise is simple, even though the engineering behind it isn't: you connect OneSignal to your product, and our software plans, sends, tests, and improves the messages that onboard, engage, and retain your users. You set the goals, the brand, and the guardrails. OneSignal does the work. All the time, continuously. While you’re outside. While you're eating lunch. Even while you sleep.
This is a real bet, and I want to be direct about it. We believe the center of gravity in our category is moving from marketers operating software to software operating the full customer lifecycle, with people and AI agents supplying goals, context, and approvals. The platforms built for the dashboard era were designed around a human clicking through every step. We're designing for a world where most of that work is done for you, and you step in where your judgment actually matters.
The Autonomy ladder
Nobody hands the keys to an autonomous system on day one, and they shouldn't. Trust is earned in steps, and so is autonomy. Here's how we think about the climb:
- L0 — Manual. The dashboard. A marketer drives every click, and the team scales by hiring. This is where the entire category has lived.
- L1 — Assistance. AI drafts copy, suggests audiences, and explains your analytics in plain language. Input is required. You review and ship. This is live today.
- L2 — Expert Assist. AI is proactive.. It recommends campaigns to fill lifecycle gaps, flags mistakes, and checks your work. Input is still required.
- L3 — Agentic Autopilot. You delegate bounded work, and AI does it within the scope and rules you've defined. You approve in large chunks.
- L4 — Full Operator. You set your goals and the guardrails. AI plans, sends, tests, and improves the campaigns against it, and reaches out only when it needs you.
L4 is the vision, not the present. But there’s real value from AI today, and we see that value growing daily.
We're already climbing
Two products we've shipped are the first real steps up that ladder, and they reflect two different ways our customers want to work.
OneSignal AI offers assistance inside the platform. You can ask how a campaign performed last week, have it draft a journey to re-engage lapsed users, tell it to send a thank you email to recent sign ups users: all in plain language. It knows the page you're on and can answer questions about what you’re actively working on at. It can build a journey, compose copy, and provide benchmarks and industry best practices. We’ve also introduced Brand Center, to ensure you can reflect your brand, logo, tone, color palette, etc. even when you’re moving at the speed of automation. Today, OneSignal AI graduates to Open Beta.
“I love OneSignal AI. If I've got a Journey with seven push notifications, I can just ask how my onboarding is performing and it spits out all the results, instead of digging through templates and reporting manually. One prompt probably saved me two hours of work, and it came back in 10 seconds.”
- Laura Allison, CRM Manager, DC Thomson
The OneSignal MCP Server brings OneSignal's capabilities to the AI tools and agents you already use. Built on the open Model Context Protocol (MCP), it lets AI assistants like Claude, ChatGPT, Cursor, and Copilot securely operate OneSignal on your behalf. You can take actions across the entire OneSignal API, with one deliberate exception: we don't allow delete actions. And as a general safeguard, high-impact actions like sending a message require explicit confirmation.
The MCP server unlocks something the dashboard never could: orchestration across your entire tech stack with no 1st party integrations required. In a single conversation, an agent can pull behavioral data from your data warehouse, find where new users drop off in an analytics tool, build a segment in OneSignal, draft the copy, and schedule the sends: coordinating across every tool that speaks MCP, not just ours. That's what an AI-native workflow actually looks like, and it's in Open Beta now with availability in both Cursor and ChatGPT marketplaces.
Updates to the Free Plan: Continuing Our Mission of Democratizing Customer Engagement
OneSignal has always allowed users to sign up and use the product for free. This enables us to collect feedback from a broad set of users and use cases that forces us to build a product that is equally intuitive and powerful. As we bring AI features to market we want to make sure those can be used by everyone.
Free plan users will have access to additional journeys, unlimited message steps, and more data tags alongside our new AI capabilities to drive powerful, personalized automated lifecycle marketing. To sustainably offer these additional features we are introducing a MAU (Monthly Active User) cap on the free plan. These changes will be rolled out in the coming months.
Our differentiated approach
Outcomes first. We believe many other customer engagement platforms are racing to win on ever-deeper 1:1 personalization. We think personalization can be made infinitely deep, but not without hitting a point of diminishing returns. What our customers actually want is the outcome: more activated users, more revenue per install, less churn. We optimize for the best result with the least effort, not the most elaborate configuration with the most.
Built for mobile-first teams. We've always built for mobile-first B2C companies with serious lifecycle needs. We started as a mobile game studio and built our own software to engage users. Today we're the best tool for companies doing that same work: driving activation, monetization, and retention with their mobile users.
The system gets smarter at scale. Our low barrier to entry means we work with more businesses, of greater variety, than our competitors do. That scale is a learning advantage: we understand the goals and tactics that achieve outcomes with AI across the spectrum of businesses that we work with. Our unique understanding of which lifecycle strategies actually produce results is an advantage we share with all of our customers.
The original version of our platform already drives the outcomes ALM is built to compound: Cashea lifted first-month activation 43% and Letgo reactivated more than 28,000 dormant users from a single campaign. Now imagine that work running continuously, on its own.
The promise
We’re building a lifecycle marketing function that's future-proofed for an AI-native world, with lower operating cost, less manual work, and outcomes that improve as both AI and our platform evolve. The send button will always exist. But more and more of what surrounds it (the planning, the testing, the tuning, the late-night "did that campaign actually go out, and how did it do?") will be handled.
OneSignal AI and the MCP are continuing to improve and we're learning from their real usage every day. More capabilities are on the way. If you're already building with AI day to day, this is the moment to get in early and tell us what works and what doesn't. We're listening, and we're building fast.
The dashboard era was about giving you better tools. This next era is about giving you the outcomes those tools were always for, and handing back the time you spent operating them.
