How does Joryio use AI to personalize customer journeys?

Updated

Joryio uses AI to decide the best channel, content, and send time for each customer, then lets you approve or constrain what actually goes out. It works on top of behavioral triggers — like an abandoned cart — and measures results against holdout groups so you can see lift rather than assume it. If you want AI to suggest next actions while you keep final control, this fits; if you want fully autonomous sending with no human review, it does not.

What the AI actually decides

Joryio's product page describes AI as tailoring three things inside a journey:

  • Channel — which of email, SMS, WhatsApp, push, in-app, or Wallet pass to use for a given person
  • Content — the message itself
  • Send time — when to deliver it

The page shows this as an "AI DECISION" step inside a journey: "Best channel each measured vs holdout." So the AI's choice is not just applied — it is compared against a holdout group to see whether it performed better.

How a personalized journey is built

The platform describes a five-step loop: Know → Understand → Decide → Engage → Measure, and then feeds the result back so "every interaction makes the next one smarter."

A concrete example from the page, an abandoned cart journey:

  1. Trigger — Cart abandoned. 12,480 people entered over 7 days.
  2. AI decision — pick the best channel, measured against a holdout.
  3. Engage — for example, email with subject "It's been a while" (42.1% open), or WhatsApp with an offer "if margin allows" (9.8% reply), or a Wallet pass with +200 points that updates live (3,104 installed).
  4. Measure — revenue attributed to the journey: $412,380, +31% lift vs control group, +6.2% measured.

Two things stand out here. First, the AI choice is bounded by your rules — the WhatsApp offer only goes out "if margin allows." Second, results are reported against a control group, so personalization is judged by incremental effect, not raw opens.

You stay in control

The page states directly: "AI to help you move forward. You stay in control." In practice this means:

  • AI proposes a channel, content, or timing; the journey logic (rules, triggers, conditions like margin) still governs whether it sends.
  • You can run a control group (the example uses 10%) to hold back a portion of the audience and measure true lift.
  • Journeys surface review prompts — the page shows a "Welcome journey · needs review" alert — so AI-driven changes are visible rather than silent.

If your team needs every message pre-approved, build that as a rule or review step. If you are comfortable letting AI choose within guardrails, the holdout measurement tells you whether it is working.

What you need before this works

Personalization depends on data flowing in first. The platform's own framing: "Know — Profiles · Events · Purchases · Preferences." The abandoned-cart example needs cart events, purchase history, and contact profiles. Joryio lists three capabilities that feed this:

Capability What it does How it's billed (per page)
Customer messaging Journeys, campaigns, messages By stored contact count
Analytics Funnels, retention, revenue attribution By event volume
Data streaming Stream events to your data warehouse By records transferred

All three are described as included in Joryio, with platform fees based on Engaged Profiles. Pricing details and a price estimator are linked from the site, but the page does not state specific figures — check the pricing page for current numbers.

When this approach fits

Choose Joryio's AI personalization if:

  • You already collect behavioral events (carts, purchases, preferences) and want them to drive sends.
  • You want AI to pick channel/content/timing but keep rules and human review in the loop.
  • You care about measuring lift against a control group, not just open rates.

Look elsewhere if you need fully autonomous sending with no guardrails, or if you have no event data to feed the system — the AI decisions depend on that signal.