NEW · AI TRUST ARC BY ÉXITO AI

Your customers aren't churning.
They're losing trust in your AI.

The first CS intelligence layer that maps every account to their AI trust stage and tells your team exactly what to do before the churn signal ever fires.

See How it Works ->

First insight in under 30 minutes · Integrates with your existing CS stack

6 wks

avg. advance warning before churn

71%

of AI CS teams detect churn <2 weeks out

arc stages mapped per account, in real time

AI-specific trust signals no other platform tracks

The problem

Your health score was built for a product that doesn't make decisions.

Gainsight. ChurnZero. Totango. They're excellent lifecycle tools at tracking SaaS health. But your product is missing signals that help predict churn, and provides success plays to mitigate churn. Now with AI, accounts churn differently. The 5 phases and associated signals can run as human in-the-loop or human on-the-loop.

Customers churn because they lost trust in the AI outputs not because of price, competition, or product quality. Standard metrics never see it coming.

Login frequency and NPS look fine while AI feature usage erodes silently. By the time your health score turns red, the customer has already decided.

There are no playbooks for the AI trust conversation explaining model outputs, handling skepticism, and rebuilding confidence after a bad recommendation.

When the internal champion leaves, AI trust doesn't transfer to the replacement the way product knowledge does. And nobody has a protocol for it.

The framework

Every AI customer follows the same trust arc.

AI Trust Arc maps exactly where each customer is on that journey and what your team needs to do at each stage to advance them, not lose them.

01 / NOVELTY

Novelty/Connect

Days 1–14

Exploring. Excited. Usage is high but trust hasn't been earned yet. The most dangerous stage for complacency.

Medium risk

02 / SKEPTICISM

Skepticism

Days 15–45

Questioning AI outputs. Most AI churn is born here silently weeks before your health score notices.

High risk

03 / CALIBRATION

Calibration/Adoption

Days 45–90

Developing a working theory of when to trust the AI. Ends in acceptance or quiet abandonment CS determines which.

Medium risk

04 / INTEGRATION

Integration/Realize

Days 90–180

AI is in the daily workflow. Trust is built. This is where expansion conversations are viable. This is where NRR grows.

Low risk

05 / ADVOCACY

Advocacy/Expand

Day 180+

Sharing AI outputs internally. Championing the product. Your most powerful growth engine left unactivated by most CS teams.

Expansion ready

The signals

8 signals no other CS platform tracks.

Built from the ground up for AI products. Each signal maps to real customer behavior that predicts trust trajectory, not just product usage.

S1 · AI Feature Activation Rate

Are they using the AI, or just the wrapper?

% of sessions where AI-specific features are used vs. non-AI features. Danger flag: below 20% by day 14.

Source: Segment · Mixpanel · Amplitude

S2 · Output Act-On Rate

Are they trusting the AI's recommendations?

% of AI outputs acted on without immediate override. Danger flag: override rate above 60% in any 7-day window.

Source: Product event data

S3 · Manual Reversion Frequency

Are they reverting to manual workflows?

How often a user starts an AI workflow then abandons it for manual. Danger flag: 3+ reversions/week for 2 consecutive weeks.

Source: Product event sequences

S4 · Support Ticket Sentiment Shift

Are tickets shifting from learning to distrust?

Ratio of "why did it" tickets vs "how do I" tickets. The earliest verbal signal that trust is breaking down.

Source: Intercom · Zendesk

S5 · AI Output Share Rate

Are they sharing AI outputs internally?

How often AI-generated content is exported or forwarded. The best proxy for internal trust propagation. Stage 4/5 entry signal.

Source: Product event data

S6 · Time Between AI Actions

Is engagement growing or shrinking?

Average gap between consecutive AI-assisted actions. An increasing gap signals workflow disengagement weeks before NPS drops.

Source: Product event timestamps

S7 · Champion Engagement Velocity

Is the champion still your advocate?

Response rate and latency on CS touchpoints. Champion silence is the single most predictive leading indicator of account regression.

Source: HubSpot · Gmail

S8 · Seat Expansion Within Account

Is trust spreading across the organization?

Distinct users accessing AI features over 30-day rolling window. Growth signals stage 4–5 progression; contraction signals regression.

Source: Product analytics · CRM

How it works

From data to intervention in under 30 minutes.

01 / CONNECT

Connect your existing data

Plug in Segment, HubSpot, or Intercom in minutes. No custom integration work. CSV import always available as a fallback. First arc stage visible in under 30 minutes.

02 / MAP

Every account gets an arc stage

The signal engine evaluates all 8 trust signals every 4 hours and assigns each account to its current arc stage. Regression risk is calculated automatically.

03 / ACT

CS team gets the right playbook

Each arc stage has a pre-built intervention playbook built from the AI Churn Benchmark data. The right action, at the right moment, for where each customer actually is.

04 / ALERT

Real-time Slack and email alerts

When an account starts regressing, your team is notified in Slack with context: which signals triggered it, ARR at risk, and the recommended intervention, before it's too late.

Integrates with your existing stack — no rip and replace - deveolpment with human-in-the-loop and human-on-the-loop

Segment

HubSpot

Intercom

Zendesk

Slack

Salesforce ↗ v2

ChurnZero ↗ v2

Gainsight ↗ v2

Mixpanel ↗ v2

Amplitude ↗ v2

Gong ↗ v2

CSV Import

FAQ

Common questions.

Does this replace ChurnZero or Gainsight?

No, and that's intentional. AI Trust Arc is an intelligence layer that sits on top of your existing CS platform. It does one thing those platforms don't: track the AI-specific trust signals that actually predict churn for AI products. Most customers use AI Trust Arc alongside their existing CS tool, not instead of it. v2 will include native embeds into ChurnZero and Gainsight so arc stage data flows directly into your existing workflows.

What if we don't have Segment or a product analytics setup?

CSV import is always available as a fallback. You can upload a spreadsheet with proxy signal data (login counts, support ticket counts, last feature usage date) and get arc stage assignments immediately. You'll have partial signal coverage, but a partial arc view is infinitely better than no arc view. Most customers start with CSV and migrate to live integrations within the first two weeks.

How is the arc stage determined? Is it AI or rules-based?

v1 is rules-based. A transparent, auditable set of signal thresholds that trigger stage transitions. You can see exactly why an account was moved to a new stage. v2 will introduce ML-based prediction trained on aggregate outcome data as we accumulate enough signal history. We prefer explainability over opacity, especially in a trust-focused product.

What types of AI companies is this built for?

AI Trust Arc is designed for B2B AI-native SaaS companies, series A through growth stage, where the CS team manages 50 to 2,000 customer accounts and the product's core value proposition involves AI-generated outputs that customers need to trust and act on. It's not designed for AI infrastructure companies, internal tooling, or consumer products.

What does "first insight in under 30 minutes" actually mean?

It means your first account's arc stage is assigned and visible in the dashboard within 30 minutes of connecting your first data source. You don't need a full integration suite to get value from day one. Connect your HubSpot or upload a CSV, and the engine will assign arc stages immediately based on the available signals, flagging which signals are active and which are pending additional data.

Stop finding out about churn at renewal.
Start seeing it six weeks early.

First arc stage in under 30 minutes.