// PYXIS ENGINE
// CUSTOMER SUCCESS
The customer's pole star. Churn before it happens.
Polaris is Ágora's Customer Success platform. It detects early churn signals before the client requests cancellation, calculating an auditable health score. It proactively manages retention and expansion through lifecycle marketing, loyalty programs and upsell recommendation. It delivers Customer Success teams the tools to operate: Customer 360, business reviews and deep retention analytics. Its Pyxis engine combines classical machine learning predictive models with LLM-based prescriptive reasoning.
Retention costs five to ten times less than acquisition.
// EARLY CHURN SIGNALS
Customers who leave don't announce it. They show signs first.
A slower response, a drop in volume, a change in tone. Polaris detects those signals while there's still time to act. It's not a Gainsight, it's customer success adapted to how an Argentine company actually works.
// WHAT IT SOLVES
Three pains it addresses.
Churn invisible until it's too late
Customers don't announce their exit. They cool down progressively: use the product less, answer fewer emails, delay payments. By the time they request cancellation, they're lost. Polaris detects the pattern earlier.
Lost expansion opportunities
A customer using the product at the max of their plan is a natural upgrade candidate. A customer who opened three tickets about a higher-plan feature is literally requesting the upgrade without knowing. Without a system detecting patterns, opportunities are lost.
Customer Success operating with Excel
Most medium LATAM companies manage retention with spreadsheets, generic newsletters and CSM memory. There is no unified Customer 360, no health score, no cohort analytics. When the CSM leaves, the knowledge is lost.
// CORE FUNCTIONS
What it does concretely.
Auditable Customer Health Score
Composite score per customer combining usage, engagement, inferred satisfaction, payment behavior and tenure. The user can see exactly why a customer scores 72 and not 89. No black box.
Unified Customer 360
Single view per customer with purchase history, all cross-channel interactions (inherited from Omnira if connected), NPS, tickets, lifecycle stage, calculated LTV and upcoming milestones.
Automated lifecycle marketing
Automatic campaigns by lifecycle stage or event: post-sale onboarding, anniversaries, usage milestones, re-engagement of inactives, win-back of churned customers.
Proactive NPS and CSAT
Automatic surveys at key moments: post-onboarding, post-purchase, post-ticket resolution. Sentiment analysis in responses. Early alerts.
Upgrade pattern detection
The Pyxis engine identifies customers with a natural expansion profile and delivers upsell or cross-sell recommendations with probability and reasoning.
Automated business reviews
Automatic business review deck generation per customer with metrics, milestones, opportunities and quarterly plan. The CSM reviews and adjusts, not builds from scratch.
// WHO IT IS FOR
Target audience.
- ›Companies with subscription or recurrence (SaaS, professional services with contracts, telcos, insurance, gyms, clubs)
- ›E-commerce with an active customer base where retention determines profitability
- ›B2B companies with formal account management and assigned Customer Success Managers
- ›Verticals with long-term relationship: clinics with patient portfolios, hotels with repeat customers, recurring consultancies, accounting firms
// FAQ
Frequently asked questions about Polaris
Does Polaris replace Gainsight, Catalyst or HubSpot Customer Hub? +
Polaris covers the same problem as those products, with three clear differences: accessible pricing for LATAM mid-market, Pyxis engine combining classical ML with LLM for prescriptive reasoning, and native integration with Omnira to execute retention actions without jumping between tools.
Do I need to connect Polaris with Omnira or does it work alone? +
It works perfectly standalone. When connected with Omnira, it executes Customer Success actions (NPS surveys, retention messages, business review calls) without going through external systems. Standalone it connects to the client's channels via their own APIs.
How does Polaris predict churn? +
The Pyxis engine trains a predictive model with the client's historical data (customers who churned vs. those who didn't) and then applies LLM-based prescriptive reasoning to understand the 'why' of each individual case and suggest the best specific retention action.