Enabled by default on every new Doppler account, reaching 4,801 accounts since launch.
DOPPLER - SOFTWARE & TECHNOLOGY
AI-powered insights that
AI-powered insights that
launch campaigns in minutes
We designed and built ECO IA, a conversational AI assistant embedded in Doppler's marketing platform that connects live account data with natural-language interaction and automated campaign creation.
The result: an AI layer live for Doppler users at no additional cost, enabled by default on every new account, operating within clear data boundaries and giving Doppler's team real-time visibility into adoption, costs, and performance.
<3 MONTHS
FROM KICKOFF TO LIVE
95%
RETURNING USERS
97%
PROMPT CACHE HIT RATE
About Doppler
Doppler is one of Latin America's most established digital marketing platforms, with over 20 years in the market. It helps thousands of businesses manage customer communications and automate campaigns across email, SMS, WhatsApp, push notifications, social media, and chatbots.
Backed by deep regional expertise and a broad customer base spanning small businesses to large e-commerce operations, Doppler continues to expand its product capabilities to meet growing demand for smarter, more automated marketing tools.
The starting point
Doppler's platform already gave users access to meaningful account and campaign data. They could review performance, identify trends, and spot areas that needed attention. The gap was in what came next: interpreting those signals, deciding what to prioritize, and building the campaign assets required to act on them. For small businesses and e-commerce owners with limited time and resources, it was much harder to sustain. The data was there; the path from insight to action was not. Doppler needed to:
-
Help users understand what was happening in their accounts faster.
-
Reduce the manual effort required to go from data to campaign draft.
-
Make performance analysis accessible to users without dedicated marketing teams.
-
Integrate AI capabilities within existing product workflows, not alongside them.
-
Build infrastructure to monitor usage, costs, and system performance at scale.
THE GOAL
Closing the gap between insight and action
Our goal was to design and deploy an AI assistant that could read live account data, identify opportunities, and help users generate editable campaign drafts directly inside Doppler's platform, without adding friction to workflows they already knew.
What we did
We designed and built ECO IA, a conversational assistant embedded directly in its platform and connected to live account data. Users can request a broad diagnostic across seven dimensions of account health, or explore specific areas, with the assistant surfacing opportunities across contacts, campaigns, and performance metrics.
When a campaign is requested, ECO IA generates branded HTML based on the company's visual identity, suggests a subject line and contact list, and saves the draft inside Doppler's existing visual editor. It also builds complete welcome automation flows, generating each email in the sequence. Everything stays as an editable draft, so the user decides what goes out. The engagement also included a custom monitoring dashboard and support for AI adoption practices, including automated testing and workflow adjustments.
Key initiatives included:
- Conversational AI interface connected to live Doppler account data, with automatic Spanish or English response based on account language.
- Account diagnostic across seven dimensions: performance, sending frequency, segmentation, automation, database health, technical configuration, and multichannel maturity.
- Automated campaign generation with branded HTML, subject line suggestions, and list recommendations.
- Multi-email welcome automation flows built from a single conversation.
- Editable draft creation within Doppler's existing visual editor.
- Prompt caching implementation to keep inference costs predictable at scale.
- Custom monitoring dashboard with real-time visibility into usage, costs, and error classification.
BEHIND THE SCENES
The tech stack
To build ECO IA, we covered the full delivery scope: AI architecture, product design, backend integration, and enablement. Every layer was designed to operate within Doppler's existing infrastructure, connecting the assistant to live account data and campaign workflows without disrupting the product experience customers already knew.
The core of the assistant runs on Claude Sonnet 4.5, Anthropic's AI model, selected for its ability to handle complex, context-dependent queries in a customer-facing product. Doppler consumes Claude on its own Anthropic account, with Making Sense responsible for architecture and integration. To keep inference costs predictable as usage grows, we implemented prompt caching over the account data retrieved through Doppler's API, sustaining a 97% prompt cache hit rate.
Results
Live in production in under three months.
ECO IA launched inside Doppler's platform in under three months from kickoff, including full integration with live account data, campaign creation workflows, and real-time monitoring. It is in active use by Doppler users today, included at no additional cost and enabled by default on every new account.
95% of users who engaged with ECO IA came back, returning across multiple months.
Complete multi-email welcome automation flows built from a single conversation, generating each email in the sequence.
97% prompt cache hit rate, keeping inference costs predictable at USD 0.066 per interaction.
Account diagnostics, performance metrics, and campaign creation are the three most frequently requested actions.
Real-time monitoring infrastructure tracks engagement, token costs, latency, error classification, and system health continuously.
AI-ready development team at Doppler, with improved practices for testing, validation, and Claude integration.
Looking forward to making
sense with you
Tell us about your project challenges.