This week: R$ 47,890 across 142 orders, up 12% vs last week.
Via Kondado MCP
You
What's the ROAS of Meta Ads this week?
Meta Ads
Meta Ads ROAS this week: 3.42x on R$ 12,400 spent.
Via Kondado MCP
You
List the top 5 products in Shopify this month
Shopify
Top SKUs in Shopify: SKU-021, SKU-188, SKU-054, SKU-301, SKU-077.
Via Kondado MCP
You
How is GA4 conversion trending?
Google Analytics 4
GA4 conversion last 30d: 2.84% (+0.42 pp MoM).
Via Kondado MCP
Works in Claude, ChatGPT and any MCP client
Plug-and-play
Connect once on Kondado, query forever in chat.
Read-only by design
The AI reads and answers, never edits your systems.
Two paths, same data
Chat for ad-hoc Qs, ready dashboards for recurring KPIs.
AI with your data on Kondado
Connect your data source once on Kondado. From that point on, any MCP-compatible AI client — Claude, ChatGPT and others — can query it in chat, in natural language. Ask about sales, leads, campaigns or finance and the assistant answers with up-to-date numbers.
The connection happens via MCP (Model Context Protocol), an open standard that bridges the AI assistant and Kondado data. In Claude (Web or Desktop), ChatGPT and any other MCP client, you authorize the Kondado server once via OAuth.
For visual, recurring monitoring, Kondado also offers free dashboard templates on Power BI and Looker Studio. Open the template, connect your source and the dashboard is ready in minutes. Use templates for continuous KPI tracking; use chat with AI for ad-hoc questions or analyses. Both paths run on the same data connected to Kondado.
Data for AI agents
Kondado is the data infrastructure layer that feeds enterprise AI agents. Instead of each agent building its own integrations with Bling, Meta Ads, GA4 or databases, it consults your data through a single MCP server — with OAuth, access control and automatic updates. The same connector base serves agents inside Claude, ChatGPT and any other MCP client.
Analytics agents
Agents that consolidate KPIs, build trend reports and answer cross-source data questions on demand.
▸ Build a sales report from Bling and reconcile it with GA4 conversions
▸ Compare CAC across Meta Ads and Google Ads, segmented by channel
Operations agents
Agents that monitor inventory, order status, receivables and operational alerts across multiple systems.
▸ Daily check on Bling + Shopify inventory; alert on critical stock
▸ Review Omie receivables due this week and prioritize collection
Marketing agents
Agents that consolidate campaign performance, follow ROAS and suggest budget shifts.
▸ Compare Meta Ads, Google Ads and Pinterest ROAS over the last 30 days
▸ Identify Google Ads campaigns with low conversion this week
How it works: from source to AI answer
1
Data source
Your systems: ERP, CRM, marketing platforms, databases.
2
Kondado
Universal connector that normalizes and serves the data.
3
AI assistant
Claude, ChatGPT or any MCP-compatible client.
4
Natural answer
You ask in chat, the AI answers with up-to-date data.
What you can ask the AI
With Kondado connecting your data to Claude, ChatGPT or other MCP clients, you turn natural-language questions into instant analyses:
E-commerce and operations
"Which products had the biggest stock drop this week?"
"Compare this month's average ticket with last month"
"What is the status of pending orders?"
Marketing and growth
"How much did we spend on ads in the last 30 days?"
"Which campaign had the best ROAS?"
"How is the sales funnel conversion rate?"
Finance and BI
"Show me the consolidated P&L for the last 3 months"
"Which receivables are near their due date?"
"What is the CAC by acquisition channel?"
AI without vs with your data
AI without your data
✗
Answers Generic, no business context
✗
Source Training data + public web
✗
Updates Knowledge frozen to training cutoff
✗
Setup Build/maintain custom integrations
✗
Security Each API with its own auth
AI with Kondado data
✓
Answers Specific, based on your real numbers
✓
Source Your ERPs, CRMs, ad platforms
✓
Updates Live data, queried on demand
✓
Setup Minutes via no-code
✓
Security Centralized and auditable
Getting started in 3 steps
1
Create your Kondado account and connect a source
Visit app.kondado.com.br/signup, pick the source (Bling, Omie, Meta Ads, etc.) and authorize the connection. Data becomes available for AI usage automatically.
2
Configure your AI client to access Kondado
In Claude (Web or Desktop), ChatGPT or another MCP client, simply add the Kondado server in the connections and authorize via OAuth. See the detailed guide per client.
3
Ask in chat
Use natural language — "compare my revenue for the last 3 months", "which Google Ads campaigns converted most this week?". The assistant queries the data live and answers with real numbers.
Frequently asked questions
Can I ask in natural language about my data?▼
Yes. Once you set up Claude, ChatGPT or another MCP client to access Kondado, you ask in chat about sales, inventory, campaigns, leads, finance — anything that lives in your connectors. The assistant answers based on the up-to-date data.
Will the AI write or modify data in my systems?▼
No. Access is read-only and analytical. The AI does not create orders, send messages or issue invoices.
Which AI clients are compatible?▼
Claude (Web and Desktop), ChatGPT and any other client that supports MCP (Model Context Protocol). In all of them, you add the Kondado server in the connections and authorize via OAuth a single time.
Kondado plans start at $19/month and include access to data via Claude, ChatGPT or another MCP client, plus ready dashboards. Free 14-day trial, no card required.
Can I use ready dashboards and AI chat at the same time?▼
Yes. The same data connected on Kondado serves both paths: dashboard for visual monitoring, chat for ad-hoc questions.
Compatible LLM clients
MCP is an open standard, so any LLM that supports it works. Detailed setup guides for the two most popular below:
Via Kondado MCP
