We’re launching version 2.0 of the Grovity Portal, with a set of improvements aimed at one thing: letting you operate your AI agents better, so you can scale your business without your operation growing at the same rate.
Here’s what’s new.
One Grovy: run everything from a single chat
What you get: you operate your whole agent from one chat interface. You don’t learn the portal, you talk to it.
We unified all the Grovys into a single Grovy. From that one conversation you can:
- Get insights. Ask what’s happening in your operation and get the answer from your real data.
- Adjust your agent’s configuration. Change the prompt and how your agent behaves, without leaving the chat.
- Create and edit tools. Give your agent a new capability, or fix an existing one, right where you noticed it was needed.
Before, seeing a problem and fixing it were two different screens. Now they’re the same conversation.
Talk to Grovy out loud
What you get: you operate your agent while you’re doing something else, at the speed you talk instead of the speed you type.
We added voice to the Grovy chat. You can speak to Grovy instead of typing, and everything else works the same: the same insights, the same configuration changes, the same tools. Describing what you want out loud is faster and more natural than writing it, especially when you’re explaining how your agent should behave.
Proactivity: your agent reaches out first
What you get: you stop depending on the customer coming back. Your agent picks the conversation up again on its own, whether that’s tomorrow or three months from now.
This is the biggest upgrade in 2.0, and it comes in three parts.
Beyond the 24-hour window on WhatsApp
WhatsApp closes the free-form messaging window 24 hours after the customer’s last message. Until now that was a hard ceiling on proactivity: if the customer went quiet, the conversation was over.
Now Grovity can reach out past that window using Meta-approved templates, so your agent can come back days, weeks, or months after the first interaction.
On the commercial side, a lead who asked about a price and never replied isn’t lost, it’s pending. Abandoned carts, quotes that never closed, customers who haven’t bought in months: your agent re-engages all of it, and because it’s the same agent, whoever replies lands in a conversation that can quote, recommend, and close on the spot.
On the support side, the follow-up that used to depend on someone remembering now happens on its own. Confirming an order arrived, checking whether an issue actually got resolved, reminding about a renewal or a pending document. All of it without an agent-hour spent on it.
Scheduled proactivity
Your agent can also schedule a message for a specific future date, from what it picks up in the conversation itself.
If a customer mentions “I have the game on March 13,” Grovity schedules the outreach for that day. It doesn’t matter whether that’s in one day or in twenty: the agent comes back at the moment it’s actually useful, with the context of what was discussed. No CRM reminder, no one writing it down.
Campaigns (beta)
You can now schedule campaigns from Grovity: mass sends to your audiences, from the same place you operate your agent.
The difference from a normal blast is what happens after it’s sent. The conversation a campaign opens doesn’t dead-end in an inbox, it lands on your agent, which answers, quotes, and follows up at whatever volume replies.
Campaigns are in beta. If you want in, book a demo.
Mercado Pago in agentic payments
What you get: you charge with the payment method your customers already use, inside the conversation.
We added Mercado Pago as a participant in agentic payments. If you already collect through Mercado Pago, or you want to implement payments with them, your agent closes the transaction in the chat, with no handoff to a separate checkout. It’s one more rail available in agentic payments, alongside the ones we already support.
Performance: know what your agents handled
What you get: you stop assuming your agent is working and start knowing what it handled.
We built a new Performance section where you monitor sessions, users, and messages in one view. How many conversations you’re handling, how many people you’re reaching, and how the volume moves day to day and week to week, without reconstructing it by hand.
Evals: governance over what your agents do
What you get: you improve your agent fast without silently breaking what already worked.
We brought Evals into the portal. An eval is your agent’s quality contract: it defines what has to be true in a given situation, so every change gets validated with evidence instead of hope.
They run in two places. Before production, against synthetic users that simulate realistic conversations, so you know in minutes whether a change works. In production, over a sample of real conversations, to confirm your agent still holds up.
We covered this one in depth in its own announcement: Introducing Evals: Certainty Instead of Hope on Every Change.
See it in the product
A walkthrough of Portal 2.0:
Try it
Portal 2.0 is live for every Grovity customer. If you want to see it on your own operation, book a demo.