Launched this week
Verly
AI agents that resolve customer support across channels
147 followers
AI agents that resolve customer support across channels
147 followers
Automate customer support with VerlyAI. Deploy intelligent AI agents for Web, Calls and WhatsApp in minutes. Reduce support costs and handle unlimited conversations simultaneously.
Hey Product Hunters 👋
I'm Raghvendra, founder of Verly.
I started building Verly after noticing something frustrating in almost every growing company: support teams weren’t struggling because they lacked effort. They were struggling because the system forced them to repeat the same conversations every day.
Hire more agents. Add more inbox tools. Patch things together.
But the repetition never really stops.
I kept asking myself:
What if support didn’t scale by hiring?
What if it scaled like infrastructure?
That idea became Verly.
Verly lets businesses train AI agents on their own data and deploy them across web chat, WhatsApp, and voice so repetitive questions get resolved automatically.
Not just replied to. Resolved.
And when AI hits a boundary, it escalates with full context so humans step in without starting from scratch.
We focused on three things while building:
• Real resolution, not canned answers
• Multi-channel from day one
• Human-in-the-loop escalation
We’re still early, still improving, and genuinely excited to learn from this community.
If you’re running support, I’d love to hear:
What’s the one support problem that drains your team the most?
Happy to answer everything openly.
Congrats on the launch.
Tried using it on a landing page, it gets the job done.
For me, the respond is a little more verbose. I would like to customize chatbot's respond styles.
@wei_yan4 Thanks a lot, Max and really appreciate you trying it out.
That’s helpful feedback on response style. You can already control the chatbot’s behavior through custom instructions inside the dashboard, including tone and response style. But we’re actively refining this to make verbosity and structure even more flexible.
Would love to hear how you’d ideally prefer responses to be structured.
Impressive mission! Since Verly trains on a business's own data, how do you handle data privacy and ensure that sensitive information doesn't leak into the AI's responses for other users? This is a huge concern for the B2B clients I work with.
Thank you @markeloff
Each Verly workspace is fully isolated. Your docs, help center, and conversations are only used to answer users for your business. Responses are generated via retrieval from your approved knowledge sources at runtime, so another company can never inherit or see that data.
We also added safeguards against prompt-injection and data leaks. The agent only receives small, relevant snippets and refuses requests for unrelated or sensitive information. If a conversation goes outside safe boundaries, it escalates to a human instead of answering.
So in short: your data stays private to your workspace, and it can’t leak into another customer’s responses.
if you are interested in seeing how verly can be helpful for your usecase, we can have a demo scheduled for you!
https://calendly.com/rdhakad2002/30min
Really love the framing around “Resolution over Responses.”
A lot of AI tools focus on generating smart-sounding replies, but grounding answers in a verified knowledge base is what actually builds trust. The fact that Verly checks structured information first instead of just “guessing” is a big differentiator.
Also, integrating with WhatsApp and Voice makes total sense — support today isn’t limited to a dashboard or web chat. It’s happening where users already are.
Curious to see how Verly handles edge cases where the knowledge base is incomplete — does it flag gaps and help improve the documentation over time? That feedback loop could make it even more powerful.
@rohan_gupta46 Thanks so much really appreciate that thoughtful take.
You’re absolutely right grounding responses in verified knowledge is key to building trust. That’s why we designed Verly to prioritize structured sources first instead of generating blind answers.
This is actually interesting — whom to contact if I want to integrate with my website. The chatbots response are really good.💥
@ketan_bajpai Thanks, Ketan! Appreciate that 🙌
You can directly sign up at verlyai.xyz and generate your chatbot in minutes. If you’d like help with integration, feel free to DM me or mail at team@verlyai.xyz happy to personally assist.
how accurate has it been in real usage so far ?
@divy_dwivedi Great question
In early usage, we’re seeing strong accuracy when trained on structured website/docs content. The key is good source data + controlled escalation when confidence drops.
hey, I just have a doubt about it.
can it assist my users inside the website or is it just for landing page?
@abby_pruh It works across your entire website, not just a landing page. Once deployed, it can assist users on any page where the widget is installed.
You can also integrate it with WhatsApp and voice if needed.
Happy to explain more if you’d like!