
Senior Solutions Engineer @hops_ai & @zencoderai | AI Agents
India
Joined October 2014
- Tweets2.6K
- Following1.3K
- Followers1.5K
- Likes13.6K
Feature requests reach support every day and mostly stay in closed tickets. A weekly job that copies each request into a shared table, links the ticket, and groups duplicates gives product something to sort and verify. I'd send that table before any summary deck.
30
Answering one support ticket can mean the helpdesk for history, the CRM for the plan, Slack for the engineering thread, and the docs to check the fix. I'd want an assistant that reads all four from the ticket and shows where each fact came from, so I can check it before replying.
2
1
36
Design feedback often lands as a Slack screenshot with a red circle and "can this be bigger?" A week later it's hard to tell which comments were handled. Pinning the comment on the image itself, with the revised version posted beside it, keeps the request and the change together.
32
Before a customer call I want a one-page brief: open tickets, last call notes, plan, recent usage. An agent can pull that from the CRM and helpdesk, as long as every line shows its source and date, and anything it couldn't find is marked missing instead of guessed.
1
47
Per-seat pricing for AI tools decides who is in the loop. When a team buys seats for a few people, everyone else gets the output secondhand, pasted into Slack. Pricing the AI's work instead of each user (Hops does this) lets the whole team use one agent in the same doc.
28
A decision made in a Slack thread often gets recorded nowhere else. Two weeks later someone asks where it was agreed, and the thread is buried. An agent in that thread could write the decision into the tracker before the conversation ends, with a link back to the messages.
1
31
For one day, count how often you copy text out of a tool, paste it into an AI chat, and paste the answer back somewhere else. Each round trip leaves the sources and the original question behind, so the next person sees only the answer.
23
When an AI draft starts in one person's chat, it reaches the team as an attachment: v2.docx in a DM, while a colleague has already edited v1. The cleanup after that is version reconciliation. Drafting in the shared doc, with the source files linked, skips that step.
1
23
Release notes usually get written by whoever merged last, and support reads them the morning customers start asking. I'd draft them in a doc where engineering, docs and support edit together, and give support one section: what customers will ask about this change.
27









