The problem this solves
The moments you most need a fast company read are precisely the moments you are away from a desk. A name comes up in a hallway conversation at an event. A prospect mentions their biggest customer on a call you took in the car. An intro email lands while you are boarding. The question is always the same - who are these people, really - and the answer is always in tools that assume you are sitting at a laptop with time to run a proper lookup.
So the check gets deferred, and deferred checks decay. By evening the name is one of nine things to look up, by tomorrow it is a half-remembered domain, and the conversational moment where the knowledge would have changed what you said is long gone. Mobile browsing a company's site is a poor substitute: marketing pages tell you what a company wants to seem, while the read you need is what they run, how big they are, and what has recently changed.
The phone-shaped answer is a bot that treats one message as a complete research request. Send a domain, get back the compressed truth in under a minute, short enough to read while walking. And because a quick read often becomes a real pursuit later, the full profile should file itself somewhere permanent without being asked, so desk-you inherits what phone-you learned.
How the mission runs
- One Telegram message starts the mission. You send "check acme.com" to your workspace bot from wherever you are. Telegram is the entire interface: no app switch, no login, no form. The agent treats the message as a research trigger with a hard speed requirement, because the answer is only useful while the moment lasts.
- The fast lookup runs across two lenses. Firmographics resolves who they are: size, model, market, footprint. Technographics reads what they actually run, which says more about a company than its homepage does. The agent favors sources it can hit quickly, keeping the reply inside the minute rather than chasing completeness.
- The compressed read arrives in chat. A handful of lines built for a phone screen: what they do, how big they are, the notable parts of their stack, and any recent move worth knowing. It is written to change what you say in the next five minutes, so anything that does not serve that gets cut from the reply.
- The full profile files to Notion unprompted. While you read the short version, the agent writes the long one - the complete profile with everything the quick reply compressed away - into your Notion workspace. When you are back at a desk, the deep version is already filed and linked, and the phone exchange lost nothing.
- Follow-ups continue in the same chat. Reply "who runs marketing there" or "compare them to the company from Tuesday" and the agent answers with memory of the whole conversation, including previous checks. The Telegram chat draws on the same workspace meter as everything else, so mobile research is metered and visible like any console mission.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A sub-minute company read delivered into Telegram: what they do, size, stack, and recent moves, compressed for a phone screen and a live conversation. Behind it, a full profile filed to Notion automatically, so every quick check leaves a permanent, searchable record. Ask five times during an event day and desk-you finds five complete profiles waiting.
Make it yours
- Prime it for an event: send the exhibitor or attendee companies you care about the night before, and get one compressed read per company before the doors open.
- Add a fit verdict: tell the agent your ICP once, and every check ends with one line on whether this company matches and why.
- Use it as a pre-reply check: before answering any cold intro email, send the sender's domain and decide from the read whether the reply is a meeting or a pass.
Frequently asked questions
How does the agent keep the reply under a minute?
The mission is scoped for speed: two fast data lenses, a live-moves check, and a compression pass, with slower deep research deliberately excluded from the reply path. The complete profile is assembled right after and lands in Notion, so speed on the phone never costs you the full picture.
Is the Telegram surface as capable as the console?
It is the same workspace, the same agents, and the same meter, reached through a different door. Thread memory works in the chat, follow-ups carry context, and the work is logged like any mission. The phone surface simply optimizes for terse requests and compressed replies.
What gets written to Notion, and can it clutter my workspace?
One page per checked company, in the location you designated when connecting Notion, with a consistent structure: overview, stack, signals, sources. Re-checking a company updates its existing page instead of duplicating it. The agent writes only to that location, so your broader workspace stays untouched.