Needle - The proactive GTM agent in Slack and Teams
by•
Most sales AI waits for you to ask. Needle is proactive. It works like a GTM engineer on your team: it watches your pipeline and acts before you do. Spots stalled deals and drafts the follow-up, preps you before calls, keeps your CRM tidy, surfaces real buying signals. It lives in Slack and Teams, wired into HubSpot, Gmail and Gong. Unlike horizontal agents, it is built for revenue teams, acts through your permissions, and your context and memory stay portable. No lock-in. Not another dashboard.
Replies
A lot of tools try to be proactive, but end up just adding another layer of noise.
The interesting part here is the judgment layer - deciding what actually deserves attention vs what can be ignored.
How do you approach that early on, before there’s enough data to really tune it?
Congrats on the launch!
Needle
@jared_salois Great question, this is exactly the right thing to push on.
Honest answer: we don't start from zero. Two things carry the cold start:
The signal already exists in your stack. Call transcripts, email threads, deal age, last activity, who went quiet after a proposal. Needle reads what's already in HubSpot/Gong/Gmail from day one, so the first stalled-deal list is useful before any tuning happens. Had a call today with a CRO sitting on 5000 old deals: the judgment there isn't ML magic, it's "no activity in X days + proposal sent + no reply" logic that's right often enough to earn trust.
So: rules + existing CRM signal first, personalization second. Being right 8/10 times with few pings beats 9/10 with spam, because after 3-4 ignored nudges reps tune you out forever.
Curious how it handles the noise though - does it learn from feedback when a flagged "stalled deal" is actually just going slow on purpose, or does it keep pinging you the same way every time?
Needle
@adem1368780 that is a good question, it has memory and learns over time from your feedback and behaviour.
AgentKey
Needle
@luki_notlowkey
Great question, and honestly the answer surprised us. We thought it was a tuning problem: find the perfect threshold, only ping when it matters. It's not. It's an ownership problem.
When we configure the agent for a team, even perfect signals feel like noise. When reps customize it themselves, decide what Needle watches, when it speaks up, how it drafts, the exact same notifications suddenly get accepted. People don't ignore an agent they built.
So our approach is less "we solved signal-to-noise for you" and more "you set the dials, Needle does the work." Adoption is a change management problem and it happens gradually, not on day one.
BlazeSQL
Great stuff! GTM is a pain, automating it is great
Needle
@justus_mulli Thanks Justus! Painfully true. The wild part is most of the "automation" isn't even fancy, it's just remembering to follow up. BlazeSQL 🙌
Fedica
Nice work Jan! Congrats on the launch
Needle
@albattran Thank you Samir, appreciate the support! 🙌
the fact that it pings you in slack before a call with a quick brief is genuinely useful, not just another ai wrapper hype thing
Needle
@neslihanamg43q Happy we have the same idea here :)
Supafax
Congrats! How are you connecting to all the 3rd party tools? @onur_o ?
Needle
Native API integrations: HubSpot, Gmail, Gong, Zoom transcripts. Needle lives in Slack/Teams so there's no new surface to learn.
The honest answer on why it works: most teams already have the signal flowing. Call transcripts land in the CRM, emails are logged. Needle just reads what's already there and acts through your existing permissions. No data migration, no IT project. Had a call today where the entire setup discussion was "so... we don't need anything new? Correct."
Each user connects their own setup, which also helps adoption. People trust an agent they wired up themselves.
Congrats on the launch @jan_heimes ! Needle feels strongest where sales teams usually lose momentum.
Curious how you recommend a team rolls this out in the first 2 weeks.
Needle
@barian_badhon
Thanks Shakhawat! Funny timing, I literally just came off a discovery call where the founder described his ideal rollout, and it matches what we see working:
Week 1: Connect the stack (HubSpot, Slack/Teams, call recording). Needle reads what's already there, no new tooling, no data migration. Then start with ONE shared channel where the whole team sees the daily briefing: which deals need updates, who to follow up with, what went stale.
Week 2: Let reps tune their own agent. What it watches, when it nudges, how it drafts. The more someone customizes it themselves, the more they accept it. Adoption is really a change management problem, and it's granular. It happens rep by rep, habit by habit, not team-wide on day one.
Tried Needle for connecting Notion + Gmail. Searching across both in one chat is surprisingly useful. The two-way actions (actually creating docs, not just reading) set it apart from other AI search tools. Some integrations still feel beta, but the workflow automation via chat is promising. Worth a try if you're tired of app-hopping.
The 'every edit is a signal' loop is the right instinct, but the thing that bit us building a similar draft-then-learn system was attribution. A rep editing a draft can mean the draft was wrong, or that the deal moved since it was generated, and those teach opposite lessons. Do you stamp each draft with the context snapshot it was built from, so an edit made because the situation changed doesn't get logged as a style correction and quietly skew the model toward the wrong voice?