Customer Relationship Agents by Clarify - The M in CRM shouldn't be you
by•
Agents that handle the CRM work you've been doing manually — pipeline digests, lead enrichment, data hygiene, call coaching, and much more.
They run on a schedule or a signal, work across your stack, and fire without you touching them. Set them to run fully autonomously or require your approval before acting.
Describe what you want to automate or start from a template. Clarify builds it from there.
Replies
Clarify
Hey Product Hunt!
When we started Clarify, we promised you an autonomous CRM. Today, we make good on that promise with our take on Agents. We’re pumped to launch it today!
They trigger on a schedule or a signal, run across your stack, and work directly with your CRM data.
You describe what you want in plain English. Clarify builds it. You control what each agent can read, write, and do.
We also built a template library so you don't have to start from scratch — pipeline digests, lead enrichment, data hygiene, call coaching. Pick one, customize it, ship it.
This is the feature I've wanted to put in front of users since we started the company.
Curious to see what you build with it!
— Patrick
Love this! Having been using it ever since I heard about it. By far the best CRM in the market today. It's not the biggest. But its by far the smartest and the one that is not bloated with a bunch of graphs just because. It shows you the essential information and works flawless!
Clarify
@mr_pft Love to hear it!
I like the idea that the CRM work finally moves into the background.
A lot of the painful part is not managing relationships, but keeping the system clean enough so those relationships don’t get lost. Pipeline digests, lead enrichment, and data hygiene feel like exactly the right things for agents to handle.
Personally, I’d probably start with follow-ups and pipeline summaries first. Curious what teams automate most often after setup: cleanup, enrichment, follow-ups, or reporting?
Clarify
@andrasczeizel the "follow-up agent" is probably the most helpul and often the first thing to go when folks get busy. It was the first agent that I built, followed by my "inbox reply drafter" that drafts replies for me automatically with our business context.
The autonomous part is exciting, but the failure mode I'd worry about in a CRM specifically is a confidently-wrong write — merging the wrong duplicate, enriching a lead from a stale source, updating a field based on a misread signal. A bad autonomous write is worse than no write because it silently pollutes the source of truth and everyone downstream trusts it. Question for you Patrick: does Clarify ever say "I'm not sure" and hold instead of acting, or does every triggered agent always write something?
Clarify
@david_marko Totally fair concern. we worry about the same failure mode.
Clarify agents don’t have to write on every run. You can configure write-capable tools to require approval (or even deny them) per tool, per agent so if the agent isn’t confident, the safe pattern is that it pauses and asks before taking the risky action, instead of “silently shipping” a change.
The idea of a CRM that updates itself is appealing because manual data entry is everyone's least favorite part of sales. I'm curious, what's the biggest habit founders stop doing once they switch to Clarify?
Clarify
@harini_mukesh having to think about updating their CRM. We've focused so much of our time on making it simple to use and the agents do the heavy lifting. So when you need to ask questions via Rep/Claude, etc. you have a reliable source of all of your customer data.
Interesting
The interesting bit is not that the CRM can update fields. It is the authority model around each update: which signal triggered it, what data it touched, whether approval was required, and what receipt proves it happened.
How are you exposing that when an agent runs fully autonomously?
Clarify
@blah_mad We have an activity log of every change on a record, including the actor. Actions from agents are associated with an agent run, which includes the entire reasoning context.
Zaro
Cool product! would like to know the guardrails for agents not completing tasks, or we can control their usage.
Clarify
@william_ross4 You can control their usage, so if you want human approval or full autonomy you can configure it at the tool/agent level.
ChatWebby AI
The "set them fully autonomous or require approval before acting" toggle is the part that stands out to me — that's usually where automation tools force an all-or-nothing choice. Is the approval gate configurable per action type (e.g. let enrichment run free but always hold writes that overwrite existing CRM fields), or is it set at the agent level?
Clarify
@zain_sheikh you can do it per action/tool on the agent.
The 'relationship agents' framing is interesting — the hard part with any AI layer on a CRM is trusting the underlying data, since reps under-log and records go stale fast. Does the agent actively enrich/verify (pulling from email, calendar, web) to fill the gaps, or does it reason over whatever's already in the CRM? The first is far more useful but also where accuracy gets tricky.
Clarify
@mikebrandswarm Yes! We connect to email, calendar, and web to fill the gaps. You can build agents to manage both the manual data entry (closed won reasons, etc.) as well as create records like our deal creation agent that looks at call transcript, email, + usage data.