# Writer uses review policies to scale personal outreach

Markdown copy of https://realalpha.app/s/writer-uses-review-policies-to-scale-personal-outreach/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [May Habib](https://realalpha.app/builders/may-habib/) · May Habib is cofounder and CEO of Writer. · Co-founded and leads Writer, an enterprise agentic AI platform serving Accenture, Mars, Uber and hundreds more.
- Host: [Alex Konrad](https://realalpha.app/people/alex-konrad/) · Founder & editor, Upstarts Media · Ran Forbes' Midas List for ten years and co-created its Cloud 100 list; founded Upstarts in 2025.
- Show: Upstarts
- Episode: [Writer's May Habib: Building The Enterprise AI That Anthropic And OpenAI Sell — Then ‘Leave’ Behind](https://realalpha.app/v/wQGIp0KouWY/), Jun 25, 2026
- Moment: 18:11 to 21:27 of the episode (3 min), playing on https://realalpha.app/s/writer-uses-review-policies-to-scale-personal-outreach/; the episode from there: https://realalpha.app/v/wQGIp0KouWY/?t=1091

## Quote

> “I'm literally already **responding to the responses faster** than me even drafting a single email.”
> — [May Habib](https://realalpha.app/builders/may-habib/) at [20:12](https://realalpha.app/v/wQGIp0KouWY/?t=1212) · Upstarts

## Takeaways

1. The agent uses attendee lists, Gong and Slack context to draft follow-ups.
2. Habib reviewed 14–15 drafts and sent the strongest 11–12.
3. Policies and audit trails support sensitive one-to-one outreach.

## About this segment

May Habib describes a Slack agent that drafts tailored follow-ups using meeting and company context, then leaves them for her review. Policies and audit trails help make that speed usable in sensitive relationships.

From “Writer's May Habib: Building The Enterprise AI That Anthropic And OpenAI Sell — Then ‘Leave’ Behind”, Jun 25, 2026. [Watch the full episode](https://realalpha.app/v/wQGIp0KouWY/)

## Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[18:23](https://realalpha.app/v/wQGIp0KouWY/?t=1103) How do you feel about preserving care, connection, respect when people might be using that tool to work faster or to reply at volume? Because I think there's a key distinction between transactional emails or messages and more relationship-driven ones. So how do you balance it? Well, we have policies in Writer, right? And this is, you know, the feature set that you build when you actually talk to customers versus talk at customers.

[18:51](https://realalpha.app/v/wQGIp0KouWY/?t=1131) And those policies help an admin. It doesn't have to be an IT admin. Sales ops or sales enablement or marketing ops say, here are the ways in which I want my people using these connectors, right? And so you can say, you have to check an email before it sends out, before it gets out. You're not allowed to build automated email flows for one-to-one messaging, right? I'll give you an example from my workflow.

[19:15](https://realalpha.app/v/wQGIp0KouWY/?t=1155) So that same advisory board meeting that I told you about, I hope everybody is listening. And you will now know where this email came from. But I don't mind because I do think this is where the world is going. I asked Ryder, and this is just talking to Slack. I talk to my phone now just about more than I talk to anybody else, everybody else combined. And I said, writer, this is an agent, a writer agent inside of Slack.

[19:37](https://realalpha.app/v/wQGIp0KouWY/?t=1177) The meeting we had two days ago, the attendees are all in. I, you know, broadly named the channel. First go get the attendee names, then go to Gong, draft an email for all those 14, 15 people that were in the meeting, try to remark on something they brought to the conversation in my email, and then invite them to the Enterprise Brain beta. And, you know, we had both writer channels, which is basically our, like, enterprise open claw, the proactive agent, as well as enterprise brain we had gone through.

[20:05](https://realalpha.app/v/wQGIp0KouWY/?t=1205) And then we had had the FinOps conversation as well. So, you know, if they wanted to be in the FinOps conversation, just remark on that. And it was, you know. 60-second voice note, and then that was on my way somewhere. And then on my way back, got to look through all of the drafts. And the ones that were perfect, so I think like 11, 12 were perfect, send, send, send, send, send, send, right? I'm literally already responding to the responses faster than me even drafting a single email.

[20:31](https://realalpha.app/v/wQGIp0KouWY/?t=1231) Now, writer knows my voice, right? It knows our company's style. It knows writer is written in all caps. And because of the custom instructions and the policies that we've set, it knows what not to reference in the other calls, right? Like other people may have said in those calls. And so I have a lot of trust, right? Plus, I've got an audit trail to understand, like, if Ryder, and this wouldn't happen because I have the policy, but let's say I didn't have the policy on and it did send something to Alex that I didn't want to get sent, right?

[20:59](https://realalpha.app/v/wQGIp0KouWY/?t=1259) Do I have a audit trail of it sending, right, that IT can look at, that I can look at? And so these are all things, these are all features that are required to actually trust AI at scale in this kind of like high-end relationship management that we build agents for that customers love.