What’s New in Reply: AI Enrichment, Agentic Capabilities, and More

Svitlana Shuliak 02 Nov 2026

Good outbound depends on two things that rarely get enough time: knowing who you’re writing to, and keeping every tool in your stack in sync with your outreach. Research eats hours per list. Data lives in one place, automations in another, and your AI agents can’t reach the platform that actually sends the messages.

This month’s release goes after both. AI Enrichment turns prospect research into data you can filter and reuse, and Reply’s agentic capabilities let your tools and AI agents run outreach without you clicking through screens. A set of smaller updates for Jason AI SDR, sequences and reporting rounds out the month.

AI Enrichment: research your prospects with one prompt

Variables, advanced variables and AI variables already tailor every message in Reply. The slow part is the research behind them. Checking each company’s news, scanning LinkedIn posts and deciding whether a contact fits your ICP can take hours for a single list, and the findings usually end up scattered across tabs where you can’t filter or reuse them.

AI Enrichment does that research for you and saves the answers as fields on each contact.

One prompt, up to six insights per contact. Describe everything you want to know in a single prompt. Reply splits it into separate AI Custom Fields, each with its own research instructions and output format:

Yes/No for qualification, such as whether a company is hiring
Numbers for filtering, such as company size
Short phrases or sentences for personalization, such as a summary of a prospect’s recent LinkedIn post

Once the fields are filled, you can put them to work across Reply:

Target smarter. Build segments from real prospect data and qualify contacts against your own ICP criteria.
Personalize at scale. Use enriched fields as variables in email and LinkedIn sequences.
Give Jason more context. Select Contact Fields as a personalization source, and Jason AI SDR uses your enriched insights in the emails it writes.

Picture a team selling to SaaS companies. One prompt asks whether each account matches their ICP, how many employees it has and what the prospect posted about last month. Once the enrichment runs, the list can be filtered to “ICP fit: Yes, 50–500 employees,” and the first email opens with a line about that post.

How to get started:

Go to Settings → Contacts and create a prompt in AI Prompts, from scratch or from a template.
Review the suggested fields and map them to existing contact fields or create new ones.
On the People page, select contacts, click AI enrichment → Enrich with prompts, choose your prompt and click Run.
Filter, segment and personalize with the results.

AI Enrichment is in beta, so turn it on first in Settings → Beta features. The step-by-step guide covers the details.

Agentic capabilities: run Reply from your stack and your AI agents

Your CRM, prospect data, automations and AI agents probably live in different tools. Getting them to work with your outreach usually means exports, imports and manual steps in between. Reply now gives you several ways to make it the execution layer for that whole system, so your other tools can trigger, manage and report on outreach directly.

Agentic Chat. Ask in plain English and the assistant manages prospects, sequences, accounts, approvals and Inbox activity, analyzes performance and troubleshoots issues. It now supports separate conversation threads: keep up to 100, run up to 3 at once, and pick them up on mobile in the Reply app.
External MCPs inside Agentic Chat. Connect tools such as Clay, Apollo, Pipedrive, Attio, Close, JustCall and ClickUp, and the agent works across Reply and those tools in one conversation.
MCP. Let your own AI agents, for example in Claude or Cursor, access Reply and take actions as part of larger workflows.
API. Over 200 REST endpoints cover contacts, sequences, Inbox, accounts, enrichment, reporting, templates and prospecting.
Integrations and webhooks. Connect Reply to your CRM, add it to n8n, Make or Zapier workflows, and send Reply events to your systems in real time.

Here’s what that can look like for a RevOps team. A new lead lands in the CRM, an n8n workflow enriches and scores it, and qualified leads go straight into the right Reply sequence through the API. When the prospect replies, a webhook updates the deal stage. On Monday, the sales lead asks Agentic Chat for last week’s reply rates by sequence, in the same thread they used the week before.

Where to start:

Open Agentic Chat in Reply and ask for something you’d normally click through.
Connect the external tools you use from inside Agentic Chat.
For custom workflows, generate an API key in Settings → API keys and check the API docs.

Jason AI SDR: sharper first messages and a desktop home

Three smaller updates help Jason write more relevant outreach and help you respond faster.

Messages built on the signal behind each lead. When Jason finds a lead through an intent signal, it now uses that signal as the reason for reaching out. If a company was picked because it’s hiring SDRs, the first email can open with their growing sales team. It works automatically, with no playbook changes.

Your call on LinkedIn connection notes. Add one instruction to your playbook, such as “Don’t add a note to connection requests” or “Write a short note that references their recent post,” and Jason follows it in every LinkedIn Connection step. Keep the first touch light for one audience and add context for senior buyers in another.

Jason AI for Mac. Replies, AI drafts and approvals now have a desktop app. Conversations from all your teams arrive in one inbox, native notifications tell you the moment a prospect writes back, and unread counts sit in the menu bar. You can review AI SDR drafts and approve them in bulk. Download it on the Mac App Store (Apple Silicon, macOS 14 or later).

Outreach that follows you, and sounds like you

This month’s updates share one goal: fewer gaps between the outreach you plan and the outreach that actually happens. When email goes quiet, a call picks up right inside the sequence. LinkedIn activity moves at a pace that protects the accounts you rely on. Jason writes differently for each campaign, team, and rep while holding the line on what your brand will and won’t say. And when you step away, Agentic Chat keeps your pipeline one message away.

Reply handles the channels and the infrastructure. Jason brings the judgment of an SDR who keeps learning. Together they give modern sales teams outbound that feels personal at every step, and there’s more on the way.