Omelo in Claude: Business Development from Your AI Assistant
Most sales software makes you go to it. You open a tab, remember which screen has the filter you want, and click until you have a list. That made sense when software was the only thing that could read a database. It does not make sense now that you have an assistant that can do the reading.
Omelo ships an MCP connector, so business development happens wherever you already think. Ask Claude who to reach, and it searches your workspace, enriches the people, reads the map, and stages the outreach for you to approve.
The short answer
Add Omelo as a connector in Claude or any MCP client, mint a key in settings with a read-only or propose scope, and your agent can research, map and draft inside your workspace. Nothing sends without a person.
What MCP is, briefly
The Model Context Protocol is the open standard for giving an AI assistant real tools instead of a copy-paste window. A connector exposes a set of capabilities, the assistant calls them, and the results come back as data rather than as a screenshot you have to retype. It means the assistant can act inside your systems, under rules you set.
What your agent can do with Omelo
The connector exposes the same capabilities the app has, grouped the way the work is grouped.
| Area | What the agent can do | Scope needed |
|---|---|---|
| Research | Search the contact universe, enrich a person or a company, continue a search, read search history | Read |
| Market maps | Read maps and missions, follow a running mission, start or refresh one | Read, start needs propose |
| Campaigns | Read campaigns, stats, timelines, upcoming sends, replies, and campaign health | Read |
| Outreach drafts | Draft an email or a DM for a specific person, with the reasoning attached | Propose |
| CRM | Read lists and opportunities, create a list, add people to it, set a status, watch a company | Read, writes need propose |
| Presence | Read the queue, settings and stats, propose a post, ask for today's drafts | Read, proposals need propose |
| Deliverability | Inspect inboxes, placement, Postmaster data and deliverability audits | Read |
| Analytics | Workspace summary, campaign intelligence, the graph and its insights | Read |
Two scopes, and the line between them
Keys are minted in your settings, and each one carries a scope you choose.
- Read-only. Search, enrich, inspect. Nothing is created and nothing is changed. This is the right scope for research, for a teammate who should not touch live sending, and for anything you are still evaluating.
- Read and propose. Everything above, plus drafts, CRM lists, watchlists and starting a map mission. Every one of those is either a draft a person approves or an action that is trivially undone.
There is no scope that sends. Sending, publishing, paying and anything else with a lasting consequence stays inside Omelo behind a human. An agent can do all the work and stage all of it, and a person still says go.
How to connect it
- Open Settings in Omelo and go to the agent and API section.
- Pick a scope. The screen lists exactly which capabilities that scope allows before you mint anything.
- Create a key, copy it once, and add Omelo as a custom connector in Claude or your MCP client, with the key as a bearer token.
- Revoke it whenever you want. Keys are listed with their scope and last use.
What this is actually good for
Research in the middle of thinking. You are working through a market in a conversation with your assistant. Ask it who fits, and it comes back with named people and verified ways to reach them, without you changing tabs.
Briefs before a meeting. Ask for everything the workspace knows about the person you are about to meet: the company, what changed recently, what has been sent, what they replied.
A second pair of eyes on the pipeline. Ask which campaigns are underperforming and why, and let the agent read the deliverability numbers alongside the reply rates.
Drafting with the evidence attached. Ask for an approach to a specific person, and the draft arrives grounded in what the workspace knows, not in what the model guessed.
Why agent-first suits business development
Business development is a research job wearing a sending job's clothes. The hard part is deciding who is worth a message and what is true enough to say. That is exactly what an assistant with real tools is good at, and it is exactly what a database with filters is bad at.
It also means the product does not have to guess which screen you wanted. You ask for the outcome, and the agent assembles it. Start a project, or read how much you let it decide.
Frequently asked questions
Does Omelo work with Claude?
Yes. Omelo ships an MCP connector. Add it in Claude as a custom connector with a key minted in your Omelo settings, and Claude can search contacts, enrich people and companies, read market maps, campaigns and Presence, and draft outreach inside your workspace.
Does it work with other AI agents?
Yes. The connector speaks the Model Context Protocol, so any MCP-compatible client can use it with a bearer key. There is nothing Claude-specific about the endpoint.
Can an agent send messages on my behalf?
No. No scope allows sending. An agent can research, map, and draft, and everything it proposes waits inside Omelo for a person to approve. Sending, publishing and anything irreversible stays behind a human.
What are the scopes?
Read-only allows search, enrichment and inspection and changes nothing. Read and propose adds drafts, CRM lists, watchlists and starting a map mission, all of which are either approved by a person or trivially undone. You pick the scope when you mint the key and can revoke it any time.
Do I have to use an agent to use Omelo?
No. The app does everything on its own: search, filters, saved lists, CSV export, campaigns, approvals, schedules and analytics. The connector is there for people who would rather work from their assistant.
Tell us who you need to reach.
Describe your target in one sentence. Get your first verified contact list in under 24 hours.
Start a project