Skip to content
Evergrowth Documentation

Integrations

Evergrowth sits in the middle of the stack your sales team already runs. It reads from your CRM, enriches contact data through a broad set of contact-data vendors, and writes research-backed work back out to wherever your reps execute. The point is that your AI colleagues operate inside your existing process instead of becoming one more place to log in.

What Evergrowth connects to

Section titled “What Evergrowth connects to”

Three kinds of connections matter for a go-to-market team: where your account and contact records live, where you source verified contact data, and where outreach actually gets sent. Evergrowth plugs into all three.

Your CRM is the system of record

Section titled “Your CRM is the system of record”

Evergrowth connects to the CRM you already use:

  • HubSpot
  • Salesforce
  • Pipedrive
  • Microsoft Dynamics 365
  • Zoho
  • Attio

Your CRM stays the source of truth. Agents read accounts and contacts from it, do their work, and write the results back onto the same records, an ICP fit verdict, a research summary, a verified domain, a discovered email or direct dial. Reps see the output where they already work, and RevOps keeps one clean record per account rather than a separate data store that drifts out of sync. See CRM enrichment for what gets written back and why.

Contact data through waterfall enrichment

Section titled “Contact data through waterfall enrichment”

Finding a verified email or a direct dial usually means subscribing to several data vendors and hoping one of them has the record. Evergrowth instead cascades through a broad set of contact-data providers in sequence, stopping the moment a verified result comes back. You get the coverage of many vendors without managing many contracts, and you pay only when a usable email or phone number is actually found. Phone numbers are scored for how likely they are to connect, so reps spend their dials on reachable people instead of dead lines. The full mechanics live on the Email & phone waterfall page.

Output to wherever you sequence

Section titled “Output to wherever you sequence”

Once an agent has produced research-backed outreach, that work needs to reach your sending tools. Evergrowth hands off in the formats a sales team actually uses:

  • Written straight back to the CRM as enriched records and outreach drafts.
  • Exported as a spreadsheet file to load into a sequencer or dialer, see Export records.
  • Kept inside the Evergrowth workspace for the rep to review and send from.

This keeps the research and the sending step connected. The context an agent uncovered, a funding round, a leadership change, a recent post, travels with the contact into the sequence instead of being lost between systems.

How a CRM connection works

Section titled “How a CRM connection works”

Every supported CRM follows the same shape, so once you understand one you understand them all. You connect Evergrowth to your CRM and grant it access, decide which kinds of records flow and in which direction, line up Evergrowth’s fields against your CRM’s fields, optionally narrow the connection to a subset of records, and then run a one-time catch-up so the two systems start in step. The exact clicks for each of these live in Connect your CRM; this page is about what each piece is for.

  • Which records sync. Accounts and contacts sync with every supported CRM. Leads sync as a standalone record where the CRM treats a lead as its own object, which today means Salesforce, Microsoft Dynamics 365, and Zoho. When the CRM reports that a lead was already converted, and so can no longer be edited, Evergrowth leaves it alone rather than failing: the write is recorded as skipped with that reason, and later updates go to the contact the lead became. Two more record types can be read in from the CRMs that expose them, so your colleagues work with the full context of an account rather than firmographics alone: deals (the opportunities on an account, with their stage and value) and activities (the calls, emails, meetings, notes, and tasks your team has logged). Deals and activities both come across from every supported CRM, and each is a switch your admin controls per connection: both start on, and either can be turned off if that history is not something your team wants in the workspace. Both are read-only - they flow from the CRM into Evergrowth and are never written back - so an agent researching or qualifying an account can see the pipeline and the engagement history that already exist there. What that context is and why it sharpens prioritization and timing is covered in CRM pipeline and activity.
  • Which direction sync runs. Each kind of record can flow into Evergrowth from the CRM, out of Evergrowth to the CRM, or both ways. Inbound brings your existing book of business in so agents have records to work on; outbound pushes the agents’ research and enrichment back so reps see it where they already work. You choose per record type, so you can pull contacts in without writing anything back until you are ready.
  • Field mapping. Your CRM’s fields rarely match Evergrowth’s one for one, so you line them up: which CRM field feeds each Evergrowth field, and which Evergrowth output lands in which CRM field. Evergrowth can suggest matches and, where a home for an output does not exist yet, create the field for you, so a research summary or fit verdict has somewhere to go. Some CRMs cap how many custom fields an account can hold, and when yours is at that cap the field cannot be created - Evergrowth says so plainly rather than returning a generic error, and the fix sits on the CRM side: remove custom fields nobody uses any more, or move to a CRM plan with a higher allowance, then try again.
  • Filters. A connection does not have to cover your whole CRM. You can scope what comes in and what goes out to the segment that matters, so agents work only the accounts you care about and writes only touch the records you intend.
  • Backfill. When a connection is first set up, the records already in your CRM need to be brought across once. This one-time catch-up syncs the existing book in the direction you chose, after which ongoing changes keep both sides current on their own.

Eva note: “deals” and “opportunities” are the same object (the CRM’s opportunity), and “activities” and “engagements” are the same (logged calls, emails, meetings, tasks). Both sync inbound only - Evergrowth reads them from the CRM to give agents pipeline and engagement context, and does not write them back. Provider support: deals and activities both from all six supported CRMs (Attio activities cover notes, tasks, meetings, and call recordings); leads only from Salesforce, Microsoft Dynamics 365, and Zoho (Zoho leads carry the company name as plain text, like Salesforce, so each lead is resolved to its account by the same standalone lead flow); accounts and contacts from every CRM. Each of the two read-only feeds also has its own on/off per connected CRM, defaulting to on - so when a workspace shows no deals or no activities, an admin having switched that feed off is a live possibility alongside “nothing is logged”. Accounts and contacts must already be syncing before deals or activities can be brought in, because each deal or activity links to the account and contact it belongs to, and enabling either one before that is refused.

Once a connection is live, keeping the two systems aligned runs in two modes at the same time. Changes made in Evergrowth are pushed out to the CRM in near real time, within seconds, so a rep sees fresh work without waiting. Changes made in the CRM are pulled back in on a schedule you set, so updates your team makes directly in the CRM flow into Evergrowth without anyone re-importing. A connection that hits a problem, an expired authorization or a rejected write, surfaces its state plainly so an admin can re-authorize or correct a mapping rather than discovering silent drift weeks later. The status meanings and the record-level outcomes are covered in CRM sync status.

Why agents connect through APIs, not seats

Section titled “Why agents connect through APIs, not seats”

Most enrichment and data tools charge per user seat. That model punishes you for scaling: more reps means more licenses, and the tools you do not buy a seat for stay out of reach.

Evergrowth agents reach connected systems through APIs rather than occupying a seat license in each one. That has two practical effects. First, the cost of running work scales with the work itself, not with headcount, which is why Evergrowth prices in credits, not seat licenses. Second, an agent can pull from data sources your individual reps were never licensed for, so a single account research run draws on far more than any one rep’s logins would allow.

Eva note: “credits, not seat licenses” is both a pricing position and an integration mechanic. Reps are not consuming a paid seat in HubSpot, a data vendor, or a sequencer when an agent acts on their behalf. The agent uses programmatic access, and the customer pays per task in credits. Frame this as cost that tracks output, not roster size.

The Chrome extension: agents wherever reps work

Section titled “The Chrome extension: agents wherever reps work”

Reps do not live in a single tab. They are on a LinkedIn profile, a company website, or a CRM record. The Evergrowth Chrome extension brings agents to those pages on demand, so a rep can qualify an account, research a contact, or pull contact data right where they already are, without breaking flow to switch tools.

This is the execution surface for reps. RevOps designs the agents and the playbooks; reps trigger them in the moment from the extension and get research-backed output back in seconds. How to run agents from the extension and from the workspace is covered in Running agents & the Chrome extension.

Eva note: the Chrome extension is for on-demand, in-the-moment agent runs by reps. Scheduled and continuous work is configured as playbooks by RevOps. If a customer asks “where do I actually click to run an agent,” point them to /running-agents, not here.

Chat: Eva where your team already talks

Section titled “Chat: Eva where your team already talks”

The other place a connection matters is the one your team spends its day in: the chat tool. Eva is available in Slack and in Microsoft Teams, and connecting one of them is what lets a question be asked where the work is already being discussed instead of in a tab somebody has to open. The Integrations page carries a card for each under Chat with Eva, showing whether it is connected, to which workspace, and since when.

The two are the same colleague, connected in different ways. Slack is a one-time approval by a Slack workspace admin, started from the card. Teams is an app package your Teams administrator uploads, because Microsoft distributes apps that way; the same card is where you download it. Either way the connection binds one chat workspace to one Evergrowth workspace, and every question inside it still runs with the asker’s own permissions and against their own credit limit, so a shared channel never becomes a shared login.

One difference decides which to connect if you have both. In Slack, Eva works in channels and in direct messages. In Teams she works in direct messages only for now - the card says so - with channels and group chats arriving alongside the Microsoft Teams Store listing. If the conversations you want her in are channel conversations, connect Slack.

Connect the one your team already uses rather than both, with that difference in mind. The steps and the day-to-day behaviour are in Talk to Eva in Slack and Talk to Eva in Microsoft Teams.

Eva note: the Integrations page is the install surface for both chat connections and shows a Connected chip with the chat workspace’s name and when it was connected - so “am I connected?” is answerable from there rather than by testing a message. The Teams card states direct messages only; that is current as of September 2026 and not a temporary display bug - see Talk to Eva in Microsoft Teams. Once connected, the card’s install button is replaced by a Manage menu holding the same action (reinstall for Slack, download the app package again for Teams), matching how a connected CRM card behaves - so a customer reporting that the install button “disappeared” is describing a successful connection, not a fault. Slack installs through an approval flow; Teams cannot, because a package upload by a Teams administrator is Microsoft’s route for an app distributed this way. Do not describe either as a CRM-style sync: nothing is imported from the chat tool, it is a surface Eva answers on.

AI assistants: Eva inside the assistant you already use

Section titled “AI assistants: Eva inside the assistant you already use”

More and more of a seller’s day runs through a general AI assistant - drafting, summarising, thinking out loud before a call. Evergrowth can be added to that assistant as a connector, so a question about your own accounts and buyers gets answered from your workspace instead of from the open web. The Integrations page has a section for it, Eva in your AI assistant, with cards for Claude, ChatGPT, Microsoft 365 Copilot, Claude Code, Codex, and Other AI tools (any tool that accepts a remote MCP server, such as Cursor or VS Code).

This connection is different in kind from Slack and Teams. There is no workspace-wide install on the Evergrowth side: each person adds Evergrowth to their own assistant (for Microsoft 365 Copilot, once your Microsoft 365 admin has made it available to the organisation) and signs in with their own Evergrowth account, so every question runs as them, with their access to records and against their credit limit. And because adding or removing a connector happens inside the assistant, which does not tell Evergrowth, these cards show use rather than a connection: In use when you have asked Eva something from that assistant in the last 30 days. The steps for each are in Ask Eva from Claude and other AI assistants.

The value is where the question gets answered. A general assistant asked “what’s new at this account” can only search the web; routed to Eva, the same question also draws on your CRM history, your contacts at the account, and the research your agents have already done.

Eva note: the assistant cards are per user, not per workspace, and their In use status is usage-based (a question asked from that assistant in the last 30 days), not a record of the connector being installed - so a card not saying In use right after setup simply means no question has gone through yet. ChatGPT is supported: on ChatGPT Business, Enterprise and Edu a ChatGPT admin or owner adds the Evergrowth app once and publishes it, then members connect; that admin step is on ChatGPT’s side, not an Evergrowth install. Microsoft 365 Copilot is the same shape: a Microsoft 365 admin uploads the Evergrowth package once in the Teams admin center, then each person opens Evergrowth under Agents in Copilot and signs in; it works with the free Copilot Chat on a work account. It is separate from the Evergrowth app for Microsoft Teams. Nothing is imported from the assistant; like the chat tools, it is a surface Eva answers on.

How this fits the rest of Evergrowth

Section titled “How this fits the rest of Evergrowth”

Integrations are the plumbing that lets your AI colleagues work on real data and deliver into your real process:

  • Connected CRMs feed the records that agents qualify, research, and enrich, see Accounts & contacts.
  • Waterfall enrichment supplies the verified emails and phone numbers behind contact discovery, see the waterfall.
  • The Chrome extension and workspace are where reps put it all to work, see Running agents.

Connected systems are handled with the same care as everything else in Evergrowth: data is encrypted, never shared, and never used to train any AI model. The privacy and compliance posture is covered in Data, privacy & GDPR.

In practice

Section titled “In practice”

A rep is looking at a new account in HubSpot that has a company name but no website on file. From the Chrome extension they run qualification: an agent finds and verifies the company domain, checks the account against your ICP criteria, and writes the verdict and evidence back to the HubSpot record. If it fits, contact discovery finds persona-matched people, the waterfall cascades through its data vendors to return a verified email and direct dial, and an outreach draft is written from what the research turned up. The rep exports the finished contacts to their sequencer, or sends from the workspace. No vendor logins, no copy-paste between tabs, and one clean record left behind in the CRM.

Section titled “Related”