Agent use case / Sales
Start the day with researched accounts, not a blank search bar.
A scheduled ZeroTwo run enriches your prospect list from public sources, scores each account against your ICP and writes the result into HubSpot with a source and a timestamp on every field. Fields it cannot source stay empty, and the fields your team owns stay untouched.
The run, as a prompt
Runs every weekday at 5:00 AMEach weekday at 5:00 AM, take the contacts and companies created in HubSpot since the last run. Research each one from public sources only: the company site, press releases, job postings and public LinkedIn pages. For every field, record the source URL, a retrieved timestamp and a confidence of High, Medium or Low; if a field cannot be sourced, leave it empty and mark it Not found instead of guessing. Match existing HubSpot records by email or company domain first, and never overwrite a field on the protected list. Score each company against our ICP criteria with a one-line reason, then post the top ten to #prospecting with a suggested outreach angle.
Example configuration. Edit the criteria, protected fields and channel to match your own CRM.
What the word verified covers, and what it does not
Three different outputs get called verified in sales tooling. This workflow does the first one, depends on your licences for the second, and leaves the third to a dedicated service.
- What a ZeroTwo research run does
Public role discovery
Finds who holds a role at a company from its team page, press releases and public profiles, and records the page it found them on. This is sourced, not verified: people change jobs and pages go stale.
- Only from a source you already license
Licensed-source enrichment
Verified-database fields such as direct dials or modelled firmographics come from a data vendor's own licence. ZeroTwo does not replace that licence; use a vendor source only where your agreement with that vendor allows it.
- Needs a dedicated verification service
Deliverable email verification
Checking that a mailbox accepts mail is a separate job from research. An address found on a public page is a lead, not a verified address, so this page never labels one as verified.
One enriched record, field by field
This is the shape to ask for. Every value names its source type, when it was retrieved and how much to trust it, and a missing value is shown as missing.
| Field | Value | Source type | Retrieved | Confidence | HubSpot |
|---|---|---|---|---|---|
| Employee range | 201-500 | Company About page | Run timestamp | Medium | Enrichment field updated |
| Tech stack | Snowflake, dbt | Public job postings | Run timestamp | Medium | Enrichment field updated |
| Funding stage | Series B | Company press release | Run timestamp | High | Enrichment field updated |
| Decision-maker | VP Data (name from team page) | Company team page | Run timestamp | High | Contact created or matched |
| Work email | Not found | None | Run timestamp | n/a | Left empty, never guessed |
| ICP fit | B: strong on size and stack, unknown on budget | Computed from the fields above | Run timestamp | Medium | Score and reason written |
Illustrative record for a fictional company, shown to define the output format. It is not a measured result and the values are not real data.
The HubSpot diff, with protected fields intact
A run is only safe to schedule if you can see what it changed. This is the same illustrative contact before and after a run: what was added, what was updated and what was deliberately left alone.
| Field | Before the run | After the run | Outcome |
|---|---|---|---|
| Owner | Sam (AE) | Sam (AE) | Protected |
| Notes | Met at a conference; asked about security review | Unchanged | Protected |
| Job title | Head of Data | VP Data (previous value kept in the run note) | Updated |
| Employee range | Empty | 201-500 | Added |
| ICP fit and reason | Empty | B, with a one-line reason | Added |
| Enrichment source and timestamp | Empty | Source URLs and run time | Added |
| Work email | Empty | Empty (Not found) | Left empty |
Rules that make a rerun safe
Deduplication and write-back behaviour are decisions you put in the run, not things to hope for. These four cover most of what goes wrong with automated CRM enrichment.
Match before creating
Look up an existing contact by email, then by company domain, before creating anything. A record that already exists is enriched, not duplicated.
Write only to an allow-list
Name the fields ZeroTwo may update and the fields that are protected (owner, notes, custom fields your team maintains). Everything else is read-only.
No source, no value
A field without a source stays empty and is marked Not found. Missing data is reported, not filled with a plausible guess.
Safe to rerun
A second run matches the same records and rewrites only fields whose sourced value changed, so a failed or repeated run does not stack duplicates.
How ZeroTwo researches and enriches your leads
Runs every weekday at 5:00 AM. Output: Enriched records in HubSpot and a ranked shortlist in #prospecting on Slack. HubSpot and Slack are the example destinations.
Step 1
Pulls firmographics, tech stack and public contact details
Each account is enriched with employee range, revenue range, tech stack and the decision-makers named on public pages. Every field carries its source.
Step 2
Researches recent funding, hires and expansion signals
Funding announcements, senior hires and new-market launches are collected from press releases, job postings and news, with a link for each.
Step 3
Scores every prospect against your ICP
Each prospect is ranked against the criteria you define, with a one-line reason and a suggested outreach angle explaining why now.
Step 4
Writes enriched contacts to HubSpot with source tracking
Existing records are matched first. Enrichment fields are added or updated; protected fields are left alone.
Step 5
Posts a prioritized shortlist to the team
The top prospects land in #prospecting with the angle for each, so the day starts from a ranked list.
The manual version is a morning of copy and paste
An SDR opens LinkedIn, searches a title, opens a profile, pastes the name into a spreadsheet, then checks the company site, a funding database and the news. That is the work this run takes over.
What it does not remove is judgement: someone still decides who to contact, and the CRM keeps the history your team wrote by hand.
| Dimension | Manual | Scheduled ZeroTwo run |
|---|---|---|
| Research throughput | One lead at a time, by hand | A scheduled run works through the list while the team is offline. How far it gets depends on list size and how many sources each lead needs. |
| Data depth per lead | Name, title and company | Company context, recent news and an outreach angle, each with a source link |
| ICP scoring | A gut check | Scored against the criteria you write down, with a reason per lead |
| Data freshness | Stale within weeks unless someone re-enriches | Re-checked on each scheduled run, with a retrieved timestamp per field |
Where this workflow stops and the next one starts
This page owns batch enrichment and CRM write-back. Preparing an individual seller, and what happens after a lead replies, live on their own pages.
Frequently asked questions
Where does ZeroTwo get prospect data from?
From public sources: company websites, press releases, job postings, review sites, public LinkedIn pages and news. ZeroTwo reads what is publicly available, the way a research analyst does open-source research, and records the source for each field. It does not read anything behind a login. You remain responsible for checking that each source's terms allow your use.
What does verified mean on this page?
Nothing in a research run is called verified unless a dedicated step verified it. Public role discovery is sourced, licensed enrichment comes from a vendor you license, and deliverable email verification needs a dedicated verification service. Each field in the record shows its source type and confidence so you can see which is which.
How does ZeroTwo handle data that is already in my CRM?
The run checks for an existing record before creating one. If a contact exists, it enriches that record with new context such as role changes and company news, and it does not overwrite fields on your protected list. You decide which fields can be updated and which cannot, and you can ask for the previous value to be kept in a run note.
What happens when a field cannot be found?
Tell the run to leave it empty and mark it Not found. That is a better CRM record than a guessed value, and it makes gaps visible in the shortlist so a person can decide whether to look manually.
Can I define multiple ICPs for different segments?
Yes. Write a separate ICP definition for each product, market or segment, and run each as its own workflow with its own schedule and output destination.
How does this compare to ZoomInfo or Apollo?
Apollo and ZoomInfo are contact and firmographic databases that you license from them. A ZeroTwo run researches public sources for context specific to your product, such as funding, hiring and launches, and drafts an angle for each account. They answer different questions, and many teams use both. Check each vendor's current pages for what its plans include.
Can ZeroTwo also write outreach emails from the research?
Yes. The research gathered during enrichment can feed drafted emails that reference the prospect's company context, role and recent news. Run research first, review it, then draft. For the seller-side workflow of preparing, qualifying and writing, see the AI for sales prospecting page.
Write your ICP down, then let a run test it on your next signups
The Free plan starts with 100 credits, and Work agent access is listed from Plus at $14.99/month. Plan limits are on the pricing page.