Source and qualify leads

The prompt

The real prompt behind the steps above — detailed enough that an agent with a people-search tool, email/phone enrichment, and CRM access can run it as written.

# Source, qualify, and push contacts to Salesforce

You have access to a LinkedIn company/people search tool, email and
phone enrichment tools, and read/write access to Salesforce. Run this
for **[Company Name]**.

## 0. Set up (skip if you're already connected)

If you don't already have live tool access for this, connect it first
— this prompt is only as real as the tools behind it:

1. Create a free account at app.tulina.ai.
2. Add Tulina's MCP server to your assistant: https://mcp.tulina.ai/mcp
3. In Tulina, connect your own API keys for the tools this prompt
   uses — LinkedIn search, email/phone enrichment, and Salesforce.

Once connected, your assistant has real tool access and the rest of
this prompt runs as written.

## 0b. Already on Tulina? Three things around the run

An account and the MCP server are not the whole setup. Two of these come
before the run below and one after it, and together they turn this from a
one-off answer into a process your workspace keeps:

1. **Activate this process's connectors first.** Call
   `oto_connector(op="list")`, match it against the tools named in step 3
   above, and `oto_connector(op="select", name="...")` every one that
   isn't active yet. Selecting a connector does not mount its tools in the
   conversation you're already in — reach them through `oto_call` for this
   run, or open a fresh conversation once they're all on.
2. **Attach the work to an existing project.** `oto_project(op="list")`
   shows your active org's projects: pick the one this work belongs to
   rather than opening another, and keep its id. The project is where this
   process, the tables it writes to and the connectors it uses hang
   together.
3. **When the run is done, save it as a process — with its graph.** Write
   the body with `oto_procedure(op="set", ...)`, then attach it with
   `oto_project(op="link", project_id=..., target_type="procedure",
   target_ref="<your slug>")`. The body has to carry a drawing: read
   `oto_guide(op="read", slug="procedure-flowchart")` and follow it
   exactly — ONE untagged fenced block in that grammar, opening with the
   trigger and a quoted example of what you'd type to start a run. Tulina
   parses that drawing back into the graph it renders as the process's
   default view, which is what makes it come out in the same style as
   every other process in the app; a drawing the grammar can't read falls
   back to raw characters instead. Saving a process needs org-admin
   rights — without them, hand the finished body to someone who has them.

## 1. Resolve the company and pull its people

- Resolve [Company Name] to its exact corporate entity — not the
  nearest name match. If you know the domain, filter on that; avoid a
  generic "website" filter if your tool has one, it tends to return
  the whole database instead of the one company.
- Discard any candidate with zero indexed employees, or that looks
  like a subsidiary/showcase page rather than the real operating
  company. If the best headquarters-level entity returns nobody, check
  its sub-organizations before giving up.
- Search people at that entity filtered to the seniority bands you
  sell into (e.g. C-suite/VP/Director/Head for economic buyers and
  decision-makers, Manager/Senior for champions). Fetch **every page**
  of results — the good candidates are often past the first screen.
- If your tool's title filter is unreliable (some silently ignore it),
  don't trust it blind: pull the broader seniority-filtered set and
  filter titles yourself against a keyword list for the functions you
  sell into.
- Never filter people by domain for a multi-location business — every
  branch or store shares the parent's domain and floods the results
  with local staff. Filter by the corporate entity's own ID instead.

## 2. Score against your ICP

- Your ICP: **[fill in — target function, seniority band, company
  profile]**.
- If you sell to a corporate/HQ buying center, exclude anyone at
  store, branch, or franchise level, even if their title looks senior.
- Sanity-check tenure: drop anyone whose headline shows a different
  current employer than the target company.
- Aim for a realistic split across buyer tiers (e.g. 1-2 economic
  buyers, 3-4 decision-makers, 4-5 champions) rather than maximizing
  headcount.
- Record a one-line reason per person for why they fit — you'll want
  it on the record later, and so will the rep who opens it.
- If this is running unattended, with no one to review the shortlist,
  be conservative: anything borderline gets dropped, not kept.

## 3. Resolve email

- Submit the **entire qualified batch in one call** to your primary
  email-enrichment tool — most batch 100-250 contacts per call, check
  yours. Never loop this one contact at a time; it costs an order of
  magnitude more calls for the same result.
- For anyone the first pass couldn't resolve, run a second batched
  pass through a fallback provider.

## 4. Resolve phone

- Same discipline as email: submit the whole batch to your primary
  phone-enrichment tool in bulk.
- Identify each contact by **LinkedIn URL + name + company** — never
  by email alone. A bare-email lookup has been known to return a
  completely unrelated person at a different company, even though it
  billed as a match.
- Fall back to a second, single-lookup provider only for the residual
  contacts the bulk pass couldn't resolve.

## 5. Deduplicate against Salesforce

- Before creating anything, query Salesforce **once** for every
  contact in the batch by LinkedIn URL — a single
  `WHERE LinkedIn_URL IN (...)`-style query covers the whole batch,
  never one query per contact.
- Split the batch into contacts that already exist (capture their Id)
  and contacts that are genuinely new.
- If Salesforce's own fuzzy duplicate rule still blocks a create (this
  happens when an existing record has no LinkedIn URL on file), look
  that record up by name + account, back-fill its LinkedIn URL, and
  treat it as "already exists" instead of fighting the duplicate rule.

## 6. Create or update, in bulk

- Bulk-create every genuinely new contact in one call (chunk at ~200
  if your tool caps batch size). Write only: first/last name, title,
  email, **mobile phone** (never the office/landline field — it's a
  different field and this pipeline never has a real value for it),
  the account, and the LinkedIn URL itself, since it's your dedup key
  for next time.
- Flag newly created records only (e.g. a status or source field set
  to something like "Newly found") — never touch that flag on a
  contact that already existed.
- Bulk-update the contacts that already existed with whatever new
  email/phone you resolved, in one batched call, not one per contact.
- Never write outside this explicit field list — no internal
  scoring/tier fields, no guessed mappings.

## Output

Report: how many contacts found, how many net-new vs. updated, how
many got email, how many got phone, and anyone dropped along the way
and why.