| Tool | What it does | When ABM marketers use it |
|---|
| Search 1st Party Account Data | Filters your Demandbase accounts by journey stage, engagement score, pipeline predict score, open opportunities, and more | Identifying which target accounts are active enough to prioritize for a send play |
| Search 1st Party Person Data | Finds engaged contacts at target accounts by name, title, job function, engagement points, and lead status | Identifying who to send to at a priority account |
| Search 3rd Party Company Data | Returns firmographics, technographics, and recent news on any B2B company | Researching a target account before building a play |
| Search 3rd Party Contact Data | Finds contacts by title, job function, seniority, and location across Demandbase's global directory | Expanding contact coverage at an account beyond your current CRM data |
| Get Reference Data | Returns tenant-configured account lists and their IDs | Pulling a named account list to use as the input for a send play |
Setup: An account has been on your Tier 1 ABM list for two quarters. Multiple SDR touches have gone unanswered. Digital engagement is low. The account fits your ICP and you know it's worth the investment.
How the MCP helps: The AI uses Demandbase's 3rd Party Company Data Global Directory to pull firmographics, recent news, and installed technologies for the account. It searches the Contact Global Directory for contacts matching your target job function who aren't yet in your CRM. It pulls available Salesforce notes from prior outreach. Based on what it finds, it searches the Sendoso marketplace for a gift within your approved budget and drafts a message tied to the account's context. You get a complete proposal before anything ships.
Pull your context first:
Run this before filling in the main prompt. It pulls what can be auto-populated so you only need to supply goal, gift type, and budget.
I want to run a cold outbound gifting play for [ACCOUNT NAME]. Before building the
play, pull the following:
1. Demandbase Company Global Directory — industry, employee count, revenue range,
installed technologies, and any recent company news for [ACCOUNT NAME].
2. Demandbase Contact Global Directory — [JOB FUNCTION OR TITLE] contacts at
[ACCOUNT NAME] not currently in our CRM.
3. Salesforce — any existing notes, logged activities, or open opportunities for
[ACCOUNT NAME].
Return the results as a context block. I'll add my goal, gift type, and budget
before we proceed.
Example prompt:
Context:
- Target account: [ACCOUNT NAME]
- Industry: [pulled from Demandbase]
- ABM tier: [TIER 1 / TIER 2]
- Outreach history: [pulled from Salesforce — or "No prior touches found"]
- Goal: [GET FIRST MEETING / REACH SPECIFIC PERSONA]
- Gift type: [EGIFT / PHYSICAL] — Budget: up to $[AMOUNT]
Steps:
1. Use the Demandbase Company Global Directory to pull firmographics, recent company
news, and technographics for [ACCOUNT NAME].
2. Use the Demandbase Contact Global Directory to find [JOB FUNCTION OR TITLE] contacts
at [ACCOUNT NAME] not currently in our CRM.
3. Pull any available Salesforce notes or open opportunity data for this account.
4. Search the Sendoso marketplace for a [GIFT TYPE] up to $[BUDGET] that fits the
account context. Recommend a specific product with a draft message tailored to
what you found.
5. Before executing, show me the recommended contact, proposed send, and draft message
for approval. Do not execute until I confirm.
Expected output: A summary of the account's context, a recommended contact with their title and the basis for their selection, the proposed send and message, and a clear confirmation prompt before execution.
Scenario 2: Acting on an Intent Spike Across a Target Account Cohort
Setup: You have 50 accounts in your Tier 1 list. Some are active. Most are dormant. Demandbase is showing that a subset have spiked in engagement this week. You want to reach them while the signal is current.
How the MCP helps: The AI queries your Demandbase 1st party account database for accounts on your list that have crossed your engagement threshold. For each qualifying account, it identifies the top engaged contacts matching your target persona. It searches the Sendoso marketplace for a send within your approved budget and proposes a gift with a message personalized per account. You review the full list and confirm before any sends execute.
Pull your context first:
Run this to get your available Demandbase account list names before filling in the main prompt. You'll need the exact list name — this gives you the options.
List all account lists configured in my Demandbase instance using Get Demandbase
Reference Data. Return the list names and IDs.
Example prompt:
Context:
- Account list: [ACCOUNT LIST NAME] in my Demandbase instance
- Signal threshold: Accounts with [X+] engagement minutes in the last [7 / 14 / 30] days
- Journey stage filter: [CONSIDERATION / DECISION / SPECIFIC STAGE]
- Target contact profile: [JOB FUNCTION OR TITLE — e.g., Director of ABM, VP Marketing]
- Gift type: [EGIFT / PHYSICAL] — Budget: up to $[AMOUNT]
- Goal: [DRIVE MEETING REQUEST / ACCELERATE PIPELINE / SPECIFIC OBJECTIVE]
Steps:
1. Search my Demandbase Internal Account Database for accounts on [ACCOUNT LIST NAME]
that meet the engagement threshold and journey stage above.
2. For each qualifying account, find the top engaged contact matching [JOB FUNCTION OR
TITLE] using the Internal Person Database.
3. Search the Sendoso marketplace for a [GIFT TYPE] up to $[BUDGET] appropriate for
this play. Use the same product across all contacts unless a specific account warrants
a different selection.
4. For each contact, draft a personalized message based on their account's engagement
context and journey stage.
5. Show me the full proposed list — account, contact, send, and message — before
executing anything. Do not execute until I confirm each row.
Expected output: A prioritized list of qualifying accounts with recommended contacts, proposed sends, and draft messages. Each row requires your sign-off before execution proceeds.
Scenario 3: Multi-Threading a Stalled Deal Using Buying Group Engagement Data
Setup: You're single-threaded on a champion at a Tier 1 account. Mid-stage deal, champion gone quiet for three weeks. Many sellers continue to focus on a single champion. The smarter move is to find out who else in the buying group is still active — and reach them before the deal dies.
How the MCP helps: Demandbase does the diagnostic work. The AI queries your 1st party Account Database in Demandbase to check the account's current pipeline predict score and total engagement minutes. If the account is still active, the deal is worth pursuing — the champion going quiet doesn't mean the buying committee did. Then AI queries the Demandbase 1st party Person Database to pull all known contacts at the account ranked by engagement points. This surfaces who is currently visiting your site, consuming content, and building a score — people you may never have talked to but who are already in the evaluation. The AI cross-references those names against Salesforce to separate known contacts from net-new ones. For the top-engaged contacts not yet in your CRM, it proposes a targeted gift to open the thread. For the champion, it pulls Gong transcripts and Salesforce notes to draft a re-engagement message grounded in prior conversations. You end up running coordinated sends into the active parts of the buying group rather than waiting on one person to resurface.
Pull your context first:
Run this to auto-populate deal stage, champion details, and Demandbase account health before filling in the main prompt.
I want to run a buying group play for [ACCOUNT NAME]. Before building the play,
pull the following:
1. Salesforce — open opportunity for [ACCOUNT NAME]: primary contact name, title,
deal stage, and last activity date.
2. Demandbase Internal Account Database — pipeline predict score, total engagement
minutes in the last 30 days, and journey stage for [ACCOUNT NAME].
Return the results as a context block. I'll add my target buying group profile,
gift type, and budget before we proceed.
Example prompt:
Context:
- Account: [ACCOUNT NAME]
- Known champion: [pulled from Salesforce — contact name and title]
- Last champion engagement: [pulled from Salesforce — last activity date]
- Deal stage: [pulled from Salesforce]
- Account engagement: [pulled from Demandbase — pipeline predict score and engagement minutes]
- Target buying group profile: [JOB FUNCTION OR TITLE — e.g., VP Marketing, CFO, IT Director]
- Gift type: [EGIFT / PHYSICAL] — Budget: up to $[AMOUNT]
- Goal: Re-open the deal by reaching active members of the buying group
Steps:
1. Search the Demandbase Internal Account Database for [ACCOUNT NAME]. Return the
account's current pipeline predict score, total engagement minutes in the last 30
days, and journey stage.
2. If the account is still showing engagement, search the Internal Person Database for
all contacts at [ACCOUNT NAME]. Rank them by engagement points, highest first.
Filter for [JOB FUNCTION OR TITLE] if multiple contacts appear.
3. Pull the Salesforce contact record for each top-ranked contact to identify which
are already in our CRM and which are net-new.
4. For each net-new contact with significant engagement points, search the Sendoso
marketplace for a [GIFT TYPE] up to $[BUDGET] and draft a short opening message
referencing the account context.
5. For [CONTACT NAME], pull any available Gong call transcripts and Salesforce notes.
Draft a separate re-engagement message that references something specific from
prior conversations.
6. Show me the full proposed list — contact, engagement score, send, and draft message
— before executing anything. Do not proceed until I confirm each send.
Expected output: A ranked list of active buying group contacts with their engagement scores, a recommended send and draft message for each, and a clear separation between net-new contacts and known ones. Each send requires your confirmation before execution.
Scenario 4: Full Demandbase-to-Send Workflow Across Your ABM List
Setup: You have a named account list in Demandbase. Some accounts are in the right stage for a send play. Most aren't. You want to run a coordinated send campaign against only the accounts showing the right signals — without manually sorting through the list yourself.
How the MCP helps: This is where both MCPs work together. The AI pulls your Demandbase account list, applies your signal filters (journey stage, qualification score, engagement minutes), identifies the best contact at each qualifying account, pulls Salesforce context to personalize the message per account, and searches the Sendoso marketplace for appropriate gifts within your approved budget. You get one consolidated view of every proposed send for review before a single one executes.
Pull your context first:
Same as Scenario 2 — get your available account list names from Demandbase before filling in the main prompt.
List all account lists configured in my Demandbase instance using Get Demandbase
Reference Data. Return the list names and IDs.
Example prompt:
Context:
- Account list: [ACCOUNT LIST NAME] in my Demandbase instance
- Signal criteria: Qualification score above [SCORE], journey stage [STAGE], minimum
[X] engagement minutes in the last [X] days
- Target contacts: [JOB FUNCTION OR TITLE] at each account
- Gift type: [EGIFT / PHYSICAL] — Budget: up to $[AMOUNT]
- Goal: [CAMPAIGN GOAL — e.g., drive meeting requests from Stage 3 accounts before
end of quarter]
Steps:
1. Pull [ACCOUNT LIST NAME] from Demandbase using Get Demandbase Reference Data,
then filter using Search Internal Account Database for accounts that meet all
signal criteria above.
2. For each qualifying account, find the top contact matching [JOB FUNCTION OR TITLE]
using the Internal Person Database. If no match exists, use the Contact Global
Directory to find one.
3. For each account/contact pair, pull Salesforce notes and open opportunity data.
4. Search the Sendoso marketplace for a [GIFT TYPE] up to $[BUDGET]. Recommend a
specific product for each contact with a personalized message based on the
Salesforce and Demandbase context you found.
5. Compile the full proposed send list — account, contact, proposed gift, and message —
and show it to me for review. Do not execute any sends until I confirm the list.
Expected output: A complete send queue with account, contact, proposed gift, and a personalized draft message for each row. You review, edit where needed, and confirm before execution. Any account that doesn't meet the signal criteria is excluded automatically.
Getting Started
- Enable the Sendoso MCP. Your org admin needs a Core, Advanced, or Enterprise plan to turn on MCP access. Once enabled, connect Sendoso as a custom connector in your MCP-compatible AI client using OAuth. Contact your Sendoso CSM to confirm your account is provisioned.
- Enable the Demandbase MCP. Contact your Demandbase CSM to request MCP enablement for your tenant. Claude, ChatGPT, VS Code, Gemini and Copilot are officially supported clients. Full setup instructions are at developer.demandbase.com/docs/mcp.
- Identify your Demandbase account lists. Know the name of the account list you want to work from before running a play. Use the Get Demandbase Reference Data tool to retrieve your configured lists and confirm the names match what you'll reference in your prompts.
- Know your approved send budget by tier. The Sendoso MCP enforces your org-level maximum send price automatically. Before running your first play, confirm with your admin what the approved budget range is for Tier 1 vs. Tier 2 sends — that's what you'll supply as the budget parameter in each prompt.
- Run Scenario 4 against a small cohort first. Start with 10–15 accounts before scaling to the full list. Review the proposed sends, check message quality, and calibrate your signal thresholds before opening it up.
