Salesforce sales cloud consultant

Sales Cloud Consultant — Skills Reference

This file is an operational playbook, not an exam outline. Every section states the rule as an actionable instruction, gives the real limits/numbers, tells you when to choose one tool over another, and flags the anti-patterns to catch in review. Read the Operational Rules Quick Reference first.

Overview

The Salesforce Certified Sales Cloud Consultant credential validates the ability to translate business requirements into scalable, maintainable Sales Cloud configurations across the full sales lifecycle (lead → opportunity → order → analytics). The lasting value is the platform-design judgment it exercises: sharing models, declarative-vs-code decisions, data migration, governor limits, and deployment discipline — applicable to any Salesforce org.

Exam code: Sales-Con-201 Credential level: Consultant (intermediate) Maintenance: Pass a short release-maintenance module each Salesforce release cycle; the credential expires if not maintained.

The exam outline was restructured in June 2024, collapsing nine topic areas into five consolidated domains. The skills tested are identical; only the labels/groupings changed. The blueprint was subsequently updated to add a Predictive and Generative AI domain (~13% [volatile — verify live]); exact current domain weights are in references/study-resources.md.

Load this skill when… scoping or implementing Sales Cloud; designing lead conversion, opportunity pipeline stages, or forecasting; picking automation tools (Flow vs Apex) for a sales workflow; modeling the sales data layer or sharing model; planning a data migration or deduplication; or building sales reports and dashboards. Not this skill: service console or case management → see salesforce-service-cloud-consultant; external portals or communities → see salesforce-experience-cloud-consultant; general org admin, profiles, or permission sets not tied to a sales implementation → see salesforce-administrator.

Deeper context: Study resources live in references/study-resources.md (loaded on demand). For org-specific applications of these rules, see a per-org appendix you maintain in your own project, referenced from a CLAUDE.md. For NPSP/nonprofit-specific guidance, see salesforce-nonprofit-cloud-consultant.

Verify steps assume nothing about your tooling — use your project's Salesforce MCP connection, the Salesforce CLI (sf), or the Salesforce setup UI, in that order of preference.

---

Credential logistics and study path: see references/study-resources.md.

---

Uncertainty & Escalation

• Always re-verify live: volatile facts in this skill include limits and caps ([volatile — verify live] inline below), license pricing and tier names, sandbox refresh intervals, feature availability (e.g. Einstein Scoring minimum record thresholds), and Apex governor-limit numbers — these can change between Salesforce releases.
• Live wins: when the live org or official documentation contradicts a statement in this file, trust the live source and log the discrepancy via the Feedback protocol below.
• Escalate to a human before proceeding on: OWD or sharing-model changes in production; any destructive data operation (delete, mass update) on production data; Connected App / ECA creation or authorization changes; deactivating managed-package automation in production; license purchases or org-type upgrades.
• Confidence taxonomy: every fact in this file is considered stable unless tagged [volatile — verify live] or [opinion — house style]. If you act on an untagged fact and the live system disagrees, that is a signal to file feedback, not to silently trust this file.

---

1. Automation: choosing the right tool (declarative vs. code)

The rule: always solve at the lowest-power tier that meets the requirement. Reach for code only when declarative tools genuinely can't do it. Over-engineering (an Apex trigger for what a validation rule does) is a review failure, not a flex.

Requirement	Use	Why / threshold
Block a bad save with a field-level condition	Validation Rule	Runs before save, no code, shows inline error. First choice for data integrity.
Field default, simple field update on same record, screen-based intake	Flow (record-triggered or screen)	Salesforce's strategic no-code engine. Workflow Rules and Process Builder are retired for new builds — do not create them.
Multi-object orchestration, complex branching, subflows	Flow	Still the default before code.
Bulk logic, recursion control, complex rollups, callouts with retry, > Flow's comfort zone	Apex trigger / class	Justify it. Triggers run in bulk and you control the transaction.
Cross-record dedupe at entry	Duplicate + Matching Rules	Declarative; no code.
Prevent two flows/triggers on the same object fighting	One trigger per object + handler class	Order of operations is otherwise non-deterministic.

Decision criteria:

• Validation rule vs. Flow: if you only need to reject a save, use a validation rule. If you need to change data or do something, use a Flow.
• Flow vs. Apex: if a Flow needs > ~3 nested loops, does DML inside a loop you can't bulkify, or needs guaranteed bulk-safe behavior at scale, move to Apex. Multi-object writes that must be transactional and bulk-safe (e.g. a record-approval handler that upserts a Contact, creates related records, and sets role flags) belong in Apex precisely because of those guarantees.

Anti-patterns / red flags:

• New Workflow Rule or Process Builder — both retired; rebuild as Flow.
• Apex doing what a validation rule does.
• Multiple record-triggered Flows + a trigger on the same object with no defined order.
• A managed-package Workflow Rule silently mutating your data. Managed packages can ship workflow rules that copy or overwrite field values on insert/update — always audit installed automation when a field value changes unexpectedly. (e.g. NPSP's npe01 package; see salesforce-nonprofit-cloud-consultant for NPSP-specific details.)

---

2. Apex governor limits & bulkification

Bulkify everything. Never put SOQL or DML inside a for loop. Full limit table, decision criteria, and anti-patterns: references/apex-limits.md — load when writing or reviewing Apex triggers, Batch Apex, or callout logic.

Key numbers to hold [volatile — verify live]: 100 SOQL / 150 DML / 50,000 rows / 10,000 DML rows / 200 records per trigger batch / 10,000 ms CPU / 6 MB heap (sync).

---

3. Field-Level Security vs. Object access — they are separate

Object access (CRUD on the object) does NOT grant field access (FLS). A user or integration can have full Read/Edit on an object and still get "Invalid field" on a SOQL query because FLS on that field is granted to no one.

The rule: every field a profile/permset/integration user needs to read or write must have an explicit <fieldPermissions> entry in a profile or permission set. There is no inheritance from object access.

Hard deployment trap: SFDX field-meta.xml does NOT carry FLS. Deploying a custom field via SFDX creates the field but grants FLS to no one — not even System Administrator. You must follow every field deploy with a permission set (or profile) that lists <fieldPermissions>.

Two more rules that bite on the permset:

• Required fields cannot appear in <fieldPermissions>. Salesforce rejects the deploy with "You cannot deploy to a required field" — required fields are always visible/editable, so omit them from the permset's field permissions.
• Restricted picklist values rejected by the API look like a field problem but are a value-allowlist problem. A write of a value not in a restricted picklist's allowed set fails even with full FLS — describe the field and check its allowed values before assuming any string is valid.

Anti-patterns / red flags:

• A field added in SFDX with no matching permset change → guaranteed "Invalid field" at query time.
• Trusting a browser-AI tool that "assigned the permission" without an API check.

---

4. Lookups, relationships & the data model

Rules:

• relationshipName must be unique per parent object. Two Lookups on the same child both pointing at the same parent object cannot share a relationshipName — deploy fails with "Duplicate relationship name". Use role-specific suffixes. SOQL parent traversal (Parent__r.Name) keys on field name, not relationshipName, so renaming the relationship is safe.
• Master-Detail vs. Lookup: Master-Detail cascades delete, locks the child to the parent, supports roll-up summary fields, and inherits sharing. Lookup is loosely coupled and optional. Choose Lookup when the child should survive parent deletion and have independent sharing — e.g. an immutable audit record that must persist forever even if its related person record is reworked.
• External IDs for upsert: mark a field External Id + Unique to upsert idempotently from an integration. Keying re-link operations on a stable External Id (e.g. a submission identifier) lets later writes find the same parent records with no new write logic.

Field-length discipline: any string an integration writes to a Salesforce field should be validated against that field's real max length at the boundary, not silently truncated downstream. When length/picklist constants are generated from the org's metadata, treat the generated files as read-only and regenerate them when the org changes — never hand-edit, which drifts from the org silently.

Anti-patterns / red flags:

• Two same-target lookups sharing a relationshipName.
• A hand-typed length/enum constraint instead of one derived from the org's actual metadata.
• Master-Detail used where the child must outlive the parent.

---

5. Sharing & security model

Design from the most-restrictive baseline up. OWD sets the floor; you open access selectively.

Mechanism	When to use
Org-Wide Defaults (OWD)	Set the baseline. Private for sensitive PII; Public Read-Only only if everyone should see everything.
Role Hierarchy	Grant managers access to subordinates' records. Vertical access.
Sharing Rules (owner/criteria-based)	Open records laterally to a role/group beyond the hierarchy.
Manual / Apex sharing	One-off or programmatic grants.
Teams (Account/Opportunity)	Per-record collaborator access without changing OWD.
Permission Sets / Permset Groups	Grant abilities (object CRUD, FLS, system perms) additively. Prefer over editing profiles.

Rules:

• OWD is the floor — you can only widen, never narrow, with sharing rules. If a requirement is "most users see nothing, a few see all," set OWD Private and share up.
• Permission sets over profiles for granting access — additive, composable, and the modern best practice (Salesforce is deprecating permissions on profiles).
• External Client App / Connected App authorization is configured on the app, not via a "permission set assigned apps" page. In modern org configs the working path is ECA detail → Policies → Edit → App Policies → Select Permission Sets. The classic "Assigned Connected Apps" page does not authorize ECA usage. Verify via API (PermissionSetAssignment, SetupEntityAccess) before trusting it.
• PII / sensitive documents: keep OWD restrictive; never log PII; keep file contents in object storage (encrypted), referenced by key, never duplicated into Salesforce field values.

Anti-patterns / red flags:

• Editing a profile to grant something a permission set should grant.
• Public Read/Write OWD on an object holding PII.
• Assuming an ECA is authorized because the permset "assigned apps" page was edited.

---

6. Data management: migration & quality

Tool selection by volume and object:

Tool	Use when
Data Import Wizard	≤ 50,000 records [volatile — verify live], standard + some custom objects, simple loads, built-in dedupe.
Data Loader	Bulk (millions), all objects, insert/update/upsert/delete, scriptable/automatable.
Bulk API 2.0	Programmatic large-volume loads; async, chunked.
Purpose-built package import tools	e.g. NPSP Data Import / Gift Entry for nonprofit-shaped imports — use the org's native import tool when one exists.

Rules:

• Use upsert with an External ID for idempotent, re-runnable migrations — no duplicates on re-run.
• Load order = dependencies first: objects before fields, parents before children, permission sets before assignments, FLS before data that needs the fields visible.
• Always have a rollback plan (keep source keys, load in reversible batches).
• Dedupe before and after: configure Matching + Duplicate Rules to prevent duplicate records.

Anti-patterns / red flags:

• Insert (not upsert) on a re-runnable migration → duplicate explosion.
• Loading child records before parents exist.
• Migrating into fields whose FLS hasn't been granted (loads silently drop or error per-row).

---

7. Large Data Volume (LDV) & performance

Data skew thresholds, SOQL selectivity rules, skinny tables, and Big Objects: references/ldv-performance.md — load when designing for millions of records or diagnosing query timeouts.

Core rules inline: avoid ownership skew (> ~10,000 records per user) and lookup skew (> ~10,000 children per parent); always filter SOQL on an indexed field to avoid full table scans.

---

8. Lead, Opportunity & sales pipeline design

Rules:

• Lead conversion maps Lead → Account + Contact + (optional) Opportunity. Configure field mapping in Setup so custom lead fields don't vanish on convert.
• Opportunity Stage drives Forecast Category and probability. Keep them aligned; a stage with the wrong forecast category corrupts pipeline reports.
• Sales Process + Record Type restrict which stages a given record type sees.
• Validation rules enforce stage gates (e.g. can't reach "Closed Won" without required fields).
• Price books: every product needs a Standard Price Book entry before it can go on a custom price book. Associate currency entries for multi-currency orgs.
• CPQ/Revenue Cloud vs. native Quotes: native Quotes for simple line items; CPQ when you need configurable bundles, complex discounting, approval-gated pricing.

Anti-patterns: orphaned lead custom fields lost on convert; stage/forecast category drift; a product on a custom price book with no standard price.

---

9. Implementation & deployment discipline

Rules:

• Sandbox → UAT → Production. Never deploy untested metadata straight to prod. Sandbox types: Developer (config only, daily refresh [volatile — verify live]), Developer Pro (more storage), Partial Copy (config + sample data, 5-day refresh [volatile — verify live]), Full (complete copy, 29-day refresh [volatile — verify live] — use for UAT/performance/staging).
• Deployment order of operations: objects → fields → FLS (permsets) → automation → assignments. Get this wrong and the deploy fails on missing dependencies.
• Use source-driven deploys (SFDX / sf project deploy, DevOps Center) over Change Sets for repeatable, version-controlled metadata.
• Run all sf project ... commands from the SFDX project root, not the surrounding repo root, or they fail with "InvalidProjectWorkspaceError".

Common deployment traps:

• Connected App creation can be gated by org policy — deploying a connectedApp-meta.xml may return "You can't create a connected app…". The modern workaround is an External Client App; stage classic Connected App metadata for the day creation is unblocked.
• External Client App Consumer Key is UI-only — no Tooling/REST/metadata path exposes it; it's behind email verification in the SF UI.
• Quick Action layout cache survives deploy + logout/login. Adding fields to an existing Quick Action's quickActionLayoutItems via SFDX updates the metadata, but the runtime QA cache (driving Lightning contextual tabs via console:relatedRecord) often does not invalidate — the new fields are silently absent with no error. Cache-bust by editing any non-field-list metadata on the QA (<description>, <label>, <layoutSectionStyle>) and redeploying; SF treats it as a structural change and flushes the org-level cache.
• Sandbox creation may need a Public Group depending on org policy — a one-time setup step.

Anti-patterns: deploying fields without their FLS permset in the same change; running sf from repo root; trusting a QA field-add without the cache-bust.

---

10. Reporting, dashboards & analytics

Rules:

• Pick the right report type. Standard for single-object; custom report types for cross-object and to expose lookup fields not in standard types.
• Reporting Snapshots capture report results to a custom object on a schedule — the only native way to trend point-in-time data (pipeline-over-time, backlog-over-weeks).
• Dashboards run as a single "running user" — every viewer sees that user's data visibility. Use a running user whose access matches the audience, or dynamic dashboards for per-viewer data.
• Multi-currency converts to the corporate currency in reports; be explicit about which currency a number is in.

Anti-pattern: a dashboard whose running user can see more than the audience should → data leak.

---

11. AI & sales productivity features

Decision criteria for Einstein Scoring, Sales Engagement, and Agentforce (formerly Einstein Copilot) — including data-volume prerequisites and when not to recommend heavyweight AI features: references/ai-sales-features.md — load when evaluating or recommending AI add-ons for a Sales Cloud implementation.

Rename: Einstein Copilot was renamed Agentforce (officially "Agentforce Assistant, formerly Einstein Copilot") in the Spring '25 release. In exam and implementation contexts, use "Agentforce for Sales" for the agent-based sales assistant. The product functionality is unchanged; only the name changed. The exam's Predictive and Generative AI domain (~13% [volatile — verify live]) covers Agentforce for Sales, predictive AI tools, and Salesforce's Trusted AI Principles (Responsibility, Accountability, Transparency, Empowerment, Inclusivity).

Core rule to hold inline: Einstein Lead/Opportunity Scoring needs history — minimum volume of closed/converted records (hundreds+) [volatile — verify live]. Do not recommend for low-volume orgs.

---

Operational Rules Quick Reference

DO:

• Solve at the lowest-power tier: Validation Rule → Flow → Apex. Justify any code.
• Bulkify: query once into a Map, DML once. Respect 100 SOQL / 150 DML / 50k rows / 10k DML rows / 200 per trigger / 10s CPU.
• After any SFDX field deploy, add explicit <fieldPermissions> in a permset — SFDX field-meta grants FLS to no one.
• Omit required fields from <fieldPermissions> (deploy fails otherwise).
• Use unique relationshipName per parent object; suffix by role.
• Validate string lengths and picklist values against the field's real metadata; treat generated metadata constants as read-only.
• Use upsert + External ID for idempotent migrations; load parents/objects/FLS before children/data/assignments.
• Prefer permission sets over profile edits.
• Authorize an External Client App via ECA → Policies → App Policies → Select Permission Sets; verify via PermissionSetAssignment API.
• Run sf project ... from the SFDX project root, not the repo root.
• Cache-bust a Quick Action by editing <description>/<label> and redeploying.
• Verify FLS, assignments, and rendered tabs against the org, not just deploy success.

DON'T:

• Never put SOQL or DML inside a for loop, or assume Trigger.new has one record.
• Don't create new Workflow Rules or Process Builders (retired) — build Flows.
• Don't rely on Apex truncation to "handle" length — validate upstream.
• Don't hand-edit generated metadata-constant files.
• Don't use Master-Detail where the child must outlive the parent (audit records).
• Don't insert (vs. upsert) on a re-runnable migration.
• Don't set Public OWD on objects holding PII/sensitive data; never log PII.
• Don't trust a browser-AI "I assigned the permission" without an API check.
• Don't assume a Quick Action field-add rendered just because the deploy passed.
• Don't deploy fields without their FLS permset in the same change.

---

12. Territory Management & Forecasting

Territory Management:

• Enterprise Territory Management (ETM) is the current model (classic is legacy). ETM uses a hierarchy: Territory Types → Territory Models → Territories. Assignment rules auto-assign Accounts (and optionally Opportunities) by field criteria.
• A territory model must be Activated before it affects record access; only one model is active at a time. Test in Draft state, activate, roll back by activating a prior model.
• Territories extend the sharing model — a rep gains access to Accounts in their territory even if OWD is Private, without a sharing rule or role hierarchy change.

Forecasting:

• Collaborative Forecasting aggregates Opportunity amounts up the role/territory hierarchy. Configure which fields roll up (amount, quantity, revenue schedule).
• Forecast Categories are distinct from Stage — Stage is a pipeline position; Forecast Category is the rep's confidence bucket (Pipeline, Best Case, Commit, Closed). A category mismatch corrupts the forecast. Review and correct in: Setup → Forecast Settings → Opportunity Stages.
• Adjustments (manager overrides on a subordinate's forecast number) do not change the underlying Opportunity data — they layer a delta on top. Useful for management confidence corrections without touching rep records.
• Quota management is separate from forecast — quotas are loaded via Data Loader or the Forecasting UI into the Forecasting Quotas data model.

---

13. Consulting Practices & Discovery

Fit/gap matrix, requirements-gathering discipline, Phase 1 scoping rules, and success-metrics definition: references/consulting-practices.md — load when running discovery or scoping a Sales Cloud implementation.

---

Executable Workflows

1. Design lead routing + conversion field mapping

1. Identify the lead sources and routing criteria (region, product, score threshold) with the business owner. → gate: criteria signed off before building any rule entries.
2. In Setup → Lead Assignment Rules, create one active rule; add entries ordered most-specific → most-general (first match wins). → gate: run a test lead through each entry and confirm OwnerId in SOQL.
3. Map custom Lead fields to Contact/Account/Opportunity fields: Setup → Lead Fields → Map Lead Fields. → gate: convert a test lead and confirm custom field values appear on the resulting Contact and Opportunity.
4. Add a validation rule on Opportunity to enforce required fields at the target Stage (e.g. "Closed Won requires Amount"). → gate: attempt to close-won an Opportunity with the field blank — confirm the error appears.
5. Verify assignment rule fires on API-created leads: set AssignmentRuleHeader in the API call, insert a test lead, query Lead.OwnerId. → gate: OwnerId = expected queue/user, not the running user.

2. Stand up Collaborative Forecasting + territories

1. Enable Collaborative Forecasting: Setup → Forecasts Settings → enable, choose forecast types (Revenue, Quantity) and hierarchy (role or territory). [volatile — verify live] → gate: at least one forecast type shows in Setup.
2. Review and correct Stage-to-ForecastCategory mappings: Setup → Opportunities → Fields → Stage picklist. → gate: every Stage value maps to an intentional Forecast Category.
3. Enable Enterprise Territory Management: Setup → Territories → enable ETM, create a Territory Model. Set model to Draft. → gate: model appears in Setup in Draft state.
4. Define Territory Types, then Territories within the model. Add assignment rules (field-criteria based). Run Rules (→ Preview Assignments). → gate: preview shows expected Accounts per territory before activating.
5. Activate the Territory Model. → gate: log in as a rep and confirm Account visibility matches their territory; no cross-region records visible.
6. Load quotas via Data Loader into the Forecasting Quotas data model. → gate: quota appears on the forecast page for the rep.

3. Configure a price book / product / opportunity stage model

1. Create Products: Setup → Products → New. Ensure each Product has a Standard Price Book entry (IsActive = true). → gate: query SELECT Id, Name FROM Product2 WHERE IsActive = true — product appears.
2. For multi-currency orgs, add the currency-specific Standard Price entry for each currency in use. [volatile — verify live] → gate: Standard Price Book entry exists for each active currency.
3. Create a Custom Price Book; add Products with custom prices. → gate: query SELECT PricebookEntryId FROM OpportunityLineItem on a test Opportunity — confirms the entry resolves.
4. Customize Opportunity Stages: Setup → Opportunities → Fields → Stage. Set Stage, Probability, and Forecast Category for each stage. → gate: no stage has an accidental Forecast Category mismatch.
5. Assign the custom Price Book to a test Opportunity; add Products. → gate: line-item amounts roll up to the Opportunity Amount field correctly.

---

Decision scenarios

Scenario 1 — Automation tool selection (Flow vs. Apex)

Situation: A client wants to automatically create a follow-up Task and update three fields on an Opportunity when the Stage changes to "Negotiation." A developer on the team suggests writing an Apex trigger because "it's more reliable."

Competent move: Build a record-triggered Flow. A Stage change → create a child Task + update three fields on the same record is exactly what record-triggered Flow handles well. No bulk-unsafe anti-patterns are involved, and Flow is the strategic declarative tool. Apex is unjustified overhead here.

Tempting-but-wrong: Agreeing that Apex is "more reliable" and coding a trigger. The reliability argument is a rationalization — for straightforward single-object updates and child-record creation at this scale, Flow and Apex are equally reliable. Adding Apex without a genuine need increases maintenance cost and violates the lowest-power-tier principle.

Verify: Open Flow in Setup, confirm the record-triggered Flow runs After Save with the Stage-change entry condition, creates the Task via Create Records element, and updates the Opportunity fields. Test in sandbox with a single record and a bulk load of 200 to confirm governor-limit safety.

---

Scenario 2 — Sharing model: Private OWD with lateral access

Situation: A B2B company has three regional sales teams. Each team should see only its own Accounts, but regional managers need to see all Accounts across all regions. The default "Public Read/Write" OWD is currently set.

Competent move: Set Account OWD to Private. Assign reps to roles under their regional manager in the Role Hierarchy — managers automatically inherit access to subordinates' records. No sharing rules are needed for this structure because the Role Hierarchy alone provides the vertical access.

Tempting-but-wrong: Leaving OWD at Public Read/Write and attempting to lock records down with validation rules or page-layout tricks. OWD is the floor — you cannot narrow access below it with anything other than OWD itself. Validation rules control writes, not visibility.

Verify: Log in as a rep in Region A and confirm Region B Accounts are invisible. Log in as a regional manager and confirm all Accounts in their sub-hierarchy are visible. Check Setup → Sharing Settings → Account OWD.

---

Scenario 3 — FLS after SFDX field deploy

Situation: A developer deploys a new custom currency field Estimated_Budget__c to Opportunity via sf project deploy. The field appears in Setup. Reps report they cannot see the field on records, and an integration user gets "Invalid field" on SOQL.

Competent move: Deploy a Permission Set that includes <fieldPermissions> for Estimated_Budget__c with readable: true / editable: true, then assign the permset to the affected profiles/users. SFDX field-meta.xml creates the field but grants FLS to no one — not even System Administrator.

Tempting-but-wrong: Re-deploying the field or opening a support case assuming the field is broken. The field itself is fine; the problem is FLS, a separate layer. Also wrong: editing the profile directly instead of using a permission set (profiles are being deprecated as the primary access vehicle).

Verify: Run SELECT Id, SObjectType, Field, PermissionsRead FROM FieldPermissions WHERE Field = 'Opportunity.Estimated_Budget__c' in the org's Tooling API or query via Data Loader. Confirm the permset assignment via SELECT AssigneeId, PermissionSetId FROM PermissionSetAssignment.

---

Scenarios 4–5 (Forecast Category vs. Stage drift; Territory model activation over-broad rules): references/scenarios.md — load for the forecasting/territory gotchas.

---

Study resources & relevance

Study resources (official Salesforce + community) are kept in references/study-resources.md. For nonprofit/NPSP-specific operational guidance, see salesforce-nonprofit-cloud-consultant.

---

Feedback protocol

Using this skill and hit a wall? If you find a claim contradicted by the live system or official docs, a missing rule that cost you a wrong attempt, or a decision this skill gave no criteria for — append an entry in the moment to .skill-feedback/salesforce-sales-cloud-consultant.md at the project root (create it if absent):

date | skill last-reviewed | claim or gap | what you observed instead | evidence (error text / doc URL / query output) | suggested fix

These are harvested back into the skill via the learning loop. When the live system and this file disagree, trust the live system.

Changelog

• 2026-06-09 — Conformed to the 12-dimension skill standard: task-vocab description + Scope block, Uncertainty & Escalation guidance with inline [volatile — verify live] marks, executable workflows, tool-agnostic verify steps, and the feedback protocol above. Exam logistics relocated to references/study-resources.md; last-reviewed set to 2026-06-09. Section 2 (Apex governor limits) condensed to a stub with full detail moved to references/apex-limits.md to keep body within word budget.
• 2026-06-10 — Cycle-4 curation (inbox): (1) Added Predictive and Generative AI domain note (~13%, third-party-sourced, [volatile — verify live]); updated blueprint domain names and weights to reflect post-2024 restructure (Consulting + Implementation merged; AI domain added). (2) Renamed Einstein Copilot → Agentforce throughout §11 and ai-sales-features.md (Spring '25 official rename confirmed). (3) Passing score updated in study-resources.md (68% → [volatile — verify live] with conflicting third-party values 69–73%). Eval probes 13–14 added.

---

Independent educational content to upskill AI agents. Not affiliated with or endorsed by Salesforce; all trademarks belong to their respective owners. "Salesforce," "Sales Cloud," "Einstein," "Flow," "Apex," and related marks are property of Salesforce, Inc., used here solely to identify subject matter. Guidance only — verify against official Salesforce documentation and live orgs before acting. No certification outcome is implied or guaranteed.