Skip to main content
GPTfy - Salesforce Native AI Platform

Salesforce AI · Platform Comparison · 2026

GPTfy vs Agentforce: Two architectures. One Salesforce org.

Both platforms are 100% Salesforce-native. The difference is how they're built, what they require, and which teams can actually get them into production.

GPTfy — Open-architecture AI agent platform

Connect any AI model you already license — OpenAI, Claude, Gemini, DeepSeek, Llama, and 10+ more — to your existing Salesforce data. Fixed per-user licensing. No separate data platform needed. Admin-deployable in hours.

Bring Your Own ModelFixed-price licensingNo Data Cloud required15+ AI models

Agentforce — Salesforce's built-in AI platform

Salesforce's own autonomous agent platform, powered by the Atlas Reasoning Engine. Deeply embedded across Sales, Service, Marketing, and Commerce Clouds. Built on Data Cloud for unified customer intelligence. Designed for teams standardizing on the full Salesforce ecosystem.

Atlas Reasoning EngineNative Salesforce platformData Cloud integrationFlex Credit billing

A note on this comparison: GPTfy is an AppExchange partner and part of the Salesforce ecosystem. This page isn't a takedown — it's a structured guide to help teams understand which architecture fits their situation. Both platforms run inside Salesforce. The right answer depends on your data readiness, team structure, and AI adoption goals.

Core architecture

The foundational difference: how each platform is built

Architecture is a long-term decision. It determines cost behavior, deployment speed, model flexibility, and compliance posture — for years.

GPTfy — Open model architecture

GPTfy runs as a Salesforce-native managed package. It connects to any AI model you configure through Salesforce Named Credentials — the same secure mechanism Salesforce uses for all external callouts. Your existing Salesforce objects are the data layer. No new platform required.

  1. 1

    Your Salesforce org is the platform

    No external servers, no middleware, no data migration. One AppExchange install.

  2. 2

    Any AI model, any provider

    OpenAI, Claude, Gemini, DeepSeek, Llama, Grok, or your own internally-hosted model — connected via Named Credentials. Switch models without rebuilding.

  3. 3

    Prompt-level model routing

    Different prompts can use different models. Summarization on a cost-efficient model. Long-context analysis on a reasoning model. No global org setting forces a single model on every use case.

  4. 4

    BYOM — your contracts, your rates

    If your organization has negotiated Azure OpenAI, AWS Bedrock, or GCP Vertex AI pricing, GPTfy routes through those agreements directly. No duplicate AI licensing.

Agentforce — Platform-native architecture

Agentforce is Salesforce's own AI platform — built directly into the product. The Atlas Reasoning Engine powers agents with chain-of-thought planning, self-correction, and real-time data retrieval. Designed for organizations running the full Salesforce stack with Data Cloud as the unified intelligence layer.

  1. 1

    Atlas Reasoning Engine

    Salesforce's proprietary reasoning system: analyzes goals, retrieves data, constructs multi-step plans, executes actions, and self-corrects when outputs fall short.

  2. 2

    Data Cloud as the intelligence layer

    Agentforce agents retrieve context through Data Cloud's unified customer profiles and vector-embedded documents. Semantic search grounds every agent response in actual organizational data.

  3. 3

    Cross-cloud native orchestration

    Agents operate across Sales Cloud, Service Cloud, Marketing Cloud, and Commerce Cloud natively. Flows, Apex classes, and APIs are all available as action types.

  4. 4

    Data Cloud as the recommended foundation

    While technically optional, most practitioners describe Data Cloud as a prerequisite for Agentforce to perform effectively. It is a separate platform SKU with its own licensing, configuration, and specialist requirements.

The key architectural question: Does your enterprise already have all its customer data inside Salesforce at high quality? If yes, Agentforce's native depth is a genuine advantage. If your data spans multiple systems, or if your team doesn't have Data Cloud specialists, GPTfy's open architecture delivers production AI faster — on the data you already have.

Side-by-side comparison

Every dimension that matters for enterprise Salesforce AI

Based on publicly available documentation, AppExchange listings, and independent analysis.

Deployment and Setup

DimensionGPTfyAgentforce
Installation method1-click AppExchange managed package. Salesforce Security Reviewed. Admin-led.First-party Salesforce feature. Enabled via setup menus within qualifying editions.
Time to first production promptSame day. Prompt Builder + Named Credential config. No data platform prep needed.Timeline varies by data readiness and agent complexity. Data Cloud configuration is typically the primary scoping variable.
Data platform requirementNone. Existing Salesforce objects are the data layer via Data Context Mapping. No extra SKU.Data Cloud recommended for full RAG and unified customer profile functionality. Separate platform with its own licensing and certified specialist requirements.
Developer dependencySalesforce Admins deploy and customize via no-code Prompt Builder. Apex optional for advanced agent skills. Admin-first.Low-code Agent Builder available for basic agents. Complex actions, external integrations, and multi-agent logic typically require Apex development or certified consultants. Dev-assisted.
External infrastructure requiredZero. GPTfy runs entirely as a managed package inside your org. No GPTfy servers. No caching outside Salesforce.Runs within Salesforce's own infrastructure. Data Cloud is an additional Salesforce platform component.

AI Model Flexibility

DimensionGPTfyAgentforce
Supported AI models15+ models: OpenAI GPT-4o, Anthropic Claude, Google Gemini, Azure OpenAI, AWS Bedrock, DeepSeek, Perplexity, Llama, Grok — plus any custom or internally-hosted endpoint. Open.Salesforce platform models plus OpenAI and Anthropic via strategic partnerships. Model selection is primarily platform-managed. Curated.
Per-prompt model routingYes. Each individual prompt has its own model configuration. Canvas Prompts can run 3 models on a single Salesforce record view.Model selection primarily a global platform setting. Per-prompt routing not exposed in current release.
Using existing enterprise AI contractsYes — core feature. BYOM routes prompts through your existing Azure, AWS, or GCP agreements. No duplicate AI licensing.Not a primary use case. Agentforce uses Salesforce-managed model access.
Model switchingUpdate a configuration field. Prompt logic, data context, security rules, and automation actions are untouched.Model updates are tied to Salesforce platform releases and partnership decisions.

Data Context and Content Ingestion

DimensionGPTfyAgentforce
Standard Salesforce objectsNative via Data Context Mapping, up to 3 levels deep.Full native access with real-time metadata awareness.
Custom Salesforce objectsIncluded in Data Context Mapping.Full metadata framework access.
Salesforce Data Cloud (DMO/DLO)Compatible when customer has Data Cloud. Not required.Core integration. Data Cloud is Agentforce's primary context layer.
RAG / Knowledge BaseGPTfy RAG Sync indexes Salesforce Knowledge into vector stores. Enterprise tier feature.Deep RAG grounding via Data Cloud semantic search. Requires Data Cloud + vector embedding pipeline for documents.
File and document analysisAI File Analysis (Enterprise tier). PDFs and attachments analyzed inside Salesforce.Agentforce 360 Intelligent Context (late 2025) enables unstructured document reading without pre-embedding.
Gmail / Exchange email syncAutomatic. AI creates contacts from CC lists and signatures. Zero rep action required.Available via Salesforce Inbox and custom action configuration.
Real-time web searchModel-dependent via BYOM (not a built-in feature).Atlas Reasoning Engine includes built-in web search as a grounding tool.

Security and Compliance

DimensionGPTfyAgentforce
Data masking before AI provider4-layer masking: field-value, regex patterns (SSNs, cards, emails), blocklist keywords, custom Apex handlers. Applied before any data leaves your org.Einstein Trust Layer: PII detection and masking, zero-data-retention policy with LLMs, toxicity filtering.
Data caching outside orgZero. GPTfy operates zero external servers. No caching. Raw data never exits your infrastructure.Zero-retention policy with configured LLM providers. Data stays within Salesforce infrastructure.
Audit trailEvery AI interaction generates a Security Audit Record in Salesforce — queryable for regulatory reporting. Full conversation context captured.Audit logging included in Einstein Trust Layer for compliance review.
Regulatory compliance coverageGDPR, CPRA, HIPAA, FINRA, PCI DSS, FEDRAMP. CheckMarx code-scan reports and source-code escrow available.GDPR, HIPAA, SOC 2. EU AI Act compliance in development. Salesforce's own certifications apply.
Third-party security validationAppExchange Security Reviewed. CheckMarx automated code scans. Source-code escrow for InfoSec due diligence.First-party Salesforce product. Covered by Salesforce's internal security posture and certifications.

Agent and Automation Architecture

DimensionGPTfyAgentforce
Agent orchestration modelJSON schema-based skills via AIAgenticInterface. Multi-agent coordination: sales, service, and CPQ agents in unified Salesforce workflows. Every action logged to Security Audit Records.Atlas Reasoning Engine with ReAct-style multi-step planning, self-correction, and goal decomposition. Agent Topics define scope; Agent Actions (Flows, Apex, API) execute business logic.
Reasoning and self-correctionModel-dependent via BYOM — use any reasoning model such as o1, Claude 3, or DeepSeek R1.Built into Atlas. Chain-of-thought evaluation loop: plan → execute → evaluate → self-correct. Native to the platform.
CPQ AI automationDedicated AI for CPQ: plain-language quote generation, pricing error detection, quote summary. Available now.Configurable via custom agent actions. Requires development effort per use case.
Voice-to-CRMGPTfy Voice: real-time voice-to-text, automatic call logging, next steps extraction. Available as add-on.Agentforce Voice announced in Agentforce 360. Rolling out through 2025 platform updates.
Microsoft ecosystem integrationFull MS 365 Copilot Agent, Outlook, Word, Excel, PowerPoint, Teams. Also on Microsoft AppSource. Governed read/write across the Salesforce–Microsoft boundary.Native Slack integration. Microsoft Office integration available via third-party connectors.

Enterprise and Platform Readiness

DimensionGPTfyAgentforce
Salesforce editions supportedAll editions — Sales Cloud, Service Cloud, Experience Cloud, all Industry Clouds.Enterprise Edition minimum for AI features. Full capabilities require Unlimited or Agentforce 1 editions.
Permission and access modelRole-based masking per Salesforce profile. Admins control feature access via standard Salesforce permission sets.Salesforce security model extended with Agentforce-specific Topics and Actions permissions.
Usage transparencyPer-model token tracking, cost-per-connection dashboards, audit trails. Built-in ROI calculator.Salesforce Digital Wallet tracks Flex Credit consumption in real time. Granular usage reporting per agent.
Coexistence with other Salesforce toolsRuns alongside any Salesforce first-party feature — Einstein, Agentforce, Data Cloud, and all Clouds. Additive.First-party platform designed as the primary AI layer for the Salesforce stack.

Deployment experience

What "going live" actually looks like

Feature lists are theoretical. Deployment experience determines whether AI ever reaches production.

GPTfy — Admin-led deployment

  1. 1

    Install from AppExchange(~15 minutes)

    Salesforce Security Reviewed managed package. No external infrastructure. No sandbox configuration needed.

  2. 2

    Connect your AI model(~30 minutes)

    Register any AI provider via Named Credential in GPTfy Cockpit. One-time setup. Use existing Azure, AWS, or GCP endpoints at your negotiated rates.

  3. 3

    Build prompts in no-code Prompt Builder(1–3 hours)

    Drag in Salesforce fields, configure masking rules, set output format. 100+ pre-built prompts available. Admins — not developers — own this step.

  4. 4

    Deploy via Salesforce permission sets(~1 hour)

    Assign by profile. Reps see AI inside the Salesforce records they already use. No new tool. No new login. No change management required.

Same-day production deployment typical
Found the deployment to be easy. Just took me 10 minutes. Pre-configured prompts and the complimentary Open AI connection helped me test this App quick.
Salman Khan, Release Manager, IBM

Agentforce — Platform-led deployment

  1. 1

    Confirm edition eligibility(Varies)

    AI features require Enterprise Edition minimum. Full agent capabilities require Unlimited or Agentforce 1 editions. License upgrade may be required before any AI work begins.

  2. 2

    Data readiness and Data Cloud setup(Varies)

    Unify customer profiles. Clean duplicates. Configure Data Model Objects (DMOs). Data readiness is foundational to Agentforce performance — the Atlas Reasoning Engine grounds its responses in unified Data Cloud profiles.

  3. 3

    Build vector database for knowledge(1–3 weeks)

    Convert PDFs, help articles, and SOPs to vector embeddings for RAG-grounded responses. Requires certified Data Cloud specialists for configuration.

  4. 4

    Configure Agent Builder, Topics, and Actions(2–8 weeks)

    Low-code for standard agents. External system integrations — ERP, custom APIs — each require custom Apex actions with authentication management and error handling.

Scope-dependent — driven by data readiness and agent complexity

Content ingestion and data context

How each platform grounds AI in your organizational data

AI quality is a direct function of data quality and accessibility. Here's what each platform can read — and how it gets there.

GPTfy — Data Context Mapping

Standard Salesforce objects
Native, no extra config
Custom Salesforce objects
Up to 3 related levels
Related lists (multi-level)
Automatic inclusion
Salesforce Data Cloud (DMO/DLO)
Compatible if you have it
Salesforce Knowledge (RAG)
GPTfy RAG Sync, Enterprise tier
PDF / file analysis (attachments)
AI File Analysis, Enterprise tier
Gmail / Exchange email auto-sync
Automatic, no rep input
External API data sources
Enterprise tier
WhatsApp / telephony
Via native integrations
Real-time web search
Model-dependent via BYOM
Separate vector DB setup required
Not required for standard use
Time to first grounded response
Same day

Agentforce — Data Cloud + Atlas

Standard Salesforce objects
Full real-time native access
Custom Salesforce objects
Full metadata framework access
Related lists (multi-level)
Cross-object native queries
Salesforce Data Cloud (DMO/DLO)
Primary context layer
Knowledge article RAG
Requires vector embedding pipeline
PDF / document ingestion
Agentforce 360 Intelligent Context (2025)
Email auto-sync
Via Salesforce Inbox / custom action
External systems (ERP, etc.)
Custom Apex action per system
WhatsApp / telephony
Via Salesforce Digital Engagement
Real-time web search
Built into Atlas
Separate vector DB setup required
Yes, for document RAG
Time to first grounded response
Weeks (data prep required)

Security and compliance

Enterprise data protection for regulated industries

Financial services, healthcare, insurance, and public sector teams cannot treat security as a post-deployment concern. Here's how each platform is built.

GPTfy security architecture

Zero GPTfy servers, zero caching

GPTfy operates as a managed package inside your Salesforce org. No external infrastructure. GPTfy has zero access to your data. Raw data never leaves your infrastructure.

4-layer PII/PHI masking

Layer 1 — Field-value masking. Layer 2 — Regex patterns: SSNs, credit cards, emails. Layer 3 — Blocklist keywords. Layer 4 — Custom Apex handlers for complex business logic. Masking applied before any payload leaves the org.

Named Credential architecture

API keys stored non-exportably in Salesforce Named Credentials. Never appear in Apex classes, version control, or debug logs. Directly addresses FINRA Rule 4370 and SEC Rule 17a-4 requirements.

Per-interaction Security Audit Record

Every AI call creates a queryable Salesforce record capturing masked input, AI response, user, timestamp, and model used — for GDPR, HIPAA, and FINRA regulatory reporting.

Compliance coverage

GDPR, CPRA, HIPAA, FINRA, PCI DSS, FEDRAMP. CheckMarx automated code-scan reports. Source-code escrow available for InfoSec due diligence. Covered under your existing Salesforce compliance posture.

AppExchange Security Reviewed

Independent validation by Salesforce's security review process for managed packages.

Agentforce security architecture

Einstein Trust Layer

Salesforce's foundational AI security system. PII masking, toxicity detection, zero-data-retention agreements with LLM providers, prompt injection defense, and audit logging — all included.

Zero-retention policy with LLMs

Data sent to configured language models is not used for model training. Salesforce contractually enforces zero-retention with AI provider partners.

GDPR, HIPAA, SOC 2 compliance

Salesforce's enterprise compliance certifications apply to Agentforce. Salesforce Hyperforce supports data residency requirements for international deployments.

Salesforce-native permission model

Agentforce respects existing Salesforce field-level security, object permissions, and sharing rules. No separate permission framework required for basic access control.

FINRA, PCI DSS, FEDRAMP

Covered through Salesforce's overall compliance posture for customers on eligible editions. Individual mapping to specific financial regulatory requirements varies by configuration and cloud edition.

Third-party code validation

Not applicable — Agentforce is a first-party Salesforce product. Security is governed by Salesforce's internal SDLC and independent audit certifications.

Use cases

Where GPTfy delivers AI across your Salesforce org

Production workflows that deploy in hours — not months — on your existing Salesforce data.

Email-to-CRM automation

Service Cloud · Sales Cloud

Gmail and Exchange sync automatically. AI reads inbound emails and attachments, creates contacts from CC lists and signatures, and transforms emails into cases, quotes, or orders — with zero rep action. 5,000+ minutes saved per day at scale.

Account 360 and deal intelligence

Sales Cloud · Revenue Intelligence

Reps walk into every meeting fully briefed in seconds — not 40 minutes of manual research. AI surfaces MEDDIC gaps, flags deals going silent, identifies renewal risks, and generates executive account briefs from Opportunities, Cases, and Contacts in one click.

CPQ AI assistance

Salesforce CPQ · Revenue Cloud

Reps describe what they need in plain language. GPTfy builds the CPQ configuration, checks for pricing errors, and generates quote summaries — eliminating the administrative overhead that costs revenue teams 72% of their time.

Einstein Bot + AI with RAG

Service Cloud · Experience Cloud

Supercharge existing Einstein Bots with GPTfy's knowledge retrieval layer. Bots gain account-specific context, intent analysis, real-time case routing, and knowledge base grounding — without rebuilding. 97% case deflection at production scale.

Voice-to-CRM logging

All Clouds · Telephony

GPTfy Voice captures calls in real time, logs automatically to Salesforce, extracts next steps, and generates call summaries. Eliminates manual note-taking — the number one source of incomplete CRM data — without requiring reps to change any behavior.

Salesforce + Microsoft AI workflows

MS 365 · Teams · SharePoint

GPTfy is the only platform on both AppExchange and Microsoft AppSource. Agents govern read/write across the full Salesforce–Microsoft boundary: retrieve Account 360 → generate a QBR PowerPoint → draft a Word proposal — in one governed, auditable loop.

Liked the easy and click/no-code way to configure GPT LLMs on any Salesforce object and go-live in days.
Gurditta Garg, Chief Salesforce Evangelist, Motorola

GPTfy's structural advantages

What open-architecture AI means in practice

These advantages compound over the life of your deployment — they're not feature checkboxes, they're architectural choices.

No prerequisite infrastructure

GPTfy works with what you have. Existing Salesforce objects, existing AI contracts, existing cloud infrastructure. No new platform to procure, configure, or maintain before AI goes live.

Predictable cost regardless of adoption

Fixed per-user licensing means the cost structure and adoption goals reinforce each other. The more your team uses AI, the better the ROI per dollar — not the higher the monthly bill.

Your existing AI contracts do the work

If your organization already has negotiated Azure OpenAI, AWS Bedrock, or GCP rates, GPTfy puts them to work inside Salesforce immediately. No duplicate AI licensing. No forced vendor lock-in.

Admin-owned, not dev-dependent

Salesforce Admins build, deploy, and maintain production AI prompts via the no-code Prompt Builder. The team that owns your Salesforce org already owns your AI deployment.

Model-agnostic by design

When GPT-5, Claude 4, or a specialized industry model launches, updating GPTfy is a configuration change — not a contract renegotiation or re-architecture project.

Spans both enterprise platforms

The only Salesforce AI platform available on both AppExchange and Microsoft AppSource. Sales and service teams on Salesforce, operations teams on Microsoft 365 — all governed by the same AI layer.

Built for regulated industries

FINRA, HIPAA, PCI DSS, and FEDRAMP compliance built into every interaction. Trusted by PayPal, Manulife, Resolution Life, and Legal & General.

Zero GPTfy data access — ever

GPTfy is a technical orchestrator, not a data processor. It has zero access to your org. Raw data never exits your infrastructure. GDPR Article 28 processor requirements satisfied by design.

Frequently Asked Questions

No. GPTfy's Data Context Mapping grounds AI responses in your existing Salesforce objects — standard, custom, and related lists up to 3 levels deep. There is no additional data platform SKU, no separate licensing, and no weeks-long configuration before your first production prompt runs. For organizations that already have Salesforce Data Cloud, GPTfy connects directly to Data Cloud DMO and DLO objects as well — giving access to unified customer profiles without additional setup. Data Cloud is an option, not a requirement.

Yes — this is GPTfy's core BYOM capability. GPTfy connects to any AI provider through Salesforce Named Credentials. If your organization has negotiated agreements with Azure OpenAI, AWS Bedrock, Google Vertex AI, or direct contracts with Anthropic or OpenAI, GPTfy routes prompts through those existing agreements at your negotiated rates. You pay your AI provider directly. GPTfy's per-user licensing covers the orchestration layer — not the AI inference — so there's no duplicate AI spend.

GPTfy installs from the AppExchange as a Salesforce Security Reviewed managed package — typically 15 minutes. Connecting an AI model via Named Credential takes another 30 minutes. The no-code Prompt Builder lets Salesforce Admins build and deploy production-ready prompts the same day. Most teams go from install to first live AI interaction in under 4 hours. There is no data platform setup, no vector database configuration, and no custom development required for standard use cases.

Yes. GPTfy is built for GDPR, CPRA, HIPAA, FINRA, PCI DSS, and FEDRAMP-aligned environments. Because GPTfy runs as a 100% native Salesforce managed package with zero external servers, it can be covered under your existing Salesforce compliance certifications. GPTfy provides documented Security Narratives, CheckMarx automated code-scan reports, and source-code escrow for enterprise InfoSec due diligence. The Named Credential architecture specifically addresses FINRA Rule 4370 and SEC Rule 17a-4. Every AI interaction creates a Security Audit Record queryable for regulatory reporting.

Yes. GPTfy is an AppExchange managed package that coexists with any first-party Salesforce feature, including Agentforce, Einstein Copilot, Data Cloud, and all Salesforce Clouds. GPTfy does not conflict with or replace Agentforce — it operates as a separate, complementary layer in the same org. Many enterprises use GPTfy for immediate BYOM AI deployments while their Agentforce implementation proceeds in parallel, or run both indefinitely for different use cases.

GPTfy supports 15+ AI models via BYOM: OpenAI GPT-4o, Anthropic Claude, Google Gemini, Azure OpenAI, AWS Bedrock, DeepSeek, Perplexity, Llama, Grok, and any custom or internally-hosted endpoint. Each model is registered as an AI Model record in GPTfy Cockpit once and is then available to any prompt in the org. The model is a configuration on each individual prompt — different prompts can use different models for cost and performance optimization. When new models launch, you update a configuration field, not a contract.

BYOM — Bring Your Own Model — means connecting your organization's own AI provider contracts to Salesforce, rather than using a bundled platform model. For enterprises, this matters for three reasons: first, most large organizations already have negotiated pricing with OpenAI, Azure, or AWS — BYOM lets you use that pricing inside your CRM rather than paying twice; second, you maintain flexibility to switch models as the AI landscape evolves without rebuilding your Salesforce integration; third, your data processing agreements, data residency commitments, and AI governance policies remain between your organization and your AI provider — GPTfy acts as a technical orchestrator, not a data processor in that relationship.

GPTfy applies 4-layer masking before any data leaves your Salesforce org. Layer 1: field-value masking — specific Salesforce fields are masked by configuration. Layer 2: regex pattern masking — automatically detects and masks SSNs, credit card numbers, email addresses, and other patterns. Layer 3: blocklist masking — custom keywords and phrases are redacted. Layer 4: custom Apex masking — complex business logic for sensitive data not covered by the above layers. Only the masked, processed payload is sent via Named Credentials to your configured AI provider. Raw data never exits your infrastructure. Every masking event is logged in the Security Audit Record for that interaction.

GPTfy supports all Salesforce editions — Sales Cloud, Service Cloud, Experience Cloud, and all Industry Clouds — without requiring Enterprise Edition or above as a prerequisite. Agentforce requires Enterprise Edition minimum for AI features, with full capabilities available only in Unlimited or Agentforce 1 editions. GPTfy's AppExchange managed package architecture makes it edition-independent: if you can install AppExchange apps in your org, GPTfy can run there.

See GPTfy running in your Salesforce org.

Your models. Your data. Live in hours — not months. No new infrastructure required.

Available on Salesforce AppExchange and Microsoft AppSource · Salesforce Security Reviewed · No commitment required

Last reviewed: 2026-05-22. This comparison is based on publicly available documentation as of 2026-05-22. Features and pricing subject to change. Agentforce and related marks are trademarks of Salesforce, Inc.. Microsoft, Copilot, and related marks are trademarks of Microsoft Corporation. GPTfy is an independent product available on AppExchange and Microsoft AppSource and is not affiliated with or endorsed by Salesforce, Inc. or Microsoft Corporation beyond marketplace partner status.