AI Integrations
EA GraphLink’s MCP Server connects your Sparx EA repository to Salesforce Agentforce. This is the MCP interface path: the same interface used by every AI assistant integration EA GraphLink supports.
Agentforce is to Salesforce-centric organizations what Microsoft Copilot is to Microsoft-centric ones: the AI assistant embedded in the tools the business already uses. For organizations where Salesforce is the primary business platform: where opportunity management, service delivery, account governance, and business processes run through the CRM: Agentforce brings AI assistance into that environment. EA GraphLink connects the architecture repository to it.
The architecture is identical to the Copilot integration: same EA GraphLink MCP Server, same MDG Technology dependency, same protocol: routed to Agentforce instead of Copilot. Business stakeholders in Salesforce can ask architecture questions and receive answers from live repository data, without leaving the CRM, without logging into EA, without sending a request to the architecture team.
The MDG Technology definition is critical. EA GraphLink transforms the physical repository schema using the MDG Technology definition. Poor MDG quality means poor Agentforce output. Assess MDG readiness in Discover. Establish it in Deploy. Then connect.
Architecture context becomes available where Salesforce-centric business decisions happen.
Your current state: Business decisions in a Salesforce-centric organization happen in the CRM. Opportunity management, account planning, service design, customer lifecycle decisions: all in Salesforce. Architecture questions that arise during those processes require leaving the CRM, finding the right architect, and waiting for a response. In practice, those questions don’t get asked. Decisions get made without architectural context. Architecture remains invisible to the business.
Your future state: A sales leader in Salesforce is managing a complex enterprise opportunity. The prospect asks about integration capabilities. The sales lead types into Agentforce: “What integration patterns does our current architecture support for this type of customer?” Agentforce queries your EA repository via EA GraphLink and returns current information: integration patterns, supporting applications, technology standards. The answer reaches the sales lead in Salesforce, in context, in seconds. No architect required for the lookup. Architect expertise is freed for the questions that require it.
Architecture context is available in the platform where you work. You don’t need to learn EA. You don’t need a separate login. You ask a question in Agentforce and get an answer from your authoritative architecture repository: the same way you would ask Agentforce about a customer record or a product specification.
Architecture visibility reaches business decisions in real time. Decisions that previously proceeded without architecture context: because the friction of getting it was too high: can now include it. The quality of technology-related business decisions improves. Architecture becomes a business input, not a specialist-only resource.
Your team’s reach extends into the business without scaling the team. Architecture data is available on demand across the organization. Architects handle complex analysis, governance, and design: not routine information requests. The repository serves the organization directly.
Architecture Q&A Within Salesforce: Capability and Technology Questions
A business development team is preparing a proposal for a customer in a regulated industry. They need to understand the current state of the compliance architecture supporting the relevant product line. In Agentforce, they ask: “What technology currently supports our customer onboarding capability and is it certified for financial services data?” Agentforce queries the EA repository and returns current information on the supporting applications, their certification status, and any known architectural concerns. The proposal is written with accurate architecture context.
Application Impact Queries from CRM Context
A service delivery manager is planning a change to a customer-facing integration. Before proceeding, they ask in Agentforce: “Which systems would be affected if we changed the integration between our CRM and the billing platform?” The query reaches EA GraphLink, returns the dependency map for that integration, and surfaces it in Salesforce. The manager understands impact before committing to the change.
Technology Lifecycle in Business Context
During account planning, a customer success manager discovers that a product their customer depends on runs on an application approaching end-of-life. They ask Agentforce: “What applications supporting this product line are approaching end-of-life, and what’s the replacement plan?” EA GraphLink returns lifecycle status and roadmap information from the repository. The customer conversation is informed. The risk is surfaced before it becomes a customer issue.
Architecture Input to Opportunity Management
An enterprise account executive is working a complex deal. The prospect has asked about the organization’s capability to support a specific technical requirement. The account executive asks Agentforce during opportunity management: “Do we have an approved architecture pattern for real-time event processing at this scale?” The architecture answer reaches the sales conversation at the moment it matters, not days later via an email chain.
Compliance Status Queries
A Salesforce-based risk function needs to understand whether the platform supporting a particular customer meets current security architecture standards. In Agentforce, they ask: “Does the platform supporting this customer meet our current security and data residency standards?” The EA repository returns compliance and standards information for the relevant architecture domain. The risk assessment has architectural grounding.
| Role | Primary Benefit | Secondary Benefit |
|---|---|---|
| Business Stakeholder (Salesforce user) | Architecture context in Salesforce: no new tools | Accurate answers without architect dependency |
| IT Executive | Architecture visibility in real-time business decisions | Fewer blind spots in technology-related choices |
| Architecture Manager | Architecture reach into Salesforce-based business processes | Architects freed from routine information requests |
| Sales / Business Development | Architecture-grounded responses to customer questions | Faster, more accurate proposal and opportunity support |
| Service Delivery Manager | Impact assessment from CRM context | Earlier identification of architecture dependencies |
| Risk / Compliance Function | Standards and compliance queries in Salesforce | Architecture visibility without leaving the platform |
| Customer Success Manager | Technology lifecycle awareness during account management | Proactive identification of end-of-life risks |
For Salesforce-centric organizations, architecture context should live where business happens.
If Salesforce is your primary business platform, Agentforce puts architecture data exactly where your people work. No new tools. No new logins. No training program. If someone in your organization uses Salesforce: which, in a Salesforce-centric organization, is most people in customer-facing, operational, and management roles: they can ask architecture questions through Agentforce.
The MCP standard means the same deployment serves multiple consumers. The EA GraphLink MCP Server that connects to Agentforce is the same MCP Server that connects to Claude, Cursor, or any other MCP-compatible tool. One deployment. Multiple downstream integrations. If you deploy EA GraphLink for Agentforce, adding Claude or Cursor later is incremental configuration, not a new deployment.
Architecture becomes relevant to Salesforce-driven decisions. Currently, architecture is largely invisible to the people making Salesforce-based decisions: because the friction of accessing it is too high. Agentforce removes that friction entirely. Architecture context arrives in the moment of decision, in the tool where the decision is being made.
Salesforce Agentforce licensing is an additional cost. Agentforce requires Salesforce licensing beyond the standard CRM subscription. Evaluate total licensing cost against the value of architecture-aware decision making in your Salesforce environment. For organizations with broad Salesforce deployments and significant architecture complexity, this is typically well justified. For smaller Salesforce footprints, evaluate carefully.
Architecture data leaves your on-premises environment. EA repository data flows from your EA GraphLink MCP Server to Salesforce’s cloud infrastructure to service Agentforce queries. For regulated industries with data residency requirements, this requires a compliance review before deployment. This is the same consideration as Copilot integration: your architecture data moves through a cloud platform to deliver conversational responses.
If you use both M365 Copilot and Agentforce, rollout complexity doubles. Both integrations can connect to the same EA GraphLink deployment simultaneously: and that’s a genuine advantage (one EA GraphLink, two AI channels). But it means double the governance decisions, double the user rollout, and double the ongoing management. Sequence them if complexity is a concern.
Data quality must be high enough to trust at business scale. Agentforce responses are visible to business stakeholders, sales teams, and executives: not just architects. Poor repository data surfaced in a customer-facing conversation creates real business risk. MDG quality must be validated before Agentforce goes live at scale. Discover assesses this. Don’t skip it.
EA GraphLink with MCP Server enabled. The MCP interface is how EA GraphLink connects to Agentforce. Confirm your EA GraphLink deployment includes an active MCP Server endpoint with your Sparx Systems account manager.
Salesforce Agentforce license. An active Agentforce deployment within your Salesforce environment. If Agentforce is not yet deployed, complete that step with your Salesforce team before configuring the EA integration.
MCP Connector Configuration. The connector links Agentforce to your EA GraphLink MCP Server endpoint. Configuration requires specifying the EA GraphLink endpoint and authentication credentials. Sparx Services configures the EA GraphLink side during Connect. Your Salesforce team configures the Agentforce MCP connector.
MDG Technology quality sufficient to trust at scale. EA GraphLink transforms the physical EA repository schema using your MDG Technology definition. If MDG quality is poor: missing element definitions, inconsistent taxonomy, deprecated content: Agentforce will surface that poor quality to business stakeholders. MDG readiness is assessed in Discover and established in Deploy. Validate before going live.
Data governance decisions made. Which architecture data is appropriate for Agentforce to surface? To whom? Are there business-unit boundaries that should restrict what different Salesforce users can query? These governance decisions affect connector configuration. Make them before deployment, not after.
Stakeholder adoption planning. Agentforce availability changes the information environment. Architects who currently gatekeep architecture information need to understand the shift. Business stakeholders need to know what kinds of questions Agentforce can answer. Plan adoption alongside technical deployment.
Identify the volume of architecture-related information requests from Salesforce users.
How often do sales, service, and operations people in Salesforce need architecture context? How long does it currently take them to get it? How often do they proceed without it because the friction is too high?
Formula:
Stakeholder architecture requests per week
× architect hours per request (research + communication)
× hourly rate
× 52 weeks
× 50–70% self-service adoption rate
= annual architect hours reclaimed
Plus:
Stakeholder hours spent waiting or working around architecture information gaps
× frequency
× stakeholder hourly rate
× self-service adoption rate
= annual stakeholder efficiency gain
Example (Salesforce context):
20 architecture requests per week from Salesforce users
× 1.5 architect hours per request
× $150/hour
× 50 weeks
× 60% self-service adoption
= $135,000 annual architect hours reclaimed
Plus:
Decisions made without architecture context: 5 per week
× 3 hours of downstream correction when issues surface
× $100/hour blended stakeholder rate
× 50 weeks
× 40% reduction from better upfront information
= $30,000 avoided rework
Total annual value: ~$165,000
Tailor these figures to your actual volume of Salesforce-based architecture requests and the scale of your architecture team. Organizations with large Salesforce deployments and complex architectures will see proportionally larger returns.
Microsoft Copilot
The direct equivalent for Microsoft-centric organizations. Uses the same EA GraphLink MCP interface. If your primary business platform is M365 rather than Salesforce, Copilot is the right choice. For organizations with both M365 and Salesforce, both can connect to the same EA GraphLink deployment simultaneously.
Kernaro AI Hub
Kernaro is Sparx Systems’ purpose-built architecture intelligence platform. It provides conversational access to your EA repository without Salesforce dependency or Agentforce licensing. Pros: architecture-specific design, no cloud platform dependency, lower licensing cost. Cons: requires navigation to a separate tool: less frictionless than Agentforce for Salesforce users, less embedded in the Salesforce workflow. Kernaro is an excellent choice if you want dedicated architecture intelligence without Salesforce dependency.
Prolaborate (Browser-Based EA Reporting)
Prolaborate provides self-service dashboards and portal access to EA data. Pros: lower cost, no AI licensing, browser-based access without special tool training. Cons: not conversational, not embedded in Salesforce, requires portal navigation separate from the CRM workflow. Prolaborate is a strong choice for organizations that need stakeholder access to architecture data without AI assistant investment.
Q: How does Agentforce connect to the EA repository?
A: EA GraphLink exposes an MCP Server: a standardized interface for AI tool connectivity. Salesforce Agentforce connects to this MCP Server via an MCP connector. When a Salesforce user asks Agentforce an architecture question, Agentforce queries the EA GraphLink MCP Server, receives the relevant repository data, and generates a response in Salesforce. The EA repository is not directly accessed: EA GraphLink mediates all queries and applies the MDG Technology transformation before exposing data.
Q: Is this the same type of integration as Microsoft Copilot?
A: Yes, technically. Both Agentforce and Microsoft Copilot use EA GraphLink’s MCP Server interface. The integration pattern is identical: same protocol, same MDG dependency, same data flow model. The difference is the AI assistant platform and the business ecosystem it serves. Agentforce is for Salesforce-centric organizations; Copilot is for Microsoft-centric ones.
Q: Can I have both Copilot and Agentforce connected to the same EA GraphLink?
A: Yes. EA GraphLink’s MCP Server can serve multiple downstream AI assistants simultaneously. If your organization uses both M365 Copilot and Salesforce Agentforce, both can connect to the same EA GraphLink deployment. Governance decisions: what each platform can access, for which users: are configured separately per connector. One EA GraphLink deployment, multiple AI channels.
Q: What Salesforce license is required for Agentforce?
A: Agentforce is a separate Salesforce product with its own licensing, layered on top of your existing Salesforce subscription. Licensing terms vary by Salesforce edition and use case. Contact your Salesforce account team for current pricing. Sparx Services does not resell Salesforce licensing: we can provide a bill of materials to support your procurement, but the license is purchased directly from Salesforce.
Q: What happens if my EA repository has low-quality data?
A: Agentforce responses are only as good as your repository data. EA GraphLink uses your MDG Technology definition to transform the physical repository schema: if the MDG definition is incomplete or inconsistent, the exposed data will reflect those gaps. If repository content is outdated or conflicting, Agentforce will surface those conflicts conversationally, visibly, to business stakeholders. This is why MDG readiness assessment (Discover) must precede Agentforce deployment. Poor data quality in an AI assistant creates real business risk: especially in customer-facing contexts.
Q: Does data leave my Salesforce environment?
A: Agentforce queries are processed within Salesforce’s cloud infrastructure. The EA repository data flows from your EA GraphLink MCP Server (typically hosted in your network or cloud environment) to Salesforce’s infrastructure to service each query. Data is not stored in Salesforce beyond the transient query processing: responses are generated and returned, not retained as architecture records in Salesforce. For regulated industries, confirm this data flow model meets your compliance requirements before deployment.
Q: What is the difference between Agentforce and Kernaro AI Hub?
A: Agentforce is a general-purpose Salesforce AI assistant that gains architecture context through EA GraphLink. It lives inside Salesforce and serves all Salesforce users. Kernaro AI Hub is a purpose-built architecture intelligence platform from Sparx Systems. It is dedicated to EA use cases, runs independently of any business platform ecosystem, and does not require Salesforce licensing. Agentforce is the right choice when you want architecture answers embedded in Salesforce workflows. Kernaro is the right choice when you want a dedicated architecture intelligence tool accessible from any browser without ecosystem dependency.
Q: How long does a Connect engagement take to deploy Agentforce + EA GraphLink?
A: Connect engagements range from $50K–$185K+ depending on scope and complexity. Timeline depends on the readiness of your EA GraphLink foundation, the status of your Agentforce environment, and the complexity of governance decisions to be configured. Organizations with a well-governed repository and an active Agentforce deployment can move faster than those requiring MDG remediation alongside integration work. Connect follows Discover, which assesses your readiness and provides a scoped deployment plan.
Agentforce integration makes architecture context available to every Salesforce user in your organization: at the moment they need it, in the tool they already use. For Salesforce-centric organizations, this is the most direct path to making architecture a business asset rather than a specialist resource.
The journey starts with assessing your readiness: repository quality, MDG governance, and Agentforce deployment status. If you haven’t completed a Discover assessment, that’s the right first step.
If you’ve assessed your foundation and you’re ready to connect:
[Start Your Connect Engagement]
Questions about whether Agentforce is the right integration path for your organization?
[Let’s Discuss Your AI Strategy]
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.