Introducing Nektar MCP: Give your AI Agents the Complete Customer Context
- Product
- 11 min
- November 3, 2026
Enterprises are building AI agents to analyze deals, identify risks, monitor account health, and automate revenue workflows. But getting these agents to work reliably requires more than just connecting them to Salesforce or a data warehouse.
Customer conversations, buying group relationships, and engagement history often contain information that never makes it into CRM. And even when the data is available, AI agents still need to figure out how every interaction connects to the right customer, contact, and opportunity.
Every token an agent reads costs money and takes up space in its context window. As deal history grows, less room remains for the agent to reason about it. Agents need a way to find the relevant context without reading every conversation in full.
We’re introducing Nektar MCP, a new way to connect AI assistants like Claude, Cursor, Gemini CLI, and your own custom AI agents directly to the customer relationship data captured and structured by Nektar.
Instead of relying on incomplete CRM records or burning tokens retrieving full information, AI applications can work with Nektar’s AI-generated email and call summaries, then retrieve the full email or transcript when a question needs it. With interactions already linked to the right people, accounts, and opportunities, they can spend fewer tokens gathering context and leave more room for reasoning.
Get our latest insights into your inbox
The missing piece in your AI strategy isn't another model. It's context.
Enterprises are investing heavily in AI. Teams are building internal copilots, account intelligence tools, sales assistants, and agents that can handle increasingly complex workflows.
But there’s a problem that better AI models alone can’t solve.
The data these tools rely on can be incomplete or disconnected. Even when it’s available, the volume of raw conversation history can exceed what a model can process at once.
Think about a typical customer account.
- Salesforce has the opportunity details, but not necessarily every interaction.
- Important customer conversations are buried in individual inboxes.
- Call transcripts contain objections, commitments, and buying signals that never make it into CRM.
- Relationships between contacts, accounts, and opportunities aren’t always captured accurately.
Connecting an AI assistant to Salesforce, Gmail, calendars, and call recording tools gives it access to those sources. Where interactions aren’t already linked, the agent still needs to establish how the records relate to each other.
Which email belongs to which opportunity? Who attended the last meeting? Who else on the team has been speaking to this customer?
Even when this information is available in a data warehouse, teams still need to build and maintain the logic that connects each interaction to the right person, account, and opportunity.
Useful AI needs both: the connections that explain the customer story and an efficient way to read it.
Meet Nektar MCP: Your customer data, connected to your AI
MCP, or Model Context Protocol, is an open standard that allows AI assistants and applications to connect to external data sources.
Nektar MCP uses this standard to make your customer relationship data accessible to AI.
Here’s how it works.
Nektar automatically captures customer interactions across emails, calendars, meetings, and calls. It then identifies the people involved and maps those interactions to the right contacts, accounts, and opportunities.
Through MCP, AI applications can retrieve this connected information and Nektar’s existing email and call summaries directly, alongside data from Salesforce and your warehouse.
Agents can use the summaries to identify relevant conversations, then retrieve full emails or transcripts when they need more detail. The account, contact, and opportunity mappings help them stay focused on the right deal.
Lower token costs on long-running deals
An enterprise opportunity can run for months, accumulating hundreds of emails and dozens of call transcripts. Loading all of that text into a model is expensive and slow. If the history exceeds the context window, the agent may have to truncate it and lose relevant context, or the request may fail.
Nektar MCP lets agents start with the AI-generated summaries Nektar has already produced for emails and calls. They can pull the full email or transcript only when a question needs it.
Because every interaction is already linked to the right account, contact, and opportunity, the agent also skips the extra calls and filtering it would otherwise need to piece the deal together from separate systems.
The result is fewer tokens per query and more room in the context window for the agent to reason about the deal.
What your AI can access through Nektar MCP
- Accounts and opportunities: Account information, opportunity details, ownership, and associated records.
- Contacts and relationships: People involved in an account or opportunity, their roles, and how they’re connected.
- Emails and meetings: Customer communications, meeting history, attendees, and relevant activity details.
- Call transcripts: Full conversations from supported sources, including Zoom, Google Meet, Microsoft Teams, and Gong, where configured.
- Salesforce record mappings: Links between Nektar’s customer data and corresponding Salesforce records.
All through a single MCP connection. And because Nektar captures and connects this information automatically, your team doesn’t need to rely on reps manually logging every interaction.
What can you Build with Nektar MCP?
The possibilities go well beyond asking AI to summarize a customer call.
Here are a few ways your teams can put connected customer data to work.
1. Build GTM agents with connected customer context
Enterprises are developing AI agents to monitor pipeline health, identify deal risks, recommend next steps, and analyze customer engagement.
These agents need access to the conversations and relationships behind every opportunity.
With Nektar MCP, teams can build applications such as:
- Deal intelligence agents: Analyze opportunity activity, stakeholder participation, and recent customer conversations to surface potential risks.
- Account health dashboards: Bring together interaction history, relationship coverage, and engagement trends to identify accounts that may need attention.
- Meeting preparation assistants: Start with recent email and call summaries, then retrieve specific conversations to prepare account briefings.
- Win-loss analysis agents: Examine historical customer conversations and opportunity context to identify patterns across deals.
Rather than building separate integrations and relationship-mapping logic for every application, teams can use Nektar as a common source of connected GTM data.
2. Give RevOps richer data without expanding Salesforce
RevOps teams often need to make additional customer information available for analytics, automation, and AI workflows.
But capturing every meeting participant, transcript, or relationship detail in Salesforce can require new fields, configuration changes, and additional storage.
Nektar MCP adds another way for agents to access that context alongside your Salesforce workflows.
- Give agents access to detailed interaction data without creating additional Salesforce fields.
- Retrieve customer context that isn’t represented in standard CRM objects.
- Use Nektar’s existing account, contact, and opportunity mappings instead of maintaining separate linking logic.
- Give AI applications direct access to supported Nektar data alongside the information they retrieve from your warehouse.
This allows RevOps teams to expand their AI capabilities while keeping Salesforce focused on the records and workflows it needs to support.
3. Ask AI about deals, accounts, and customer relationships
Nektar MCP also allows sales and customer success leaders to explore customer information through compatible AI assistants such as Claude, Cursor, and Gemini CLI.
Instead of manually reviewing CRM records, emails, and meeting history, leaders can ask questions like:
- “Summarize our interactions with Acme over the last 90 days.”
- “Who is involved in this opportunity, and when did we last engage with them?”
- “What concerns came up in our recent customer calls?”
- “Which stakeholders have participated in meetings for this deal?”
- “What were the key discussion points and next steps from our last meeting?”
The AI assistant can start with relevant summaries through Nektar MCP and retrieve full conversations when the question requires more detail, grounding its response in customer interactions.
Teams can also use this data to build internal dashboards that present account summaries, engagement trends, and relationship coverage.
Here is an example of a dashboard you can create connecting Nektar MCP:
4. Support customer success workflows outside Salesforce
Not every organization manages its customer success activities in Salesforce. Some teams use dedicated CS platforms or maintain separate workflows for post-sales engagement.
Nektar MCP can make eligible customer success interaction data available to AI applications without requiring every activity to be written into Salesforce.
For example, a customer success agent could retrieve recent conversations, account relationships, and meeting history to prepare for a quarterly business review or renewal discussion.
This gives organizations more flexibility in how they use customer interaction data across their AI applications and operational systems.
Why choose Nektar MCP?
Add conversation context to Salesforce workflows
Salesforce provides account, contact, and opportunity records that support your revenue workflows. Nektar MCP complements those records by giving AI agents access to captured interaction details and summaries.
Alongside Salesforce records, agents can retrieve supported Nektar context, including:
- Complete internal and external meeting participation.
- Broader buying group and selling team relationships.
- Confidence-scored activity-to-opportunity links.
- Full call transcripts and available conversation summaries.
- Supported data that hasn’t successfully reached Salesforce.
AI applications can retrieve this information directly from Nektar without requiring teams to expand their Salesforce schema.
Use Nektar's existing relationship mappings
Data warehouses bring information from multiple sources together and can include relationship mappings and transformations your team has built.
Nektar maps interactions to contacts, accounts, and opportunities as part of its data capture and processing.
Through MCP, agents can use those existing mappings to find the relevant deal history, reducing the extra calls and filtering needed to connect separate records.
Where your warehouse already provides equivalent mappings, agents can continue using them. Nektar MCP adds direct access to Nektar’s interaction data and summaries alongside your existing warehouse queries.
Retrieve the level of detail each question needs
An agent can use Salesforce for opportunity records, your warehouse for broader analysis, and Nektar MCP for linked customer conversations and summaries.
For a deal briefing, it can begin with summaries. For a specific objection or commitment, it can retrieve the relevant email or transcript in full.
This lets teams combine their existing data sources with a way to retrieve customer context that uses fewer tokens than loading the full conversation history for every query.
Built for Enterprise security
Customer conversations contain sensitive business information. Giving AI access to that data requires the right controls.
Nektar MCP is designed with enterprise requirements in mind.
- Read-only access: MCP can retrieve data without modifying customer records.
- Admin-managed setup: Organizations control how their MCP connection is configured.
- Data isolation: Each organization’s data remains isolated.
- Configurable data access: Access to sensitive content, including email bodies and transcripts, can be restricted during setup.
- Works with MCP-compatible AI tools: Connect supported assistants and custom applications through the MCP standard.
Your customer data stays under your organization’s controls while becoming more useful to the AI tools your teams rely on.
Better AI starts with customer context it can use
Revenue agents need to understand the conversations, people, and commitments behind a deal. As that history grows, reading all of it for every question becomes costly and can exhaust the model’s context window.
Nektar MCP gives agents a more efficient starting point: existing summaries, interactions linked to the right deal, and full conversations available when needed.
That means fewer tokens spent assembling and reading deal history, with more room left for the agent to reason about what to do next.
Connected customer context, retrieved at the right level of detail. That’s what Nektar MCP brings to your AI assistants and applications.
Frequently Asked Questions
Enjoyed our content? Follow Nektar on LinkedIn
