What Is MCP (Model Context Protocol), and Why It Matters If You Sell Services
MCP, or Model Context Protocol, is an open standard that lets AI systems connect to external tools, databases, and live data sources through a single, universal interface. For service providers (businesses where pricing, availability, and scope change with every client) MCP is the infrastructure layer that makes AI genuinely useful for dynamic, consultative selling. An AI assistant connected through MCP can actually quote, configure, and sell services in real time.
MCP in 30 Seconds: The "USB-C for AI" Analogy
Think of how USB-C replaced a tangle of proprietary cables with one universal connector. MCP does the same thing for AI applications. Before MCP, every AI tool needed a custom-built integration for every data source: your CRM, your pricing engine, your availability calendar. Each connection was bespoke, fragile, and expensive to maintain.
MCP standardises that connection into a single protocol. An AI assistant that speaks MCP can plug into any MCP-compatible data source without custom engineering for each new tool.
Since Anthropic open-sourced MCP in November 2024, adoption has been extraordinary. MCP server downloads grew from roughly 100,000 at launch to over 8 million by April 2025, with more than 10,000 active public MCP servers running by late 2025. As of December 2025, Anthropic reported over 97 million monthly SDK downloads across all languages. OpenAI, Google DeepMind, and Microsoft have all adopted the protocol. In December 2025, Anthropic donated MCP to the Agentic AI Foundation under the Linux Foundation, co-founded with Block and OpenAI, and backed by Google, Microsoft, AWS, and Cloudflare.
MCP is becoming the standard infrastructure for how AI connects to the real world.
Why Static-Product Companies Had It Easy (and Service Providers Didn't)
If you sell a physical product with a fixed price (say, a €29.99 phone case) AI integration is straightforward. The product has a name, a price, a description, and an image. An AI assistant can look it up in a catalogue and present it. Done.
Service providers operate in a fundamentally different reality. Consider:
- An insurance broker whose premiums depend on property size, location, coverage level, claims history, and a dozen regulatory variables
- A consulting firm whose project scope and pricing shift based on team composition, timeline, and deliverables
- A logistics company whose quotes fluctuate with fuel costs, route availability, and cargo weight
- A marketing agency whose retainer varies by channel mix, market, and campaign complexity
In all these cases, there is no single "price" to display. The price is the output of a calculation that depends on context the customer provides. Before MCP, AI assistants could not perform this calculation. They had no standardised way to reach into pricing engines, underwriting models, or availability systems to pull live data.
This is why so many service businesses watched the AI revolution from the sidelines. The technology couldn't connect to the systems that mattered.
What MCP Actually Changes for Service Businesses
1. Real-Time Pricing Without Custom Integrations
MCP's core capability is dynamic tool discovery. When you expose your pricing engine as an MCP server, any MCP-compatible AI client (a chatbot on your website, Claude, ChatGPT, or a custom agent) can query it in real time. The AI doesn't need to know your pricing logic in advance. It discovers what tools are available, what inputs they need, and how to call them.
For a service provider, this means an AI assistant can ask a potential customer the right questions, pass those answers to your pricing model, and return a tailored quote, all within a single conversation. No forms. No "we'll get back to you in 24–48 hours." No manual intervention.
According to research from Gartner, 85% of enterprises are expected to have implemented AI agents by end of 2025. For service providers who haven't started, the window to be an early mover is closing fast.
2. Context That Travels With the Customer
One of MCP's most underappreciated features is contextual persistence. When a customer interacts with an MCP-enabled AI assistant, the context of that interaction (what they asked, what they were quoted, what options they explored) can travel across touchpoints.
A customer who starts a conversation on your website at 10pm can pick it up with a sales rep the next morning without repeating themselves. The AI has already gathered the relevant information, structured it, and made it available to your team. For service businesses where the sales process is consultative and multi-touch, this eliminates the friction that kills conversions.
3. Multi-Source Intelligence in One Conversation
Service pricing rarely depends on a single system. An insurance quote might pull from an underwriting model, a geographic risk database, a claims history system, and a regulatory compliance engine, all in one interaction.
MCP supports multi-server architectures natively. A single AI conversation can query multiple MCP servers simultaneously, aggregate the results, and present a coherent answer to the customer. Think of it as having your most knowledgeable employee, the one who knows every system, available 24/7 for every prospect.
4. No Vendor Lock-In
MCP is an open standard, now governed by the Linux Foundation, so adopting it doesn't tie you to Anthropic, OpenAI, Google, or any specific AI provider. Your MCP servers work with any compliant client. If you switch AI providers next year, your integrations still work.
For service businesses that have been burned by proprietary vendor lock-in before, this matters enormously.
The Service Provider MCP Playbook: Where to Start
If you're running a service business and considering MCP, here's a practical starting framework:
Step 1: Identify your "quotable moment." What is the single most common question prospects ask that requires a dynamic answer? For insurers, it's "how much will it cost me?" For agencies, it might be "what's included in your retainer?" That question is your first MCP server.
Step 2: Expose your pricing or scoping logic as an MCP server. This doesn't mean rebuilding your systems. MCP servers are lightweight wrappers around existing APIs or databases. If your pricing model already lives in an API, you can have an MCP server running in days, not months.
Step 3: Connect it to a client. Start with an AI assistant on your website or an internal tool your sales team uses. MCP clients are available for Claude, ChatGPT, and dozens of other platforms.
Step 4: Measure what changes. Track time-to-quote, conversion rates, and customer satisfaction. Early adopters in insurance and professional services are reporting quote delivery times dropping from days to seconds, and conversion rates rising as a result.
What Happens If Service Providers Ignore MCP
The risk of inaction is concrete. AI-powered search engines like ChatGPT, Google AI Overviews, Perplexity, and Claude are increasingly where customers start their buying journey. When someone asks "how much does home insurance cost in Madrid?" or "what does a marketing agency charge for SEO?", the AI that can provide a specific, real-time answer wins the interaction.
If your competitor's AI assistant can give a personalised quote in 30 seconds and yours says "fill out this form and we'll get back to you," the choice is obvious. MCP is what makes the former possible at scale.
The businesses that move now, while the protocol is still maturing and competition is thin, will have a structural advantage. They'll have trained their systems, refined their prompts, and built the muscle memory of AI-assisted selling before MCP becomes table stakes.
Where Waniwani Fits
At Waniwani, we build AI infrastructure for service businesses with dynamic pricing, with a particular focus on regulated industries where accuracy and compliance matter. The shift to MCP-native architecture aligns with what we've seen firsthand: service businesses need AI that connects to their actual business logic and can handle real, context-dependent conversations with customers. We'll be sharing more about how we're using MCP in our own stack in the coming months.
Frequently Asked Questions
What does MCP stand for?
MCP stands for Model Context Protocol. It is an open standard created by Anthropic in November 2024 that allows AI models to connect to external tools, databases, and live data sources through a single, standardised interface.
Is MCP only for developers?
No. While developers build MCP servers, the business impact is felt across sales, operations, and customer service. Service providers benefit from MCP through faster quoting, smarter AI assistants, and reduced manual work, none of which require the business team to write code.
Which AI platforms support MCP?
As of early 2026, MCP is supported by Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), Microsoft Copilot, Visual Studio Code, Cursor, and hundreds of other AI tools. Enterprise infrastructure support comes from AWS, Cloudflare, Google Cloud, and Microsoft Azure.
How is MCP different from a regular API?
A regular API is a point-to-point connection: one tool talking to one data source. MCP is a universal protocol that allows any AI client to discover and connect to any MCP server dynamically. A helpful analogy: APIs are like having a dedicated cable for each device, while MCP is a single USB-C port that works with everything.
How long does it take to implement MCP for a service business?
A basic MCP server wrapping an existing pricing API can be built in days. More complex implementations involving multiple data sources, authentication, and multi-step workflows typically take 2 to 6 weeks depending on the complexity of your existing systems.
Is MCP secure?
MCP supports authentication, access controls, and encrypted communication. Since December 2025, governance has been managed by the Agentic AI Foundation under the Linux Foundation, with security standards co-developed by Anthropic, OpenAI, Google, Microsoft, and AWS.