MCP: The Interface Between AI and Software

Artificial intelligence is already part of everyday business. Software vendors are increasingly competing on how effectively their products use AI, and users are becoming accustomed to interacting with software through questions and requests rather than menus and reports.

Model Context Protocol (MCP) provides a standard way for AI applications to work with existing software. An MCP server acts as a translator between the AI and the application, allowing the model to access information and use functions much as a user would.

The architecture is simple:

User

↓

AI Application

↓

MCP Server

↓

Business Application

And in most cases one can view the AI application as an additional user alongside the regular user.

This means AI applications can work with the systems businesses already use rather than requiring entirely new software. Customer records, invoices, projects, products and workflows become available to the model, allowing it to retrieve information, perform actions and help users accomplish tasks.

Whether the underlying application is an ERP system such as Odoo, a CRM platform, a document repository or a source code management system, the principle is the same. MCP provides a common language between AI and software, allowing organisations to benefit from advances in artificial intelligence while continuing to use the applications they already depend on.

What Does MCP Mean in Practice?

For most users, MCP is interesting because it allows them to get more from the systems they already use. Business applications such as ERP, CRM and accounting systems are often extremely capable, yet most organisations only use a fraction of what they have purchased. Advanced reporting, forecasting and analysis frequently require specialist knowledge or depend on a handful of experienced users.

MCP changes that. By allowing AI applications to work with business systems directly, users gain access to capabilities that previously required detailed knowledge of the application or hours spent collecting information and preparing reports.

Instead of navigating screens and menus, users can simply ask questions such as:

·          Which customers have not placed an order in six months?

·         Show me all invoices raised this month and estimate month-end revenue.

·         Forecast cash flow for the next quarter.

·         Identify customers showing signs of churn.

·         Which products are running low and need replenishment?

·         Summarise everything that happened with this customer during the past month.

·         Create a new opportunity in the CRM system for this customer.

·         Prepare quotations for the customers attending next week's trade show.

The result is that more people can benefit from the capabilities already contained within their existing business systems. Managers spend less time gathering information and more time using it. Reports that once required spreadsheets and manual effort can be produced in seconds and checked for accuracy rather than assembled from scratch.

This does not require every employee to become an expert user. Instead, AI helps make the capabilities of complex software more accessible. Organisations gain greater value from the systems they already own, while users spend less time wrestling with software and more time focusing on customers, decisions and the activities that make a difference.

Why MCP Matters Today

General-purpose AI applications like ChatGPT and Claude have already found their way into everyday business. People are using them to write documents, analyse information and answer questions. At the same time, organisations continue to rely on ERP systems, CRM platforms, accounting systems and collaboration tools that contain years of business information and established processes.

MCP matters because it brings these worlds together. Instead of treating AI as a separate tool, organisations can use it alongside the applications that already run the business. Questions become reports, reports become forecasts and information becomes action.

For businesses this coming together creates opportunities that were previously difficult or expensive. A sales manager can ask for a forecast for the entire team without spending hours collecting spreadsheets. A finance manager can investigate declining margins without manually assembling reports and the frequency of reporting this information to shareholders is less of a challenge. A customer service manager can identify trends and recurring issues across thousands of interactions. Users spend less time preparing information and more time making decisions, putting the information into action.

For many organisations, this represents a practical opportunity to benefit from AI without replacing existing systems or embarking on large transformation projects. Businesses can begin experimenting immediately and incorporate AI into their software strategy.

Software vendors have recognised the same opportunity. Applications such as Odoo have rapidly expanded their own AI capabilities, and similar developments are taking place across CRM, productivity and collaboration platforms. Intelligence is becoming part of the value proposition of software itself.

This creates an interesting transition period. MCP provides immediate access to the capabilities of today's general-purpose AI while software vendors develop increasingly capable, application-specific intelligence. 

Some applications will integrate AI in useful ways, while others will use it to create genuinely transformative experiences for their users.

For organisations, this means MCP represents an opportunity that exists today. It allows businesses to explore how AI can improve reporting, forecasting and decision-making while gaining a glimpse of where business software is heading. Over time, many of these capabilities are likely to become increasingly embedded within the applications themselves, making intelligence a natural part of the overall software experience.

The New Competitive Advantage

The rapid adoption of AI inside business applications reflects something more important than technology. Software providers are competing to give their customers a market advantage.

Smaller organisations can increasingly compete with larger businesses by using intelligent systems to achieve more with fewer resources. Activities that once required specialist skills, analysts or dedicated teams can now be performed in seconds. Sales forecasts, revenue projections, stock planning and customer analysis become available at the push of a button.

Intelligent systems amplify people. 

Sales managers spend less time preparing reports and more time coaching their teams. Finance managers spend less time assembling information and more time making decisions. Operations managers spend less time reconciling spreadsheets and more time improving delivery. People focus more attention on customers, opportunities and decisions that move the business forward, while software takes on a growing share of analysis, recommendations and routine work.

This changes the economics of scale. Ten people can achieve the output that previously might have required twenty. Growing businesses gain access to capabilities that were once reserved for larger organisations with specialist departments and greater resources. AI becomes another way for organisations to multiply the effectiveness of their people.

That means that there's a new competitive battleground for software providers. Features, workflows and user interfaces still matter, but increasingly the question customers ask is simpler:

"How much difference does this software make to my business?"

The answer to that question determines who wins.

For companies like Odoo, HubSpot, Microsoft and Salesforce, intelligence becomes both an opportunity and a necessity. If another platform can help customers forecast better, serve customers better and operate more effectively, it creates an opportunity for competitors to displace existing systems. AI therefore becomes both an offensive capability and a defensive one.

That's a set of strong incentives for software vendors to keep intelligence close to the application itself. The software already understands its own workflows, terminology and business processes. It already knows who the users are and what information they should have access to. Governance, support, accountability and commercial relationships are already in place. Applications have built in guardrails, and know what should or shouldn't be editable in the database. Delivering AI as part of the application experience reduces barriers to adoption and gives customers a simpler, more trusted way to benefit from intelligent systems.

This relationship, internal vs external 3rd party, plays a role in defining the relationship between AI providers and software vendors. Customers increasingly experience the benefits of AI through the applications they already use, while the underlying models become part of the platform rather than a separate destination.

The result is a new generation of business applications in which intelligence becomes another source of competitive advantage and software increasingly helps organisations accomplish more with the resources they already have.

What This Means for Users and Specialists

For organisations looking at AI today, MCP represents an opportunity. It provides a practical way to experiment, learn and begin extracting more value from the systems already running the business. Reports become easier to generate, information becomes easier to access and routine tasks become easier to automate. Businesses can start small, learn quickly and incorporate AI into their software strategy without waiting for the next generation of applications to arrive.

At the same time, software vendors are investing heavily in AI and competing to make intelligence part of the value they deliver to customers. Increasingly, users are likely to experience AI through the applications they already use. ERP systems will understand finance, inventory and manufacturing. CRM platforms will understand pipelines and customer engagement. Collaboration platforms will understand communication and teamwork. Intelligence becomes increasingly specialised around the problems each application exists to solve.

For many organisations, this creates a more natural way to consume AI. Governance, support and accountability remain part of existing commercial relationships. Businesses gain confidence in how information is handled, and users benefit from intelligent capabilities without needing to become experts in AI or adopt additional tools and services.

This does not diminish the importance of MCP. On the contrary, it accelerates the transition and provides immediate value. Organisations operating across multiple systems benefit from AI that can work wherever information resides. Consultants can accelerate implementations and configurations. Developers can understand complex systems and build integrations more quickly. System integrators gain a common way to connect intelligence with business processes.

As intelligence becomes increasingly embedded into applications, the role of MCP naturally becomes more specialised. Most users will care far more about outcomes than the mechanism delivering them. They will judge software by the difference it makes to the business, not by whether the intelligence comes from an external AI application or from capabilities built directly into the platform.

For specialists, however, MCP remains a powerful tool. Much as APIs became essential infrastructure for connecting applications, MCP provides a common language for connecting intelligence with systems. Its role evolves from something visible and exciting to something increasingly foundational.

The result is a software ecosystem in which intelligence becomes woven into the fabric of business applications, completing the evolution from software that people operate to software that people collaborate with.