simpleAI Engine Our Vision for Your Shopping Experience

No searching through catalogs. No calling the purchasing department. No orders placed outside the system. Not just because a policy prohibits it, but because, for the first time, the right way is also the easiest.

This is exactly the vision we’re building. And we’ve already gotten started.

This post describes where we’re taking simple system over the next 18 months. We call this goal the simpleAI Engine. And we’re showing you the first building block. It’s live as of today.

Who will do the procurement work in the future?

For 26 years, we’ve been simplifying indirect procurement for small and medium-sized businesses. For a long time, that meant making work easier for people. With AI, something even more fundamental is changing. It’s changing who does the work. It’s no longer just the person clicking through the system. Instead, it’s agents working in the background.

But agents don’t click through menus. They read data, access interfaces, and take action. Software built solely for humans remains invisible to agents. That’s why we’re consistently developing simple system in both directions: as a platform for humans and as infrastructure for agents.

From simpleAI to the simpleAI Engine

Many of our customers are already working with simpleAI today. simpleAI turns free-form text into structured queries, parses quotes in any format, assigns items to the correct product category, suggests suitable suppliers, and keeps processes up to date via follow-up documents. Each of these functions solves a task that, until recently, had to be done manually. None of them is a chatbot. All are deeply integrated into the processes.

The simpleAI Engine is the overarching goal behind it all: you build your own custom agents. All data and all interfaces interlock and handle every request all the way through to the order. Just like in an engine room , no part works on its own. Every search, every order, and every connected system makes the engine a little better. And the better it gets, the more requests flow through it voluntarily—because it’s the easiest way.

The AI model doesn’t decide the outcome.
Your data is what decides.

A general language model knows half the internet. But it doesn’t know your purchasing process. It doesn’t know what a normal delivery time is for you. It doesn’t know which supplier delivers reliably. And it doesn’t know which rule should stop an order. It is precisely this knowledge that matters in purchasing. It comes from real transactions, not from a larger model.

That’s why we combine both worlds. You use the AI model of your choice, with your data and your guidelines. Through the simple system MCP, you connect it to your purchasing rules and your product assortment in simple system. Users choose the approach that works best for them. Administrators can analyze data wherever they want. And everyone benefits from the strengths of both worlds.

Two sentences describe where we’re headed:
simple system brings AI to procurement.
simple system prepares your purchasing for any AI.

The first building block is here: the simple system MCP

There are many visions. That’s why we’re not starting with a promise, but with a product. As of today, the first version of the simple system MCP is available.

MCP stands for Model Context Protocol, an open standard that allows AI assistants to access tools and data sources. Simply put: The MCP is the outlet to which you connect your AI.

Starting now, your AI assistant works directly within simple system. It searches your shared product range across all connected suppliers and puts together shopping carts. You describe your needs in your AI environment; the rest happens in simple system—in the background. Users log in with their existing credentials. No IT project is required.

The key point here: Your rules still apply. Product range restrictions, price limits, and approval processes work exactly as they would if a human had clicked. The agent doesn’t get a master key. It gets a work pass.

Administrators have access to settings, can view the configuration, analyze reports, and create users.

Approvers are presented with a list of shopping carts requiring approval and can check for anomalies and assess whether the orders deviate from the order history.

Buyers can search for recurring orders, query product ranges, and view RFQs.

The MCP is intentionally just the beginning. It demonstrates what agent-based procurement feels like at simple system: open access, strict rules.

You decide how much of the user interface you want

In the future, simple system will support all stages of procurement. You can run simple system in headless mode and integrate it into your own AI ecosystem via open interfaces like the MCP—without using our user interface at all. Or you can continue to use the user interface that over 28,000 buyers rely on today—with a clear AI-centric focus for maximum efficiency in the future. From the full user interface to headless procurement: You decide how much automation you want. We deliver both.

The Vision for 2027

Every Search Yields a Match

No matter what’s being searched for: simple system takes care of the rest. Here’s an example.
A requester is looking for a specialty part. A generic model makes a suggestion and displays something. The simpleAI Engine knows better. It finds the part in the existing product portfolio. And if it’s not available there, it shows where it can be procured. It detects when a rule is blocking the order and forwards the information to the right place. It also finds the part in the product range of a supplier who is not yet known and points out suppliers who carry such items or have already delivered them. And if there’s nothing off the shelf, it initiates special procurement or a service request.

This turns every search into a path to procurement. The fruitless search—currently the most common reason for placing orders outside the system—disappears.

Product Assortments That Proactively Suggest Optimization

We don’t just show you where to source which items. We show you how to improve. Where can product ranges be consolidated? How can you safeguard procurement against shortages? Where is there money to be saved? Your data is transformed into concrete suggestions that make your purchasing cheaper, more reliable, and easier. Not as an annual report, but as ongoing recommendations from the system. Your product range alerts you when it identifies opportunities for optimization. It’s all in your hands, with your agents and settings in your AI tool.

A Centralized Database for the AI Era

Strong AI in procurement requires a strong data foundation. AI without your procurement data is like a new colleague without training: hardworking but clueless. That’s why centralization is the way forward. Route all orders through simple system—from the ERP, from vending machines, from external systems, and from your users. This creates a central database for the AI era.

That’s also why this decision can’t be put off. The data your AI will need in 2027 is being generated right now. Every order that bypasses the system today will be missing as knowledge tomorrow. If you want AI in 2027 that understands your own purchasing, you’ll need to lay the foundation for your data in 2026.
So talk to our consultants today about how we can process all orders through simple system to prepare you for your entry into the world of AI.

What’s Changing and for Whom

For requesters, ordering will become as easy as asking a question. Describe instead of searching; approve instead of following up with a phone call.

For procurement management, the focus shifts from processing to oversight. The engine handles searching, requesting, and recording. Procurement defines rules, negotiates terms, and decides on cases that truly require judgment.

For IT, the use of open standards like MCP means no new integration project for every AI idea. One platform, clear interfaces, and centralized permissions.

Three principles guide us

Especially when agents read data and act on behalf of our customers, clear principles are essential. Ours are these:

Customer benefit. Always and everywhere. You should choose the best agents and tools to help you. We build our own first-class agents. Sometimes the ecosystem provides the best agent for a specific use case. That’s a good thing.

Open and neutral. simple system is neutral and does not engage in proprietary trading. Whatever you can do in simple system, you should also be able to do via interfaces. You can access our data and our intelligence wherever you work: in simple system or in your own AI environment.

Trust and control. AI in the engine room, not in the showcase. Human in the Loop, AI Supported. When simpleAI acts on your behalf, you can see at any time what’s permitted and what’s happening. You determine the level of autonomy: from approving every single transaction to agents you connect via MCP. Your rules and approval processes apply at every stage. No AI can bypass them.

What’s already possible today and what’s coming by 2027

Already a reality today: the simpleAI features. They turn free-form text into structured requests, parse quotes in any format, assign items to the correct product category, suggest suitable suppliers, and keep processes up to date via follow-up documents. Already a reality today: the ability to manage all orders centrally via simple system. And available starting today: the first simple system MCP.

From here, we’ll gradually unlock more capabilities, step by step, until we reach full autonomy. You build agents and tools according to your requirements, processes, and guidelines. We provide the infrastructure. You define the AI.

The Opportunity

The shift toward AI has been underway for some time. It’s evident in every procurement department that’s asking where the work is headed in the future. The best AI will be the one that understands a company’s procurement process just like an experienced buyer: what’s normal, what’s working well, what’s changing, and what has worked for comparable companies. This is precisely the knowledge that simple system has been building for 26 years—with 28,000+ buyers, 1,000+ suppliers, and over 26 years of transaction data.

We don’t replicate every use case. We build the engine that can support every use case.

The first building block is right in front of you. The data foundation for your 2027 AI is being built from today’s orders. Let’s get started together.