The Architect’s Dilemma – O’Reilly

The agentic AI panorama is exploding. Each new framework, demo, and announcement guarantees to let your AI assistant ebook flights, question databases, and handle calendars. This speedy development of capabilities is thrilling for customers, however for the architects and engineers constructing these methods, it poses a basic query: When ought to a brand new functionality be a easy, predictable software (uncovered through the Mannequin Context Protocol, MCP) and when ought to it’s a classy, collaborative agent (uncovered through the Agent2Agent Protocol, A2A)?

The frequent recommendation is commonly round and unhelpful: “Use MCP for instruments and A2A for brokers.” That is like telling a traveler that automobiles use motorways and trains use tracks, with out providing any steerage on which is healthier for a particular journey. This lack of a transparent psychological mannequin results in architectural guesswork. Groups construct advanced conversational interfaces for duties that demand inflexible predictability, or they expose inflexible APIs to customers who desperately want steerage. The end result is commonly the identical: a system that appears nice in demos however falls aside in the actual world.

On this article, I argue that the reply isn’t discovered by analyzing your service’s inner logic or know-how stack. It’s discovered by trying outward and asking a single, basic query: Who is asking your product/service? By reframing the issue this manner—as a consumer expertise problem first and a technical one second—the architect’s dilemma evaporates.

This essay attracts a line the place it issues for architects: the road between MCP instruments and A2A brokers. I’ll introduce a transparent framework, constructed across the “Merchandising Machine Versus Concierge” mannequin, that will help you select the correct interface based mostly in your client’s wants. I will even discover failure modes, testing, and the highly effective Gatekeeper Sample that exhibits how these two interfaces can work collectively to create methods that aren’t simply intelligent however actually dependable.

Two Very Completely different Interfaces

MCP presents instruments—named operations with declared inputs and outputs. The caller (an individual, program, or agent) should already know what it needs, and supply a whole payload. The software validates, executes as soon as, and returns a outcome. In case your psychological picture is a merchandising machine—insert a well-formed request, get a deterministic response—you’re shut sufficient.

A2A presents brokers—goal-first collaborators that converse, plan, and act throughout turns. The caller expresses an end result (“ebook a refundable flight below $450”), not an argument record. The agent asks clarifying questions, calls instruments as wanted, and holds onto session state till the job is completed. If you happen to image a concierge—interacting, negotiating trade-offs, and infrequently escalating—you’re in the correct neighborhood.

Neither interface is “higher.” They’re optimized for various conditions:

• MCP is quick to purpose about, straightforward to check, and powerful on determinism and auditability.
• A2A is constructed for ambiguity, long-running processes, and desire seize.

Bringing the Interfaces to Life: A Reserving Instance

To see the distinction in observe, let’s think about a easy job: reserving a particular assembly room in an workplace.

The MCP “merchandising machine” expects a wonderfully structured, machine-readable request for its book_room_tool. The caller should present all vital data in a single, legitimate payload:

{
  "jsonrpc": "2.0",
  "id": 42,
  "technique": "instruments/name",
  "params": {
    "identify": "book_room_tool",
    "arguments": {
      "room_id": "CR-104B",
      "start_time": "2025-11-05T14:00:00Z",
      "end_time": "2025-11-05T15:00:00Z",
      "organizer": "consumer@instance.com"
    }
  }
}

Any deviation—a lacking discipline or incorrect information kind—leads to a right away error. That is the merchandising machine: You present the precise code of the merchandise you need (e.g., “D4”) otherwise you get nothing.

The A2A “concierge,“ an “workplace assistant” agent, is approached with a high-level, ambiguous aim. It makes use of dialog to resolve ambiguity:

Person: “Hey, are you able to ebook a room for my 1-on-1 with Alex tomorrow afternoon?”
Agent: “In fact. To verify I get the correct one, what time works finest, and the way lengthy will you want it for?”

The agent’s job is to take the ambiguous aim, collect the mandatory particulars, after which seemingly name the MCP software behind the scenes as soon as it has a whole, legitimate set of arguments.

With this clear dichotomy established—the predictable merchandising machine (MCP) versus the stateful concierge (A2A)—how can we select? As I argued within the introduction, the reply isn’t present in your tech stack. It’s discovered by asking an important architectural query of all: Who is asking your service?

Step 1: Determine your client

1. The machine client: A necessity for predictability
   Is your service going to be referred to as by one other automated system, a script, or one other agent appearing in a purely deterministic capability? This client requires absolute predictability. It wants a inflexible, unambiguous contract that may be scripted and relied upon to behave the identical method each single time. It can not deal with a clarifying query or an sudden replace; any deviation from the strict contract is a failure. This client doesn’t need a dialog; it wants a merchandising machine. This nonnegotiable requirement for a predictable, stateless, and transactional interface factors on to designing your service as a software (MCP).
2. The human (or agentic) client: A necessity for comfort
   Is your service being constructed for a human finish consumer or for a classy AI that’s attempting to meet a posh, high-level aim? This client values comfort and the offloading of cognitive load. They don’t wish to specify each step of a course of; they wish to delegate possession of a aim and belief that it will likely be dealt with. They’re snug with ambiguity as a result of they count on the service—the agent—to resolve it on their behalf. This client doesn’t wish to observe a inflexible script; they want a concierge. This requirement for a stateful, goal-oriented, and conversational interface factors on to designing your service as an agent (A2A).

By beginning with the patron, the architect’s dilemma typically evaporates. Earlier than you ever debate statefulness or determinism, you first outline the consumer expertise you’re obligated to supply. Generally, figuring out your buyer will provide you with your definitive reply.

Step 2: Validate with the 4 elements

After getting recognized who calls your service, you will have a robust speculation to your design. A machine client factors to a software; a human or agentic client factors to an agent. The following step is to validate this speculation with a technical litmus take a look at. This framework offers you the vocabulary to justify your selection and make sure the underlying structure matches the consumer expertise you propose to create.

1. Determinism versus ambiguity
   Does your service require a exact, unambiguous enter, or is it designed to interpret and resolve ambiguous targets? A merchandising machine is deterministic. Its API is inflexible: GET /merchandise/D4. Every other request is an error. That is the world of MCP, the place a strict schema ensures predictable interactions. A concierge handles ambiguity. “Discover me a pleasant place for dinner” is a legitimate request that the agent is predicted to make clear and execute. That is the world of A2A, the place a conversational movement permits for clarification and negotiation.
2. Easy execution versus advanced course of
   Is the interplay a single, one-shot execution, or a long-running, multistep course of? A merchandising machine performs a short-lived execution. The whole operation—from cost to allotting—is an atomic transaction that’s over in seconds. This aligns with the synchronous-style, one-shot mannequin of MCP. A concierge manages a course of. Reserving a full journey itinerary may take hours and even days, with a number of updates alongside the way in which. This requires the asynchronous, stateful nature of A2A, which may deal with long-running duties gracefully.
3. Stateless versus stateful
   Does every request stand alone or does the service want to recollect the context of earlier interactions? A merchandising machine is stateless. It doesn’t do not forget that to procure a sweet bar 5 minutes in the past. Every transaction is a clean slate. MCP is designed for these self-contained, stateless calls. A concierge is stateful. It remembers your preferences, the small print of your ongoing request, and the historical past of your dialog. A2A is constructed for this, utilizing ideas like a session or thread ID to take care of context.
4. Direct management versus delegated possession
   Is the patron orchestrating each step, or are they delegating the whole aim? When utilizing a merchandising machine, the patron is in direct management. You’re the orchestrator, deciding which button to press and when. With MCP, the calling utility retains full management, making a sequence of exact operate calls to attain its personal aim. With a concierge, you delegate possession. You hand over the high-level aim and belief the agent to handle the small print. That is the core mannequin of A2A, the place the patron offloads the cognitive load and trusts the agent to ship the end result.

Desk 1: 4 query framework

These elements will not be impartial checkboxes; they’re 4 sides of the identical core precept. A service that’s deterministic, transactional, stateless, and immediately managed is a software. A service that handles ambiguity, manages a course of, maintains state, and takes possession is an agent. Through the use of this framework, you possibly can confidently validate that the technical structure of your service aligns completely with the wants of your buyer.

No framework, irrespective of how clear…

…can completely seize the messiness of the actual world. Whereas the “Merchandising Machine Versus Concierge” mannequin offers a strong information, architects will ultimately encounter providers that appear to blur the strains. The secret is to recollect the core precept we’ve established: The selection is dictated by the patron’s expertise, not the service’s inner complexity.

Let’s discover two frequent edge instances.

The advanced software: The iceberg
Think about a service that performs a extremely advanced, multistep inner course of, like a video transcoding API. A client sends a video file and a desired output format. It is a easy, predictable request. However internally, this one name may kick off a large, long-running workflow involving a number of machines, high quality checks, and encoding steps. It’s a vastly advanced course of.

Nevertheless, from the patron’s perspective, none of that issues. They made a single, stateless, fire-and-forget name. They don’t must handle the method; they simply want a predictable outcome. This service is like an iceberg: 90% of its complexity is hidden beneath the floor. However as a result of its exterior contract is that of a merchandising machine—a easy, deterministic, one-shot transaction—it’s, and needs to be, applied as a software (MCP).

The easy agent: The scripted dialog
Now think about the alternative: a service with quite simple inner logic that also requires a conversational interface. Think about a chatbot for reserving a dentist appointment. The inner logic is perhaps a easy state machine: ask for a date, then a time, then a affected person identify. It’s not “clever” or significantly versatile.

Nevertheless, it should bear in mind the consumer’s earlier solutions to finish the reserving. It’s an inherently stateful, multiturn interplay. The buyer can not present all of the required data in a single, prevalidated name. They should be guided by the method. Regardless of its inner simplicity, the necessity for a stateful dialogue makes it a concierge. It have to be applied as an agent (A2A) as a result of its consumer-facing expertise is that of a dialog, nevertheless scripted.

These grey areas reinforce the framework’s central lesson. Don’t get distracted by what your service does internally. Deal with the expertise it offers externally. That contract along with your buyer is the final word arbiter within the architect’s dilemma.

Testing What Issues: Completely different Methods for Completely different Interfaces

A service’s interface doesn’t simply dictate its design; it dictates the way you validate its correctness. Merchandising machines and concierges have basically totally different failure modes and require totally different testing methods.

Testing MCP instruments (merchandising machines):

• Contract testing: Validate that inputs and outputs strictly adhere to the outlined schema.
• Idempotency assessments: Be certain that calling the software a number of instances with the identical inputs produces the identical outcome with out unintended effects.
• Deterministic logic assessments: Use commonplace unit and integration assessments with mounted inputs and anticipated outputs.
• Adversarial fuzzing: Take a look at for safety vulnerabilities by offering malformed or sudden arguments.

Testing A2A brokers (concierges):

• Aim completion fee (GCR): Measure the proportion of conversations the place the agent efficiently achieved the consumer’s high-level aim.
• Conversational effectivity: Monitor the variety of turns or clarifications required to finish a job.
• Software choice accuracy: For advanced brokers, confirm that the correct MCP software was chosen for a given consumer request.
• Dialog replay testing: Use logs of actual consumer interactions as a regression suite to make sure updates don’t break current conversational flows.

The Gatekeeper Sample

Our journey up to now has targeted on a dichotomy: MCP or A2A, merchandising machine or concierge. However essentially the most subtle and strong agentic methods don’t pressure a selection. As an alternative, they acknowledge that these two protocols don’t compete with one another; they complement one another. The last word energy lies in utilizing them collectively, with every enjoying to its strengths.

The best option to obtain that is by a strong architectural selection we are able to name the Gatekeeper Sample.

On this sample, a single, stateful A2A agent acts as the first, user-facing entry level—the concierge. Behind this gatekeeper sits a group of discrete, stateless MCP instruments—the merchandising machines. The A2A agent takes on the advanced, messy work of understanding a high-level aim, managing the dialog, and sustaining state. It then acts as an clever orchestrator, making exact, one-shot calls to the suitable MCP instruments to execute particular duties.

Think about a journey agent. A consumer interacts with it through A2A, giving it a high-level aim: “Plan a enterprise journey to London for subsequent week.”

• The journey agent (A2A) accepts this ambiguous request and begins a dialog to assemble particulars (precise dates, price range, and so on.).
• As soon as it has the mandatory data, it calls a flight_search_tool (MCP) with exact arguments like origin, vacation spot, and date.
• It then calls a hotel_booking_tool (MCP) with the required metropolis, check_in_date, and room_type.
• Lastly, it’d name a currency_converter_tool (MCP) to supply expense estimates.

Every software is an easy, dependable, and stateless merchandising machine. The A2A agent is the good concierge that is aware of which buttons to press and in what order. This sample offers a number of important architectural advantages:

• Decoupling: It separates the advanced, conversational logic (the “how”) from the easy, reusable enterprise logic (the “what”). The instruments may be developed, examined, and maintained independently.
• Centralized governance: The A2A gatekeeper is the proper place to implement cross-cutting issues. It might deal with authentication, implement fee limits, handle consumer quotas, and log all exercise earlier than a single software is ever invoked.
• Simplified software design: As a result of the instruments are simply easy MCP features, they don’t want to fret about state or conversational context. Their job is to do one factor and do it properly, making them extremely strong.

Making the Gatekeeper Manufacturing-Prepared

Past its design advantages, the Gatekeeper Sample is the best place to implement the operational guardrails required to run a dependable agentic system in manufacturing.

• Observability: Every A2A dialog generates a singular hint ID. This ID have to be propagated to each downstream MCP software name, permitting you to hint a single consumer request throughout the whole system. Structured logs for software inputs and outputs (with PII redacted) are vital for debugging.
• Guardrails and safety: The A2A Gatekeeper acts as a single level of enforcement for vital insurance policies. It handles authentication and authorization for the consumer, enforces fee limits and utilization quotas, and might preserve a listing of which instruments a specific consumer or group is allowed to name.
• Resilience and fallbacks: The Gatekeeper should gracefully handle failure. When it calls an MCP software, it ought to implement patterns like timeouts, retries with exponential backoff, and circuit breakers. Critically, it’s liable for the ultimate failure state—escalating to a human within the loop for evaluation or clearly speaking the difficulty to the tip consumer.

The Gatekeeper Sample is the final word synthesis of our framework. It makes use of A2A for what it does finest—managing a stateful, goal-oriented course of—and MCP for what it was designed for—the dependable, deterministic execution of a job.

Conclusion

We started this journey with a easy however irritating drawback: the architect’s dilemma. Confronted with the round recommendation that “MCP is for instruments and A2A is for brokers,” we have been left in the identical place as a traveler attempting to get to Edinburgh—figuring out that automobiles use motorways and trains use tracks however with no instinct on which to decide on for our particular journey.

The aim was to construct that instinct. We did this not by accepting summary labels, however by reasoning from first ideas. We dissected the protocols themselves, revealing how their core mechanics inevitably result in two distinct service profiles: the predictable, one-shot “merchandising machine” and the stateful, conversational “concierge.”

With that basis, we established a transparent, two-step framework for a assured design selection:

1. Begin along with your buyer. Probably the most vital query will not be a technical one however an experiential one. A machine client wants the predictability of a merchandising machine (MCP). A human or agentic client wants the comfort of a concierge (A2A).
2. Validate with the 4 elements. Use the litmus take a look at of determinism, course of, state, and possession to technically justify and solidify your selection.

In the end, essentially the most strong methods will synthesize each, utilizing the Gatekeeper Sample to mix the strengths of a user-facing A2A agent with a set of dependable MCP instruments.

The selection is not a dilemma. By specializing in the patron’s wants and understanding the basic nature of the protocols, architects can transfer from confusion to confidence, designing agentic ecosystems that aren’t simply practical but in addition intuitive, scalable, and maintainable.