The Architect’s Dilemma – O’Reilly

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

The widespread recommendation is usually 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 steering on which is best for a particular journey. This lack of a transparent psychological mannequin results in architectural guesswork. Groups construct complicated conversational interfaces for duties that demand inflexible predictability, or they expose inflexible APIs to customers who desperately want steering. The end result is usually the identical: a system that appears nice in demos however falls aside in the true 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 wanting outward and asking a single, basic query: Who is looking 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 fitting interface based mostly in your client’s wants. I may also discover failure modes, testing, and the highly effective Gatekeeper Sample that exhibits how these two interfaces can work collectively to create programs that aren’t simply intelligent however really 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 desires, and supply a whole payload. The device validates, executes as soon as, and returns a consequence. 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 (“e-book a refundable flight beneath $450”), not an argument listing. The agent asks clarifying questions, calls instruments as wanted, and holds onto session state till the job is finished. When you image a concierge—interacting, negotiating trade-offs, and infrequently escalating—you’re in the fitting neighborhood.

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

• MCP is quick to motive about, straightforward to check, and robust on determinism and auditability.
• A2A is constructed for ambiguity, long-running processes, and choice seize.

Bringing the Interfaces to Life: A Reserving Instance

To see the distinction in follow, let’s think about a easy activity: 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 mandatory data in a single, legitimate payload:

{
  "jsonrpc": "2.0",
  "id": 42,
  "methodology": "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 area or incorrect knowledge sort—leads to a direct 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 purpose. It makes use of dialog to resolve ambiguity:

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

The agent’s job is to take the ambiguous purpose, collect the mandatory particulars, after which possible name the MCP device 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 a very powerful architectural query of all: Who is looking your service?

Step 1: Establish 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 performing 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 approach 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 desire 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 device (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 satisfy a posh, high-level purpose? This client values comfort and the offloading of cognitive load. They don’t need to specify each step of a course of; they need to delegate possession of a purpose and belief that will probably be dealt with. They’re snug with ambiguity as a result of they anticipate the service—the agent—to resolve it on their behalf. This client doesn’t need to comply with 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 buyer, the architect’s dilemma typically evaporates. Earlier than you ever debate statefulness or determinism, you first outline the consumer expertise you might be obligated to supply. Generally, figuring out your buyer gives you your definitive reply.

Step 2: Validate with the 4 components

After you have recognized who calls your service, you’ve gotten a powerful speculation to your design. A machine client factors to a device; a human or agentic client factors to an agent. The subsequent step is to validate this speculation with a technical litmus check. This framework provides 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 objectives? 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 anticipated 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 complicated 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. Your entire operation—from fee 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 best way. This requires the asynchronous, stateful nature of A2A, which might 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 keep in mind 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 main points 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 keep up context.
4. Direct management versus delegated possession
   Is the buyer orchestrating each step, or are they delegating all the purpose? When utilizing a merchandising machine, the buyer is in direct management. You’re the orchestrator, deciding which button to press and when. With MCP, the calling software retains full management, making a collection of exact perform calls to attain its personal purpose. With a concierge, you delegate possession. You hand over the high-level purpose and belief the agent to handle the main points. That is the core mannequin of A2A, the place the buyer offloads the cognitive load and trusts the agent to ship the end result.

Desk 1: 4 query framework

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

No framework, regardless of how clear…

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

Let’s discover two widespread edge instances.

The complicated device: The iceberg
Take into account a service that performs a extremely complicated, 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 an enormous, long-running workflow involving a number of machines, high quality checks, and encoding steps. It’s a vastly complicated course of.

Nonetheless, from the buyer’s perspective, none of that issues. They made a single, stateless, fire-and-forget name. They don’t have to handle the method; they simply want a predictable consequence. 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 ought to be, carried out as a device (MCP).

The straightforward agent: The scripted dialog
Now contemplate 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 interior logic is likely to be a easy state machine: ask for a date, then a time, then a affected person identify. It’s not “clever” or significantly versatile.

Nonetheless, it should bear in mind the consumer’s earlier solutions to finish the reserving. It’s an inherently stateful, multiturn interplay. The patron can not present all of the required data in a single, prevalidated name. They must be guided by way of the method. Regardless of its inner simplicity, the necessity for a stateful dialogue makes it a concierge. It have to be carried out 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. Concentrate on the expertise it supplies externally. That contract together with your buyer is the last 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 essentially completely different failure modes and require completely different testing methods.

Testing MCP instruments (merchandising machines):

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

Testing A2A brokers (concierges):

• Aim completion charge (GCR): Measure the share of conversations the place the agent efficiently achieved the consumer’s high-level purpose.
• Conversational effectivity: Observe the variety of turns or clarifications required to finish a activity.
• Instrument choice accuracy: For complicated brokers, confirm that the fitting MCP device 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 to date has centered on a dichotomy: MCP or A2A, merchandising machine or concierge. However probably the most subtle and strong agentic programs don’t power 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 method to obtain that is by way of a robust architectural selection we will 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 complicated, messy work of understanding a high-level purpose, 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.

Take into account a journey agent. A consumer interacts with it through A2A, giving it a high-level purpose: “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, finances, 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 device 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 supplies a number of important architectural advantages:

• Decoupling: It separates the complicated, conversational logic (the “how”) from the straightforward, 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 will probably deal with authentication, implement charge limits, handle consumer quotas, and log all exercise earlier than a single device is ever invoked.
• Simplified device 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 nicely, 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 novel hint ID. This ID have to be propagated to each downstream MCP device name, permitting you to hint a single consumer request throughout all the system. Structured logs for device 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 charge limits and utilization quotas, and may 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 device, it ought to implement patterns like timeouts, retries with exponential backoff, and circuit breakers. Critically, it’s accountable for the ultimate failure state—escalating to a human within the loop for assessment or clearly speaking the difficulty to the top consumer.

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

Conclusion

We started this journey with a easy however irritating downside: the architect’s dilemma. Confronted with the round recommendation that “MCP is for instruments and A2A is for brokers,” we had 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 purpose 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 together with your buyer. Probably the most vital query shouldn’t 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 components. Use the litmus check of determinism, course of, state, and possession to technically justify and solidify your selection.

Finally, probably the most strong programs 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 now not a dilemma. By specializing in the buyer’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 additionally intuitive, scalable, and maintainable.