Agent, workflow or automation: what you are buying

The three words name three places a decision can live

An AI agent, an AI workflow and an automation are not sizes of the same thing; they differ in where the next-step decision lives. In an automation, every step is decided in advance by a rule: when X happens, do Y, identically every time. In an AI workflow, the path is fixed but individual steps use a model to handle input a rule cannot - reading a document, drafting a reply, classifying a message - with the output flowing back into the fixed path. In an agent, the model itself chooses the next step: it is given a goal and a set of tools, and decides at run time what to do, in what order, and when it is finished. One disclosure is owed before comparing them: Foxnut Studios builds these systems for clients, so it is a seller in this category; the page’s defence is that no row below favours the most expensive option.

The comparison that matters is where mistakes happen

The table compares the three on the axes a buyer actually pays for later: who decides, what it costs to run, how it fails, and what it takes to test and hand over. The vocabulary matters less than these rows; buy the row behaviour, not the label.

Read bottom-up, the table is a cost-of-being-wrong ladder. Each step right adds capability and subtracts predictability, and the price of the added capability is paid in testing and oversight, not in the licence fee.

When to choose each

Each category wins real situations. The failure modes below are the ones practitioners report, not hypotheticals.

Choose an automation when the rule is stable

If a person can write the rule and the inputs do not drift, an automation is the correct buy and the cheapest thing in this comparison to own. Its failure mode is brittleness at the edges: the rule silently stops matching reality, and nobody assigned an owner to notice. Most of the reliable value in most businesses still lives here, which no one selling the shinier categories is paid to say.

Choose an AI workflow when the steps need judgment but the path does not

Most briefs that arrive saying “agent” resolve here: the process is known and repeatable, and what was missing was a way to handle fuzzy input inside it - the kind of somewhat-repeatable judgment work that previously needed a person. This is where the practitioner corpus reports its plainest wins, in the register this territory prefers: a measured “11 times faster” on a defined process, not a transformation story. The failure mode is silent step-level degradation: the path completes, the output is worse, and only an eval set can tell you - which is why the eval set, not the prompt file, is the artefact to insist on.

Choose an agent when the path genuinely cannot be fixed

When the space of situations is too varied to enumerate - the system must decide what to look at next based on what it just found - an agent is the honest fit, and the only category that can do the job. It is also the minority case: practitioners who build these systems for a living report that most briefs asking for agents do not need them. The failure mode is compounding: an open-ended cost shape plus path-level unpredictability plus the hardest testing burden, all inherited on day one, whether or not the problem required them.

What this comparison usually gets wrong

It is argued as a maturity ladder, with the agent as the destination and everything else as a stepping stone. That framing is a sales artefact: the categories are tools with different failure economics, not stages of enlightenment, and “which is most advanced” is a vendor’s question where “where does a wrong decision cost me least” is the buyer’s. The surrounding vocabulary makes this worse - the adjectives sellers attach to “agent” have no settled definition, and buyers in the corpus now read them as a warning sign rather than a credential, to the point where the word itself has become a purchase blocker. The practical reading of the table is direction, not destination: start at the most predictable category that solves the problem, and step right only when a measured limitation - not a label - forces it. Whichever category is built, insist on the artefact list a complete AI handoff carries, because the handover burden in the table’s last row is where the territory on what a consultant leaves behind says the real price of each category shows up. Foxnut Studios’ own record runs in both directions, which is the point. The studio has built and shipped real agents where the brief needed one: a customer-support agent for a SaaS company selling accounting software to small businesses in California, and a project-manager “whip” agent for a design agency, which turns meeting discussions into actionables, breaks them into tasks, assigns them to the right projects and tracks them until they are actually done. And it has advised the least fashionable category of all when that was the honest answer: a small industrial machinery manufacturer in Texas asked for AI to track working capital, and the studio told them not to use AI at all - a one-time-license software met every requirement at a fraction of the cost.

The part most pages leave out

When not to choose Foxnut Studios

Situations where another option is the better call, and where we say so in the first conversation rather than the fourth.

• You have chosen the category and want a product: which vendor's platform to license. Your own team, trialling two or three products on a real workflow for a week. Tool selection is not judgment work worth consulting rates, and a studio that took it on would be selling you its reading list.
• The work is high-volume integration across legacy systems - classic enterprise automation at scale. An RPA or systems-integration firm with a delivery bench. That is a mature category with real specialists; a two-person judgment studio is the wrong shape for it.
• You want to explore what is technically possible at the frontier, without a named workflow or a measurable outcome attached. Your own engineers, a research partner, or a lab. Exploration is legitimate work, but it is not an operations engagement, and billing it like one is how pilots that were never going to ship get funded.
• Procurement requires named client case studies in your category before anyone signs. A firm with a public client list. Foxnut Studios has no publishable named client case study yet - better said here than discovered in procurement.

Foxnut Studios works on briefs like this one from Bengaluru and Paris. If you want the shape of that before you talk to anyone, here is what an AI engagement covers and what you keep.