Every question on this page came from a real conversation with an operator, and every answer links to the page it came from. Cost first, then how to choose what to deploy, then the roles, then the definitions.

99 questions. If yours is not here, ask it and we will answer it properly.

Cost and pricing

How Much Do AI Agents Cost?

Is there a minimum number of agents?

No. Most companies start with one role, on purpose, because a single role done properly proves the model and produces a number you can take to the next decision.

What does the $1,000 deployment cover?

The build for that specific role: reading from your systems, the rules, the exception queue, and the approval gate. It is a one time charge per role, and it is the entry point rather than a fixed rate.

Is $500 a month the price for any agent?

No. It is the entry point for one narrowly scoped role, which is the scope we recommend starting with. A role covering several systems, high volume, or a complex approval matrix costs more, and you get that number in writing before anything is built.

Can we cancel?

Yes. Nothing is on a long contract. Your software stays in your name and the rules we wrote stay with you, so the work is not lost if you stop.

Why is this so much cheaper than the enterprise products?

Because you are not buying a platform licence or a transformation programme. You are buying one role, on your existing systems, with a person still approving the output.

What if the agent does not work out?

Then you have spent a deployment fee and a few months of running cost, not a headcount, and you have process documentation you did not have before. That is the honest downside, and it is why we start with the workflow where the return is measurable fast.

How Do AI Agents Make Money?

How do AI agent companies make money?

Four main ways: per seat, per role running, usage based, or a share of the outcome. Larger firms also sell agent work as consulting programmes.

Is per seat or per role better for the buyer?

Per role, if the goal is fewer hours spent on a workflow. Per seat pricing pays the vendor more as more people use the tool, which works against that goal.

How do AI agents make money for a business?

By recovering money that was already leaving and by returning hours. Recovery is usually the larger and more defensible number, because it does not depend on anyone's salary.

What is outcome based pricing?

The vendor takes a share of what the agent recovers. It aligns incentives well and it requires a measured baseline, which is why it is uncommon.

Can an AI agent save money?

It can, and savings cases are typically smaller than recovery cases and harder to prove, because they depend on a headcount decision that somebody has to own.

How do I know if the agent is worth it?

Measure one full month before deployment, then compare coverage, recovered value, and hours returned. If there is no baseline, the comparison cannot be made.

How to decide what to deploy

Should You Hire Another Person or Deploy an AI Agent?

Is it cheaper to deploy an AI agent than to hire someone?

A first agent role starts at $1,000 to deploy plus $500 a month, about $7,000 a year for a narrowly scoped role. A mid-market operations hire is commonly $70,000 to $90,000 fully loaded. The comparison only holds if the work is rule-driven enough for an agent to own it end to end. If it is not, you will pay for both.

What work should never go to an AI agent?

Negotiation, ambiguous judgment, relationship ownership, and anything where a named person must be answerable for the outcome. Keep those human and let the agent prepare the work.

Can we start with just one agent role?

Yes, and one role is the normal starting point. It proves the workflow on a single measurable process before you redesign more of the team.

How long does it take to see a result?

The first role is usually running inside a few weeks, because the work is already written down somewhere: contracts, retailer portals, an ERP, a shared inbox. The rulebook exists. It just needs to be applied consistently, which is what the agent does and what people run out of hours to do. ---

How to Decide Which Roles Your AI Agents Should Own

How many tasks does a role usually contain?

Between twenty and forty for most mid-market operations and finance roles. The count matters less than the shape. Usually seventy to eighty percent of the hours sit in a handful of rule-driven tasks.

Should an AI agent replace a person or work alongside them?

It depends on how much of the role is rule-driven. Where a role is almost entirely rule-driven, the headcount can go down, which is what happened on a loan operations team that went from six people to two. Where judgment is half the job, the same person does more of the work that matters.

What happens when the agent gets something wrong?

It escalates rather than guesses. That is what the approval rule is for. An agent with a clear escalation path is safer than a person doing a thousand line items at the end of a long week.

Do we need to replace our existing software?

Usually not. The agent typically works inside the tools the team already uses, with monday.com as the command center for visibility. The change is who owns the work, not which software is installed. ---

AI Agents vs Stacked Point Solutions: Why Piecemeal AI Costs More

Are AI agents just another tool?

No, and the difference is where the intelligence sits. A tool executes the step it was built for. An agent owns a sequence of steps and applies judgment across all of them, escalating what it should not decide.

Should we cancel our existing tools?

Not by default. Most agent deployments run inside the tools the team already uses. The point is to stop adding tools for every new fragment, and to let one agent own the sequence that the tools currently split up.

How do we know if our stack is too big?

Count how many tools in one process hand work back to a person. If most of them do, people are acting as the integration layer, and that cost is invisible on the invoice but real in the payroll.

Is an agent more expensive than a point solution?

Per tool, yes. Per outcome, usually no, because you stop paying for the integration tax, the stale data, and the attention of someone hired to reconcile six systems. A first agent role starts at $1,000 to deploy plus $500 a month, sized to the scope. ---

What Is Workforce Design for the AI Era?

What is workforce design in the AI era?

It is the discipline of deciding which roles AI agents own and which stay human, before any tool is purchased. The deliverable is a role map with approval rules, not a software licence.

How is it different from hiring a consultant?

The output is operational. A consultant produces analysis. A workforce design engagement produces agents running in production inside the tools your team already uses, with a documented approval process.

How is it priced?

A first role starts at $1,000 to deploy plus $500 a month to operate it. Each role is visible on its own, so you can see what it costs and what it recovers.

Do we need AI agents to do workforce design?

You can run the task inventory and assignment exercise on your own. What you cannot easily do alone is decide the approval thresholds without having built agents before, which is where the judgment sits.

Which roles should we look at first?

The one where volume is highest and the rulebook is already written down. High-volume reading and matching work, with a measurable output, is the reliable starting point. ---

Roles we deploy: operations and supply chain

AI Agents for Supply Chain Managers

Can an AI agent replace a supply chain manager?

No, and nobody who has run one would claim that. It replaces the high volume checking work inside the role, so the manager spends their time on suppliers and planning instead of chasing documents.

What should a supply chain team automate first?

Invoice against terms matching. The volume is high, the rules exist, and the leakage is already sitting in your records waiting to be measured.

How is this different from supply chain software we already own?

Existing systems report what was entered. An agent reads the source documents, which is what finds the things that were never entered correctly.

Does it work with our ERP?

Where an API exists, yes. Where it does not, the agent reads the same screens your team reads. That works and it costs more to build.

What stops it making a bad decision?

Limits you set plus a person approving exceptions. No serious deployment lets an agent release a payment on its own in month one.

How long before it pays for itself?

For a well scoped matching workflow, the first measured cycle is usually enough to see whether the leakage is there. If it is not, you have learned that cheaply.

AI Agents for CPG Operations Teams

What is the difference between a deduction and a chargeback?

A deduction is an amount a retailer withholds from an invoice, often for a reason stated in the agreement. A chargeback is a disputed or retroactive claim, frequently arriving after the fact and with a filing deadline attached.

How much of CPG deductions is actually invalid?

It varies by retailer and by how well the agreement is understood, which is exactly why the first project should be one retailer and one deduction type. The measurable metric is your win rate on filed disputes, before and after.

Can this work with the portals retailers already use?

Yes. The agent reads the same portals and documents your team reads. Where a portal has an API it helps, but it is not a requirement.

Do we need to replace our ERP?

No. Agent deployments typically run inside the systems already in place. monday.com serves as the command center so the exceptions queue is visible. ---

AI Agents for Accounts Payable Teams

Does an AP agent need to replace our ERP?

No. It reads the systems already in place, and where an API exists it helps but is not required. monday.com serves as the command center so the exception queue is visible to the team.

How is this different from the three-way match we already run?

A three-way match compares the invoice, the PO, and the receipt. It does not read the contract terms, the freight agreements, or the vendor statement. Most of the recoverable money in AP sits in those three documents.

What if our contracts are all PDFs with different formats?

That is the normal starting condition, not an obstacle. It is the work of the first engagement, and it is smaller than most teams expect.

How fast does this pay for itself?

It depends on volume and how long the leaks have been running. On a mid-market AP function the first recovery cycle usually covers the deployment by a wide margin. Ask for the math on your own invoice volume before you decide. ---

Roles we deploy: professional services

AI Agents for Accounting Firms

Do agents replace staff accountants?

No. They replace the portion of the week spent assembling and chasing, which is the part nobody joined the profession to do. The judgment, the client relationship, and the signature stay with people.

Which software does this work with?

The agent reads the systems already in place, including common write-up, tax, and document platforms. Where an API exists it helps, but it is not required. monday.com serves as the command center so open items are visible to the whole engagement team.

Is client data safe?

This is the first question a firm should ask any vendor, and the answer should be specific: where data lives, who can see it, and how long it is retained. Ask for it in writing before a pilot.

What does it cost?

Deployment is a one time setup of about $1K per role, then a monthly fee per role that depends on volume and scope. The firm keeps its own software. Compare that against one week of unbillable seasonal hours before you decide. ---

AI Agents for Legal Transaction Coordination

What does an AI paralegal actually do?

It owns the process half of the role: deadlines, document collection, checklist completeness, drafting from templates, and assembling binders. It does not advise clients or interpret the law.

Can a law firm use AI agents without breaching confidentiality?

Yes, with defined data handling and a deployment that respects your obligations. Which arrangements are acceptable is a call for the firm, so get it in writing before a document is read.

Will this replace our paralegals?

In most firms it changes what they spend their day on. The hours that move are the chasing and tracking hours, not the judgement. Where a role is almost entirely process, capacity is the honest conversation to have.

Is this the same as legal research AI?

No. Research tools help find the law. This owns the operational process of moving matters from open to closed.

What is the cost compared to hiring another paralegal?

A fully loaded paralegal runs about $90,000 a year. A narrowly scoped role starts around $7,000 in the first year.

How long does it take to deploy?

The first matter cycle is days. Writing down the checklist the firm actually uses is the longer piece, and most firms have never had it in one place.

Roles we deploy: real estate

AI Transaction Coordinator for Real Estate

Can an AI agent be a transaction coordinator?

It can own the tracking, chasing, and compliance completeness across every open file. A licensed person still handles judgement, negotiation, and anything that needs advice.

How much would this save compared to a per file coordinator?

At $400 a file, fifteen files a month is $6,000 a month in fees. A role of this scope starts at $500 a month, so the comparison is between a fee you stop paying and a fee you keep paying.

Is it compliant to have an agent handle contract documents?

The agent tracks completeness and deadlines. It does not interpret the contract or advise anyone. Your broker and your state rules still govern, and a person signs off.

Does it work with our CRM and transaction management software?

Yes where there is an API, including the common transaction management platforms. Where there is not, it reads the same screens your coordinator uses.

Will it catch a missed contingency date?

That is the main reason to deploy it. Continuous deadline tracking across all open files is the single thing a busy coordinator cannot reliably do by hand.

How long does setup take?

Days for the first file cycle. The longer part is writing down your checklist, which most teams have never done in one place.

Roles we deploy: small business

AI Agents for Small Business: What to Deploy First

Do I need technical staff to run an agent?

No. Deployment is a setup project, and after that the agent runs on the systems you already use. The ongoing work on your side is approving what it flags.

What does it cost per month?

It depends on the role and the volume. Fees are priced per role, not per seat, and monday.com licensing is typically paid by the client so you keep control of the account and the data.

Will it work with the tools I already have?

In most cases yes. An agent reads email, spreadsheets, portals, and common business systems. Where an integration does not exist yet, that is part of the scoping conversation rather than a blocker.

How do I know if it is working?

Pick one number before you start. Missed calls answered, days sales outstanding, quotes sent. If that number does not move in thirty days, the role was wrong, not the technology. ---

What AI agents actually are

What Are the 7 Types of AI Agents?

How many types of AI agents are there?

Seven in the common business taxonomy. Five are academic, taken from Russell and Norvig's Artificial Intelligence: A Modern Approach (simple reflex, model-based reflex, goal-based, utility-based, learning). Two are architectural: hierarchical and multi-agent systems.

What are the 5 types of AI agents?

The classic five come from Russell and Norvig: simple reflex, model-based reflex, goal-based, utility-based, and learning agents. The seven type list adds hierarchical agents and multi-agent systems.

Are hierarchical and multi-agent the same?

No. A hierarchical setup has a manager agent that splits work and hands pieces to specialist agents. Multi-agent means several agents working together, and they can be peers rather than a hierarchy. Both add coordination problems, which is why small deployments rarely need either.

Which type should a business use?

Almost always a combination of simple reflex for the clear cases, goal-based for the rest, and a learning layer on top. Start with the least autonomy that covers the work.

Do I need a multi-agent system?

Only when the work genuinely splits into parts that need different skills. For one role on one workflow, a single agent plus a person approving exceptions is faster to build and easier to trust.

Is agentic AI the same as an AI agent?

No. Agentic AI describes systems with agent-like behaviour. An AI agent is the specific thing doing the work. The terms get used interchangeably in marketing, which is usually a sign that nobody has defined the scope.

What Does an AI Agent Do Exactly?

Does an AI agent replace a person?

It replaces the part of a role that is rule driven and repetitive. The person stays, approving exceptions and handling judgement. In practice teams shrink only when the volume of that repetitive work was the whole job.

Does the agent work at night?

Yes. Work arriving at 11pm is processed at 11pm. For compliance and finance workflows that arrive overnight, this is usually the first visible change.

How is this different from automation I already have?

Traditional automation follows fixed steps and breaks when the input changes shape. An agent reads the document, understands it, and applies rules to messy input.

Does it read every line or sample?

It should read every line. If a vendor is sampling, ask what share, because a sample finds the problems rather than recovering the money.

What happens when it makes a mistake?

It goes to the exception queue and a person corrects it. The correction should become a rule. A deployment without that loop does not improve.

How long until it is useful?

The first useful cycle is usually days, not months, for a well scoped role. The first trustworthy cycle takes longer, because you need to see it handle a month end.

Is ChatGPT an AI Agent?

Is ChatGPT an AI agent?

Not on its own. It is an assistant that can act agentically when connected to tools, given a persistent goal, and allowed to act without a prompt each time.

Is ChatGPT a chatbot or an AI agent?

Both descriptions fit depending on configuration. The default product is a chatbot. Agent behaviour requires tools, autonomy, and a way to run without a person starting it.

Can ChatGPT do agentic things?

Yes. It can call tools, run multi step tasks, and operate from a schedule in some setups. What it still needs is a goal, limits, and someone reviewing exceptions.

Is Claude an AI agent?

The same answer. It is a model and an assistant, and it becomes part of an agent when someone builds the surrounding system.

Is Gemini an AI agent?

Same. These are models with chat interfaces. The agent is the system built around them.

What is the difference between a chatbot and an AI agent?

A chatbot responds when asked. An agent sees work arrive, decides what to do, acts within limits, and runs without a person starting it.

Can I use ChatGPT to build an agent?

Yes, and it is a reasonable way to prototype. When the prototype needs to run every day without anyone watching, that is a different build.

The platform landscape

What Are the Top 5 AI Agents?

What is the best AI agent?

There is no best one. There is a best fit for the job you have, and the fit depends on whether you want to build, buy a seat, or have a role owned for you.

Is ChatGPT a top AI agent?

ChatGPT is the most used AI assistant, and it can act as an agent when given tools and autonomy. Out of the box it is an assistant that waits to be asked.

Is there a free AI agent?

Yes, for personal tasks, and open source frameworks are free to download. Free means your time is the cost. A framework that runs is not the same as a role that has been designed, tested, and handed over.

Should I choose one platform and standardise?

Only if your work is already concentrated on that suite. Standardising on a suite you do not use will cost more than the agent saves.

How often does this list change?

Faster than you would like. Treat any ranking, including this one, as dated the day it is written. The jobs above have been stable; the products have not.

Who Are the Big 4 AI Agents?

Who are the Big 4 firms?

Deloitte, EY, KPMG, and PwC. The four largest professional services networks by revenue, traditionally audit and advisory.

Do the Big 4 sell AI agents?

Yes. Each has a platform, and each sells agent work as part of advisory engagements rather than as a product you operate on your own.

What is the difference between Zora AI and ChatGPT?

Zora AI is a platform of pre-built enterprise agents covering finance, tax, and audit functions, delivered by Deloitte on NVIDIA infrastructure. ChatGPT is a general purpose assistant that can be given tools. One is a programme, the other is a tool.

Are the Big 4 agents open source?

No. All four are proprietary platforms, though EY.ai and KPMG Workbench are built on Microsoft infrastructure and PwC agent OS is built with Salesforce, CrewAI, and AWS.

Should a small business buy a Big 4 AI agent?

Almost never. The entry point is a readiness assessment that costs more than a full year of running a single agent role on your own systems.

Is it worth paying for the brand?

If the purchase needs board level sign off, regulated audit trails, and an outside name on the risk, yes. If the goal is recovered hours in one workflow, the brand is not doing the work.

Building it yourself

Can I Build an AI Agent on My Own?

Can I build an AI agent with no coding experience?

For simple tasks attached to tools you already use, yes. Anything touching finance, compliance, or customer data eventually needs someone who can test it and handle the failures.

How long does it take to build one?

A useful prototype is an afternoon. A role that runs unattended every day, handles exceptions, and can be audited takes weeks and keeps needing attention after that.

Do I need to know how to code?

It helps a lot. No-code platforms get you started, and they get difficult exactly when the workflow gets messy, which is where the value usually is.

What is the biggest mistake people make building their own?

Building the happy path and calling it done. The exceptions are the job.

Can I build one with ChatGPT?

Yes, for prototyping. When it needs to run on a schedule and act on real data, you are building a system rather than using a chat window.

Is it cheaper to build it myself?

Only if your time is genuinely free and nobody else depends on the output. Otherwise count your hours and compare them against the deployment fee.

Where this goes next

Use cases: the roles our agents own

How much do AI agents cost?

Can I build an AI agent on my own?

Frequently asked questions