← / Return to monday.com

AI Transformation Package

Manual work, on an AI workflow

monday.com's AI Transformation Package puts one AI workflow on a job that already lives in monday.com, agrees the number that has to move, and stays until the team is the one running it.

How the package moves

01 AssessFirst quarter
02 AlignThe middle
03 AdoptLast quarter

Built on monday.com, using only the pieces that process needs.

SidekickAI blocksVibeAgentsWorkflows
0%

of companies have adopted AI, or are planning to

Glide, State of AI 2025
0%

report a real change in how the work runs

Glide, State of AI 2025
0×

output reported by 45% of companies using generative AI

Google, ROI of generative AI
0%

of companies have AI running across the operation

McKinsey, 2025

What you get

A workflow your team actually runs

monday.com's AI Transformation Package finds the job that keeps eating the week, builds it on the account, and leaves a person on the team who can own it.

01
The bottleneck, namedThe manual work gets named, and the process that should go first.
02
The workflow, built for how you workBoards, automations, and the AI pieces that fit, with a proof on a small slice before the rest is built.
03
A before and an afterHours, errors, or cycle time, written in your words and then reported on a dashboard.
04
A champion who owns itTraining, written instructions, and a named person, so the workflow keeps running after the package ends.

Methodology

Assess, align, adopt

These are monday.com's three phases. The work stays with the team until people are using the workflow on a normal day, with a weekly check-in, and nothing goes live until it is signed off.

01

Phase one, about a quarter of the time

Assess

The business is learned before a board is touched, across one to three sessions, and a short conversation with one or two leaders sets the outcomes in their own words.

  • Readiness scored 1 to 5 before the build
  • What the team owns, this year's priorities, and what great looks like
  • The current process, or the target if it doesn't exist yet
  • Each pain sorted: classify and route, extract and enrich, assist and generate, or automate and act
  • An agent first, and a written reason when an agent is the wrong tool
  • A SMART objective: specific, measurable, achievable, relevant, time-bound
02

Phase two, about half the time

Align

The target workflow is designed and proved on a small slice of real items before the rest is built.

  • Solution design tied to the objective, not to a feature list
  • A proof on items from your account
  • Edge cases tested, including what the agent must not do
  • Credit use checked against the budget you agreed
  • Written sign-off before the full build
  • A weekly check-in until go-live
03

Phase three, about a quarter of the time

Adopt

People need to know why it exists, want to use it, know how, and still be using it on a normal day after the build steps back.

  • Training for the people who use it, and a deeper session for the champion
  • Instructions for the workflow, the agent, and the prompts
  • Go-live watch, with issues fixed while the package is still open
  • The baseline read back against the target
  • Handover, and a date for the next conversation

Before it starts

Is the organisation ready

Five things are scored from 1 to 5. A low mark doesn't stop the work. It shows what gets fixed first, usually the data, before an agent is allowed to run.

The bars are a normal starting point, not a pass mark. Most accounts look like this, and the gaps become the plan.

Where a normal account starts

0/ 5
Data3 / 5

Are the boards clean, current, and kept up

Talent2 / 5

Is there someone inside who will learn to run it

Governance2 / 5

Do you know who may use the agents, and what they can see

Culture3 / 5

Is there room to try a new way of working in these weeks

Tooling3 / 5

Can the account support the workflow without a long wait

Packages

Sized to your scope

Three sizes, set by the team and the work. Essential, Advanced, and Premier. The price for each is on monday.com.

Tier 1

Essential

For one team that wants one process running, with a person trained to keep it going.

Up to 8 weeks, one team, up to 50 people

  • 1 basic use case
  • Discovery and an AI opportunity score
  • Up to 5 boards and 1 dashboard
  • Up to 2 AI workflows and 1 basic agent
  • 1 Vibe app, plus Sidekick set up
  • Up to 5 working sessions and 1 training session
  • 30 days of watch after go-live
Most teams start here

Tier 2

Advanced

For two processes, or one harder one, across up to two teams, so intake connects to the work that follows it.

Up to 12 weeks, up to two teams, up to 250 people

  • Up to 2 basic use cases, or 1 advanced
  • Full discovery and a readiness score
  • Up to 10 boards and 2 dashboards
  • Up to 4 AI workflows and 2 basic agents
  • Up to 2 Vibe apps
  • Up to 8 working sessions and 2 training sessions
  • Champion training included
  • 60 days of watch, with check-ins

Tier 3

Premier

For up to three teams, through the build until the connected processes are running, then handed to a champion.

Up to 16 weeks, up to three teams, up to 500 people

  • Up to 5 basic use cases, or 2 advanced
  • The build is included, not only the handover
  • Up to 20 boards and 5 dashboards
  • Up to 6 AI workflows
  • Up to 4 basic agents, or 2 advanced agents
  • Up to 3 Vibe apps, plus a written control plan
  • Up to 12 working sessions and 3 training sessions
  • 90 days of watch, then a quarterly look at the next process

The price for Essential, Advanced, and Premier is on monday.com.

See the price

AI credits

What a credit is

A credit is what monday.com charges when the AI runs. Sidekick, agents, blocks, Vibe, and the notetaker all spend them. The budget is agreed before anything goes live.

The pieces

What each tool actually does

Not all of these are switched on. The process uses the ones it needs, and the credit cost is known before they run.

Agents

For work that needs a judgment across steps, like routing a request, chasing a missing field, or updating the next board. An agent comes first, with a basic one as a short chain and an advanced one reasoning across boards.

Basic, 10 to 50 credits. Advanced, 50 to 250 or more

Sidekick

The assistant inside the account, for when someone needs an answer from the boards right now rather than a standing automation, and it spends credits from the first message.

30 to 150 or more credits a message

AI blocks

One step on one item, such as summarising, pulling a field out of an email, tagging a request, or reading the tone, which is enough when a full agent is more than the job.

8 credits an item

Vibe

A small app on monday.com for when the board is the wrong screen, built for that process and either public or private, The prompts spend the credits.

About 30 to 500 credits a prompt

AI workflows

The chain that connects agents, blocks, and the rules you already have, so a job can move from one team to the next without a handover meeting.

Credits per run, by complexity

AI notetaker

A meeting becomes notes and actions on the item, which is the right pick when the lost work is what was agreed in the room rather than the board itself.

120 credits an hour

Use cases

Where the first workflow usually sits

The package is the same shape for professional services firms, clinics and practices, businesses that run projects, and established retail and wholesale. The first workflow comes from the week, not from a menu.

Professional services

Enquiry in, the right person on it

New work gets classified and routed, the weekly status is drafted from the board, and meeting actions land on the item.

Agents, workflows, notetaker

Clinics and practices

The chase after the first enquiry

New enquiries get sorted and followed up, and referrals or missing paperwork stay visible until someone has actually finished them.

AI blocks, agents, workflows

Project businesses

The job, without the status meeting

Stage, variations, and client updates sit on the board, and a short status comes from the live items while people still make the decisions.

Sidekick, AI blocks, agents

Retail and wholesale

Exceptions, not every order

Ordinary orders stay where they are. The requests that land in an inbox get classified and routed, and a person still handles the ones that need a decision.

AI blocks, workflows, Vibe

What good looks like

By the numbers

One to three outcomes are agreed with one or two leaders, in their words. There is no promised return. The baseline is written first, then read back on a live dashboard at the end.

Productivity

Time back on the repeated job

This is the clearest thing to measure, and it doesn't need a finance model.

Example target: status reporting down by a set number of hours each week

Quality

Fewer errors, the same output

Missed fields, rework, and jobs that come out of the process in a different shape every time.

Example target: exception rate on that workflow down against the baseline

Adoption

People using it on a normal day

Who is active, and how often the agent runs, because a workflow nobody opens is not finished.

Example target: the champion's team on it through the watch period

Business impact

The result you named

Cycle time or response time, only if it was raised. Headcount is not the goal.

Example target: the number from the alignment conversation

Why it works

One workflow, then the next

The package is monday.com's. On the call, the question is whether it is the next step, or whether the workspace comes first.

One workflow first

Not every AI feature is turned on. The next process waits until the first one is in daily use.

An agent first, and a reason when it isn't

If an agent is the wrong tool, the reason is written down, and Sidekick, a block, a workflow, or Vibe is used instead.

The rest of the business, said out loud

Native boards sit inside the package. If the work has to meet the CRM, the inbox, or the reporting already in use, that sits outside this package.

A person on the team

The champion is named before the package ends, so questions have a person on the team.

The before is written down

Hours and error rate are recorded in your words before the build, so the dashboard at the end can show whether it made a dent.

A champion and a next date

The package doesn't close without an internal owner and a follow-up already booked.

Book a call

Find the first workflow

On the call we look at one process and whether Essential, Advanced, or Premier is the right size, or whether the workspace comes first. The price for the three sizes stays on monday.com.

Book a call