TODAY, FOR MOST TEAMS
Without MCP
Claude can write a good plan. Then you copy it into messages, chase each person yourself and paste their answers back.
THE PLANNER AND MCP
Plan with your AI.
Efforti gets it done.
Tell Claude, or any AI assistant that supports MCP, what you want done. Efforti gives every person their own part, ten people or a hundred, follows up with each one until it is done, and brings the answers back to your assistant.
- Planner, assignments, proof, WhatsApp, vendor linksEnterprise Early access
The planner is in early access on Enterprise. We switch it on with you and set it up together, and details may change as we learn. Everything else on this site works today.
ONE SENTENCE, EVERY PERSON
Tell your AI what you want done.
Efforti gets every person on it.
Write the intent once to your AI assistant. It drafts the plan and checks it against your real team, then Efforti gives each person their own part and follows up with every one until it is done. Ten people or a hundred: the same sentence. Everyone can still push back before a part becomes theirs, which is what makes a hundred at once safe.
Your AI assistantEfforti connected through MCP
Get all six sites ready for the December handovers. Every site engineer and supervisor gets their part. Preview it first.
Preview ready: 18 parts for 112 people across 6 sites, and every name matches someone in Efforti. 3 people are away, so theirs will wait until they are back. Send?
Send.
112 people · 6 sites · 18 partsEach person sees only their own part
Wakad · 22
Baner · 18
Hinjewadi · 24
Kharadi · 16
Moshi · 20
Talegaon · 12
- 97 accepted
- 9 still reviewing
- 3 asked a question
- 3 away, theirs waits until they are back
1sentence from you
112people, each with only their own part
0reminders you send yourself
3people away, held until they are back
Sample company and sample data.
MCP, IN PLAIN WORDS
A safe way for your AI
to use Efforti.
MCP, the Model Context Protocol, is the open standard behind connectors in Claude and ChatGPT. It lets an AI assistant use other software through a set of named tools, the way you use apps on your phone. Connect Efforti, and your assistant can work with your team through Efforti, inside Efforti's rules.
WITH THE PLANNER
With Efforti connected
Claude can see who is on your team and who is away, give each person their part, ask questions and read back what people said.
YOU STAY IN CHARGE
Inside Efforti's rules
Claude can only reach your team through Efforti, so working hours, a daily limit per person and out of office all still apply.
HOW IT WORKS
From a conversation with your AI
to work your team agreed to.
Connect once
During setup we connect Claude to your Efforti workspace with you. Efforti then appears in Claude as a set of tools Claude can use when you ask.
Plan as you normally would
Talk the plan through with Claude: goals, owners and dates. Claude checks your team in Efforti, so names and leave are right.
Preview, then send
Claude runs a dry run first: every name matched to a person, anyone away flagged, nothing written. Then each person gets their part to review.
The answers come back
Accepts, pushback, questions, blockers and check-in answers flow back. Ask Claude "how is the Q4 plan going?" and it reads them from Efforti.
A PLAN FROM CLAUDE
Preview first. Then each person gets their part.
Your assistant checks the plan against your real team before anyone sees it.
Claudewith Efforti connected
Send the Q4 plan to the team in Efforti. Show me a preview first.
Here is the preview. Six parts for five people, and every name matches someone in your workspace. Farhan is away Thursday and Friday, so his part reaches him when he is back. Shall I send it?
Q4 · Growth and ProductPreview
- Pricing page v2 liveTanay · Fri
- Final pricing copy, signed offMeera · Thu
- Webinar page formDev · Fri
- Funnel teardownFarhan · back Mon
- +2 more
Yes, send it.
- A dry run before every plan. Claude can preview a plan without sending anything: each name matched, and anyone out of office flagged.
- A proposal, not an order. Each person reviews their part before it becomes theirs. By default, nothing recurring starts until they accept.
- Working hours and a daily limit. Efforti only messages people inside the working hours you set and under a daily limit per person. Only something you mark urgent skips the hours.
- Out of office is respected. Someone who is away gets their part when they are back, not on their holiday.
WHAT YOUR AI CAN AND CANNOT DO
A short list of tools.
A longer list of limits.
Your assistant only reaches your team through these tools. Efforti enforces the limits, whatever the conversation says, for ten people or a thousand.
WHAT YOUR AI CAN DO THROUGH EFFORTI
- See who is on your team, their teams and who is away
- Preview a plan: match every name and flag anyone on leave
- Give each person their part of a plan to review
- Ask one person one question, with a deadline
- Propose a task, with what done looks like and who checks it
- Ask a lead to break a piece of work into tasks
- Answer a person's question about their part, or revise it
- Read progress: answers, blockers, missed check-ins and scores
WHAT IT CANNOT DO
- Message anyone outside working hours, unless you mark it urgent
- Go past the daily message limit for any person
- Keep more than three open questions with one person at a time
- Reach someone who is out of office before they are back
- Read raw chat transcripts: it gets answers and blockers, not the conversation
- Watch anyone: no screenshots, keystrokes, activity or location
- Make a score final: a lead approves every score
WHAT YOUR TEAM GETS
Their part, not your whole plan.
People can push back before they commit. Once they accept, Efforti follows up until it is done, or tells you why before the date.
- Their part, with the why. Each person sees the goal, the date and the reason it was given to them. Nobody sees anyone else's part.
- Push back before they commit. They can accept, ask for a change, push back with a question, or decline with a reason. Nothing is silently dropped.
- Questions answered from your context. Efforti answers questions about the plan from the background you gave it. What it cannot answer comes back to you.
- Revised with a reason. When a part changes, the person gets the new version with the reason for the change.
DONE MEANS PROOF
Nothing closes without the proof.
Every assignment says what done looks like. The review checks it, item by item.
INBOX · REVIEW
Round 1 of 2Slab 13 pour
Rohit · due Sat · Block A handoverPour photo from bay 2IMG_2314.jpg, 4:12 PM
Quantity poured, in m³14 m³, 3 batches
Cube test sample taken and labelledRohit says taken. No label photo.
AcceptReturn with a noteReturn sends it back to Rohit for round 2.
- Done is defined up front. A photo, a file or a number, agreed before the work starts.
- No "done, unverified". Done without its evidence keeps the work open, with what was given saved.
- Checked by someone else. The person who checks the work is never the person who did it.
- Rounds, with notes. A return says what is missing and starts the next round. Client-facing work needs a senior reviewer's accept.
ALSO IN EARLY ACCESS
Ways in that do not need an AI assistant.
Early access
Plans from a file or a sentence
No Claude needed: drop in a spreadsheet or PDF, paste a plan or type a sentence. Efforti shows what it understood before anything goes out.
Early access
WhatsApp, permission first
The first message on WhatsApp is a hello from you that asks permission. One word stops WhatsApp messages, and the app still works.
Early access
Vendors answer by link
Suppliers, clients and volunteers answer an ask from a link. No login, no seat, and they never see the plan or the team.
Early access
A shared pool, with a limit
Open work sits in a pool anyone on the team can pick up, with a limit so nobody holds more than they can finish.
WALKTHROUGHS
A week with the planner, in under 80 seconds.
Sample teams and sample data. These walkthroughs show early-access features as they are designed today, so details may change before general release.
QUESTIONS
About the planner and MCP.
What is MCP?
The Model Context Protocol, an open standard for connecting AI assistants to other software. A connected app shows up in the assistant as a set of tools with clear names and limits. Efforti's tools let Claude plan with your real team and hand the plan to Efforti.
Do we need Claude to use Efforti?
No. Check-ins, follow-ups, blockers and the morning summary work on their own, on every plan, today. Claude is one way to bring in a plan, and in early access a plan can also come from a file or a typed sentence.
Can Claude message my team without me?
Claude acts when you ask it to in a conversation, and only through Efforti's tools. Every plan can be previewed first, people review their part before it becomes theirs, and Efforti's working hours, daily limit and out of office rules always apply.
What does Claude see about my team?
Names, roles, teams and who is away, plus the answers, blockers and scores on the work it handed over. Not raw chat transcripts, and nothing Efforti does not collect: no screenshots, keystrokes or location.
Which assistants work with it?
We set it up with Claude today. MCP is the open standard behind connectors in Claude and ChatGPT, and other assistants support it too, so they can connect the same way. Tell us which one you use and we will say what we have tested.
Is there a limit to how many people one request can reach?
One conversation can send several plans, and one plan can reach a whole department. What Efforti limits is how often each person is contacted: working hours, a daily limit per person and out of office apply to everyone, however many people a plan reaches.
How do we get early access?
Request it below. Early access is on Enterprise, and we set it up with you on a call. Bring a real plan.
REQUEST EARLY ACCESS
Bring a real plan.
We will connect it with you.
Early access is on Enterprise. Check-ins and follow-ups work today on every plan, including Free.
No credit cardNothing sent until you say soAsked, not watched