Short answer: OUTMO is a six-step method for rewiring one result once AI is in the room: define the outcome, agree the commitments, choose the contributions (people, AI, software), delegate the authority, deliver and check the evidence, then learn and begin again. It is written for founders, CEOs, managers and HR leaders, in India and elsewhere. A tool chosen before the outcome is named speeds up the old cut of the work.
I am Vinay Pasricha, Founder & CEO of GoodSpace AI and Founder & Chairman of WLCI. GoodSpace has served 1,200+ corporate customers. Hiring was the first proven use of outcome-delivery AI agents there. The same shape of work applies to any result a customer would miss. This page is the practical account of the six steps in OUTMO: How AI Will Rewire Work, Careers and Companies. The book page is OUTMO. The longer argument about departments and outcomes is Collapse or Collapse. Together with the six earlier books, these are eight.
What the six steps are for
Most companies I sit with have already bought AI. The sales desk has a draft button. Quality has a checker. The founder has a dashboard. Each desk moves faster. The customer still waits for the thing they were promised.
That is the failure I describe in why AI is not improving productivity. The tool landed inside the old structure. OUTMO is the sequence I want a leadership team to run on one result before they buy the next tool. Outcome-based AI agents is the same idea at the level of the company. This page stays on one result, and on the order of the six steps.
The order is the same in a promoter-led plant in India and in a company anywhere else. What changes is local: who can actually say no, which system holds the record, and which promise cannot be undone. The steps stay.
HR leaders belong in the room because contributions change jobs. Once AI holds a step, the role that used to hold that step has to be rewritten. That rewrite is a manager’s job and an HR job. It is a poor subject for a software announcement.
The six steps
The OUTMO book page states the method in this order. The lines below are that published sequence, then the test I use with a leadership team.
| Step | What you do | The test |
|---|---|---|
| 1. Outcome | Define the outcome. | A person outside the company would recognise the sentence, and would miss it if it failed. |
| 2. Commitments | Agree the commitments. | Named people have agreed what they will keep, in writing, before anyone picks a tool. |
| 3. Contributions | Choose the contributions: people, AI, software. | Each step has one kind of contributor. People, AI and software are three choices, kept distinct. |
| 4. Authority | Delegate the authority. | The person who owns the result can stop a promise and stop a delivery. A tool cannot overrule them. |
| 5. Evidence | Deliver and check the evidence. | The check is what was delivered and what was paid, read against the promise. |
| 6. Again | Learn and begin again. | A miss changes the commitment or the contribution before the next cycle. |
1. Outcome
Define the outcome in the receiver’s language. A correct part on the date promised, accepted, and an invoice that matches the shipment. A person who joins and stays, and an invoice that is paid. A claim that closes with the customer informed.
“More quotes”, “AI in quality” and “a copilot for the team” are activities. They can be contributions later. They are a weak place to start. If the sentence still names a department, go back and name the result that department was hired to help produce. AI organisation design is that cut, at the scale of the whole company.
2. Commitments
Agree the commitments before you choose the tools. A commitment is a promise a named person will keep: sales will promise a date the plant can meet; quality will release a batch that matches what the customer specified; accounts will invoice what shipped.
Write the promises down. A meeting that ends with “we are aligned” has not finished this step. The people who must keep the promises have to see their names on them. In a promoter-led company this step is short when the promoter is in the room, and it is fiction when the promoter is not.
3. Contributions
Choose the contributions: people, AI, software. Keep the three apart.
People hold judgement, care, and the act that cannot be undone: accepting a date, releasing a batch, signing a price exception, telling a customer the truth.
AI can draft, match, chase and assemble a status from records that already exist. It is a contribution to the result. It is a poor owner of the result.
Software is the system that already holds the record: the ERP, the inspection log, the ledger. AI that keeps a second set of numbers beside that system has created a new argument, and the argument will win on the day someone is audited.
A role changes when its old contribution moves. HR’s work at this step is to rewrite the role around the contribution that remains human, and to say so to the person in the role. Do you need a Chief AI Officer? is the question of who owns that rewrite across several functions. In a smaller company the founder owns it for the first year. AI for MSMEs in India is where I would start if the company is still one plant and one promoter.
4. Authority
Delegate the authority. The person who owns the outcome needs the power to stop a promise and to stop a delivery. A dashboard does not have that power. A committee that can reverse the stop on a phone call has not delegated it.
I use a plain test, the same one I use for any result a person signs. If a senior person can overrule the signature without taking the result onto their own name, the signature was a recommendation. Name the person who owns the result up to the next gate, and give them the authority the gate requires.
This is the step founders skip, because it is the step that moves power. The other five can be written as a memo. This one changes who is allowed to say no on a Tuesday afternoon.
5. Evidence
Deliver, then check the evidence. The evidence is the thing the customer received and the money that followed, read against the commitment. A rise in quotes sent, tickets closed or messages drafted is a record of motion. It becomes useful when you can tie it to the outcome. On its own it is the old scoreboard, moving faster.
Sample. Read a few promises against a few deliveries every cycle. Call the customer when the record and the delivery disagree. The person with authority does this. An agent can lay the two records side by side. The judgement of a mismatch stays with the person.
6. Again
Learn and begin again. A missed date is a fact about the commitment or about the contribution. Change one of them before the next order, and write down what changed. A cycle that produces a lesson and leaves the promises untouched has not reached this step.
One pass is a project. The method is the loop. Companies that stop at step five have a pilot. Companies that return to step one with a sharper outcome have a way of working.
A worked example: one result in a components business
OUTMO includes a fictional case, Aster Components, and a Practical Implementation Handbook. The book page names both. It does not publish the case, and I will not invent it here. The sketch below is a separate components business, so a founder can see the six steps on one result. It is an illustration for this page. It is not a chapter, and it is not a customer of GoodSpace.
Take a components maker that machines parts for two vehicle makers. The result the customer feels is a correct part, on the date that was promised, accepted at their gate, and an invoice that matches what arrived.
Outcome. Write that sentence on one line and stop. “AI for the quality team” comes off the page. So does “more RFQs”. The outcome is the part, the date, the acceptance and the invoice.
Commitments. Three promises, named. The sales lead will not confirm a date the planner has not accepted. The quality lead will not release a batch that misses the drawing the customer sent. The accounts lead will invoice the shipment, in the quantity that left the gate. Each name is a person, not a department.
Contributions. People accept the date, release the batch and sign a price exception. AI drafts the status from the order and the inspection record, matches the drawing revision to the order, and chases a missing certificate overnight. The ERP and the inspection log remain the record. AI reads them. It does not keep a parallel sheet that sales trusts and the shop ignores.
Authority. The person who owns this outcome can refuse a date and can hold a shipment. A message from a director that releases the shipment anyway has to carry that director’s name as the new owner of the outcome, or it does not release. That is the whole of the authority step in this sketch.
Evidence. Each Friday, a sample of promises against receipts: what date was confirmed, what the customer accepted, what the invoice said. The score is that sample. Quotes issued that week stay in the record and stay off the score.
Again. A late part changes one thing before the next order is confirmed. Either the commitment (the date rule) or the contribution (who checks the revision, or which record the draft status reads). The change is written where the three named people can see it.
That is one result, one cycle. A second result, collections or a new customer, waits until this loop has run. Starting six outcomes in the same month is how the method becomes a slide.
What each seat does on Monday
Founder or promoter. Pick one outcome the customer would miss. Sit in the room when the commitments are named. Delegate the authority in a sentence other people can repeat, and keep your hands off the stop you just gave away. In many Indian companies this is the whole job of the first month. The tools can wait.
CEO. Keep the six steps in this order for that one outcome. Refuse a tool purchase that arrives before step three, and refuse a tool purchase that skips step four. Ask for the evidence sample, and ask what will change because of it.
Manager who holds the result. You are the name on the authority step. Your week is the commitments, the gates you still sign, and the sample. Status that an agent can assemble from the record does not need a meeting. A mismatch does.
HR leader. When a contribution moves to AI or already sits in software, rewrite the role. Say what the person still decides, what they no longer type, and how their work will be read: against the outcome, through the evidence sample. Hire for the contribution that remains. AI for Business Leaders is the earlier book on designing that company before the tools arrive. OUTMO is the six-step pass on one result.
A smaller company can run this with the founder in the authority seat and one manager beside them. A larger one can run it on a single plant, a single customer, or a single product family. The method gets worse when the first outcome is “the whole company”.
What I would leave for the book
This page is the sequence and one illustration. OUTMO is the full method, the fictional case Aster Components, and the Practical Implementation Handbook. The Kindle edition is ASIN B0H7L56HNY, ₹399 on Amazon.in. Read OUTMO on Kindle, or start at the book page. All eight books on Kindle are on the Kindle shelf.
If the question in front of you is the company rather than one result, read Collapse or Collapse next. If the question is who should own AI across several functions, read the Chief AI Officer guide.
FAQ
What is the OUTMO six-step method?
It is the method in Vinay Pasricha’s book OUTMO: How AI Will Rewire Work, Careers and Companies: define the outcome, agree the commitments, choose the contributions (people, AI, software), delegate the authority, deliver and check the evidence, then learn and begin again.
Who is OUTMO for?
Founders, CEOs, managers and HR leaders. The steps are the same in India and elsewhere. What changes is who can say no, which system holds the record, and which promises cannot be undone.
What is Aster Components?
Aster Components is the fictional case in OUTMO. The book also includes a Practical Implementation Handbook. This page does not retell the case. The components sketch above is a separate illustration of the six steps.
Where does hiring sit in this method?
At GoodSpace AI, hiring was the first proven use of outcome-delivery AI agents, across 1,200+ corporate customers. The outcome there is a person who joins and an invoice that is paid. The six steps are for any result of that kind. Hiring is the proof we have already run.
Where can I buy OUTMO, and what does it cost?
OUTMO is a Kindle book, ASIN B0H7L56HNY. The primary listing is Amazon.in at ₹399. The same Kindle is also on Amazon.com. The buy link on this page uses the site’s Kindle redirect, which sends you to the store for your own country.
Read OUTMO on Kindle
The six steps, the fictional case Aster Components, and the Practical Implementation Handbook are in the book. Kindle on Amazon.in is ₹399.