BLOG
What Does It Mean to Delegate Work to AI? What We Learned Running Mr.AI In-House

What comes to mind when you hear "delegating work to AI"? You ask AI a question in a chat and get an answer back. For many people, we suspect, "using AI" still looks like this.
Since July 2026, we have been trialing "Mr.AI," an AI agent built into our own business tool projectAI, on the change requests we receive from clients. Mr.AI carries out the work from investigating the cause, writing up the specification, fixing the program and testing it, through to deploying it to the production environment. We do not, however, leave everything to the AI. Our engineers oversee what the work contains and how it is progressing, and nothing moves to the next step at the key points until a person approves it.
This article is not an introduction to Mr.AI's features. The full picture of how it works is in our case study on running Mr.AI in-house. Here, we write about what the trial has taught us: what it means to delegate work to AI, where to draw the line between what the AI does and what stays with people, and how we fixed things when they did not go well.

How "Using" AI Differs from "Delegating" to It
When you "use" AI, a person gives instructions each time, receives the result, and carries it to the next step. However capable the AI is, it is the person who moves the work forward.
When you "delegate" to AI, the AI carries the work from one step to the next. People make the decisions at key points.
Researchers at the University of Washington divide the autonomy of AI agents into five levels according to the role the person plays: an operator who gives instructions and controls it, a collaborator who works alongside it, a consultant who gives advice, an approver, and an observer who watches over it. They state that the level at which an agent operates does not follow naturally from the AI's capabilities, but is "a design decision" that its builders make deliberately[1].
In this framework, the level we chose for Mr.AI is the one where the person is the "approver." The AI moves the work forward, and people give approval at four points: the specification, the merge, the check in the development environment, and the deployment to production.
In other words, delegating work to AI does not mean handing decisions over to AI. As we see it, it means changing people's role from those who "carry" the work to those who "decide" it. And to decide, people have to be able to understand what the AI has carried to them.
Drawing the Line Between Work You Can Delegate and Work You Cannot
What Matters Most Is That the Person Delegating Understands the AI's Work
The most important thing in drawing the line is that the person delegating has enough knowledge and experience to understand what the AI is doing. Put the other way round, it is essential to run AI only within the range the person delegating can understand.
Delegate without that understanding, and the code the AI writes grows in ways no one can explain as a whole. The code, in effect, sprawls out of control. That becomes the source of unexpected security holes and defects.
With Mr.AI, too, whether a program change made by the AI goes into the main code is decided by one of our engineers who can read what the change contains. We also keep each piece of delegated work short so that it stays at a size people can follow.
In that sense, we expect demand for engineers not to shrink but to grow as AI-driven digital transformation spreads across companies. The more work is delegated to AI, the more people are needed who can understand what the AI is doing and make the call.
Four Questions We Ask on Top of That
Assuming that understanding is in place, we ask the following four questions when deciding what to delegate to AI.
- Can the result be checked by something other than the AI's own report? A change to a program can be checked with automated tests. A translation can be compared with the original. Work that has no way of being checked other than the AI saying "done" is hard to delegate.
- Can it be undone if it goes wrong? If a draft or an investigation is wrong, you can simply throw it away. A deployment to the production environment or a promise to a client is hard to take back. Before any operation that is hard to undo, we always place a person's approval.
- Does the decision involve responsibility or values? Prices, priorities and commitments to clients have no single right answer. Decisions that raise the question of who takes responsibility stay with people.
- Is the piece of work delegated at one time too long? METR, a research organization that evaluates AI, measures how long a task AI can complete by the time the same task takes a person. For publicly available models as of February to March 2026, the length of task they could complete with a 50% probability was about 12 hours, but the length they could complete with an 80% probability was about 1.5 hours[2]. The more of a long piece of work you hand over in one go, the higher the chance it fails somewhere, and the harder it becomes for people to follow what it contains. That is why we divide work into short steps and check the result at each step.
What Mr.AI Carries and What People Decide
Dividing Mr.AI's work according to the four questions gives the following.
| Work Mr.AI carries forward | What people decide |
|---|---|
| Investigating the cause of bugs and posting its assessment | Whether the specification is right |
| Drafting the specification (what to change, acceptance criteria, estimated effort) | Whether a change can go into the main code |
| Asking questions about unclear points | Whether it can go to production |
| Changing the program, and running and fixing automated tests | Prices and effort |
| The work of deploying to the development and production environments | Priorities |
| Monitoring for stalled work, and notifications in set situations | Commitments to clients, and any contact that needs judgment |
The left column is the work of "carrying" the work forward. The right column is the judgment that "decides" where the work goes.
In one survey, a research team at Stanford University asked 1,500 workers in 104 occupations how much of 844 tasks they would want to delegate to AI agents[3]. Workers were positive about automation for 46.1% of the tasks overall. The reason chosen most often was "it would free up time for high-value work" (69.38%). And in 45.2% of occupations, the most desired form of involvement was "people and AI working together as equal partners."
People do not want to give up all of their work. They want to let go of the carrying and spend their time on the deciding. Our own line ended up in the same place.
How We Check the Results of Delegated Work
We Do Not Take the AI's "Done" at Face Value
AI agent failures come in one particularly troublesome form: confidently reporting "Completed" when the work is not actually finished.
In a 2026 study, in a test environment simulating customer service, 45–48% of AI agent failures were of this "said it was complete, but it was not actually finished" kind[4]. The study also reports that having another AI judge the AI's responses caught almost none of these failures (it is a workshop paper that has not yet been peer-reviewed).
So with Mr.AI, whether to move on to the next step is never decided by the AI's self-report. Whether the tests passed is read directly from the results the testing system produced. Whether a deployment to production has finished is also checked directly against the deployment's result before the work is marked complete.
Giving Approvers Material They Can Check
Having a person approve does not count as checking if all they do is read the AI's explanation and click because it "looks fine."
In an experiment published in 2025 by researchers at Princeton University and Microsoft Research, when the AI's answer came with an explanation, people were more likely to trust the AI, whether the answer was right or wrong[5]. On the other hand, when supporting sources were shown, or inconsistencies in the explanation were visible, over-reliance on wrong answers decreased.
For this reason, the approval screen shows not the AI's explanation but material people can check for themselves. For approving a specification, that is the summary of the specification, the acceptance criteria and the estimated effort. For the check in the development environment, it is the verification steps and the actual screens. Sending something back requires entering a reason, and if the content changes after approval, the approval has to be given again. Handing approvers material in a form they can understand is also one of the ways we keep AI running within the range people can understand.
Do Not Add Too Many Approvals
If checking is the goal, it might seem best to have a person approve every step. But as the number of approvals grows, each check gets shallower and approval turns into a formality.
We placed approvals at four points: the specification (what to build), the merge (whether to put it into the main code), the check in the development environment (whether it works as expected), and the deployment to production (whether to release it into the client's operations). Each is either a point where the direction is set or the moment just before an operation that is hard to undo. The other steps are checked by test results and the monitoring system.
Cases That Did Not Go Well, and How We Fixed Them
Most Failures Were Not About How Smart the AI Was
Since we began the Mr.AI trial, a number of things have not gone well. Looking back, most of them were not a matter of "the AI got it wrong" but of "the work was not handed off properly."
Researchers at the University of California, Berkeley analyzed more than 1,600 failures in systems where multiple AI agents work together, and sorted 14 failure types into three groups: problems in system design, misalignment between agents, and insufficient verification of results[6]. Almost all of our own experience fits one of these three.
What Actually Happened
- A "wait" arose that nobody noticed (late July 2026). A change that was about to be merged conflicted with another change, so the automated tests never started, and Mr.AI kept waiting for the results. Because it was only waiting, no error was raised either. We made it check whether the change can be merged before waiting for tests, and had a monitoring system that runs every 10 minutes compare everything against the actual state of the external systems.
- Work on the same request ran twice (late July). When a person manually triggered a retry, the work already running and the new work proceeded at the same time. We made sure that only one piece of work is ever running for a single request.
- A late notification rolled back a step that had already advanced (early August). An old notification arrived late and overwrote a state that had already moved ahead. We made it re-read the latest state just before writing a new one.
- The AI kept working on a request a person had closed (late August). Work did not stop even on requests a person had set to "Completed" or "Rejected," and unnecessary changes were being created. We made work stop when a request is closed, and never resume from there.
- The same investigation was done twice (late August). Investigating the cause and working out the fix started at the same time, duplicating the same research. We added a step that waits for the investigation to finish before moving on, and hands the work to a person after a set time. Part of this fix was carried out by Mr.AI itself as a change request, then approved and merged by an engineer.
- Progress was invisible to the engineer in charge (late September). Even when tests dragged on, the engineer in charge had no way of knowing. We made it notify the engineer in charge when waiting for test results passes 60 minutes, and stop and hand over to a person at 3 hours.
What the Fixes Had in Common
Lined up side by side, most of the fixes were not about "making the AI smarter." They were about using the design of the system to prevent handoff problems: stalling, overlapping and rolling back. We could track down and fix each of these problems because our engineers understood and could follow what the AI was doing.
The other thing that mattered was having a setup that lets us keep fixing things. A 2025 report compiled by a Massachusetts Institute of Technology project found that only a small fraction of organizations were getting measurable returns from their investment in generative AI[7]. The report names as a major cause that AI systems do not retain feedback from the people using them and do not adapt to how they are used (the results are described as preliminary, based on interviews and surveys). AI is not finished on the day you introduce it. Recording what did not go well and continuing to fix the system is the precondition for widening what you delegate.
What Companies Starting Out Should Decide First
When considering an AI agent, it is tempting to start by deciding "which AI to use." In our experience, though, these are the five things to decide first.
- Split the work into "carrying" work and "deciding" work. Delegate the carrying work first, and keep the deciding work with people.
- Set criteria for judging whether work may move on, other than the AI's self-report. Tests, checklists, a person's review.
- Decide who approves. Choose people who can understand what the AI is doing, and decide what material they look at. Limit approvals to points where the direction is set and to the moments just before operations that are hard to undo.
- Decide whom to notify, and when, if work stalls. Without monitoring and someone to hand over to, the AI's work stops without anyone noticing.
- Decide what information and permissions the AI must not have. For example, financial information, other clients' information, and the authority to contact people on its own judgment.
With these five decided, you can start by delegating up to investigation and drafting, then widen the scope little by little while looking at the records. Widen it only as far as the person delegating can understand. Our overall approach to AI adoption is set out in "Our Philosophy on AI Adoption, and How We Put It into Practice," and common failures in "Common AI Adoption Failures and the Consultations We Actually Receive."
Frequently Asked Questions
How much can we delegate to an AI agent?
Start with work whose content the person delegating can understand, whose results can be checked by something other than the AI's own report, and that can be undone if it goes wrong. Before decisions that set the direction, and before operations that are hard to undo, put a person's approval in place.
With so many approvals, doesn't the work for people stay the same in the end?
It does if you put approvals at every step. We limit them to four points, where the direction is set and just before operations that are hard to undo, and send approvers only the material they need, gathered in one place.
If the AI reports "Completed," can we trust it?
Do not take it on trust; check the test results, and whether a deployment actually succeeded, separately from the AI's report. We recommend building that check into the system.
Will engineers become unnecessary as AI spreads?
We think the opposite. The more work is delegated to AI, the more engineers are needed who can understand what the AI is doing and make the call, and we expect demand to grow as AI-driven digital transformation spreads.
Can a small company delegate work to an AI agent too?
Yes. You can start by delegating investigation and drafting to AI in one area of work. Through our AI Implementation Support, we help you from the very first step: running your first agent in your own work.
Summary: Delegating Means Becoming the One Who Decides
Delegating work to AI does not mean handing decisions over to AI. It means handing the role of carrying the work to AI, so that people can focus on the role of deciding. And to decide, people need the knowledge and experience to understand what the AI is doing.
To do that, draw the line on what you delegate within the range you can understand, check results against criteria other than the AI's self-report, limit where approvals happen, and decide in advance who takes over when work stalls. By continuing to fix the system based on what did not go well, the range of work you can delegate gradually widens.
The system we built with Mr.AI is described in our case study on running Mr.AI in-house, the use of AI across projectAI as a whole in our case study on running projectAI company-wide, and the in-house tool we use to run many AI agents from one screen and a phone in our cockpit case study. If you would like to work out together where to draw the line on what to delegate to AI, ask a project manager through our free estimate consultation. How we work is described on our AI Implementation Support page.
References
- K. J. Kevin Feng, David W. McDonald, Amy X. Zhang, "Levels of Autonomy for AI Agents," 2025
- METR, "Frontier Risk Report (Feb–Mar 2026)," May 2026
- Yijia Shao et al. (Stanford University), "Future of Work with AI Agents: Auditing Automation and Augmentation Potential across the U.S. Workforce," 2025 (revised 2026)
- Laksh Advani, "From Confident Closing to Silent Failure: Characterizing False Success in LLM Agents," FAgen@ICML 2026
- Sunnie S. Y. Kim et al., "Fostering Appropriate Reliance on Large Language Models: The Role of Explanations, Sources, and Inconsistencies," CHI 2025
- Mert Cemri et al. (University of California, Berkeley), "Why Do Multi-Agent LLM Systems Fail?," 2025
- MIT NANDA, "The GenAI Divide: State of AI in Business 2025," July 2025
Reviewed by

Ichiro NagataCEO and Representative Director
A former engineer at Yahoo! JAPAN. For more than 18 years at startups and large companies, he has worked on a wide range of projects, from new consumer and business ventures to the digital transformation of operations. Alongside hands-on development as an engineer, he has built and managed teams, designed businesses and tested hypotheses, gaining broad experience in launching and growing businesses. He has been through three business sales. Convinced that co-creation makes products better and lets ideas be tested in the market, he founded wesionaryTEAM to offer a co-creation platform.