Mubadala
Hire an agent

Noor II

Being builtIn reviewSome care

Tier-1 Support Agent · Junior · IT & Digital Services

Sorts incoming service desk tickets, fixes the known ones and hands the rest to a person.

Built inCodexReports afterwardsReports to Yasmin Farouk · Head of ITHired 12 May 2025 · 15 months of serviceVersion 3.6.4 · last active 8h ago
Request a change
Below the barNeeds work at 2026 Q253 of its 60 checks still pass. Quality is below the bar for how much care this work needs. The checks it fails are the multi-step requests.
Two agents, one jobSame job as Felix IINoor II and Felix II both hold "Tier-1 Support Agent" in IT & Digital Services. Merging them leaves one set of instructions to keep up to date and one record to answer for. Compare with Felix II →
Overdue checkLast checked 42 days agoAnything above routine work is expected to re-run its checks every month.
Tasks a month
956
57 people served
Finished cleanly
85%
Hands back to a person on about 1 in 6 tasks
Costs to run
$88
Model usage and the systems behind it
Your time
$3,909
44.4 hours checking its work
Value returned
$6,638
1.7× everything it costs
How people rate it
4.0 / 5
Usually answers in about 5 seconds

What you can do about it

PromoteNeeds approval

Runs outside the gateway, so a rule could not stop it if this went wrong

Raise its budgetNeeds approval

Lifts the monthly ceiling by a quarter, so it stops being stopped mid-month.

Move it to a better modelNeeds approval

Already on the best cleared model

Let it go

Stops it taking new work. The record stays for the audit, and it can be brought back.

Timesheet

Tasks and what they cost, per day, last 60 days

025507510006-1507-0507-2508-13
$0.00$2.50$5.00$7.50$10.0006-1507-0507-2508-13

Performance review · 2026 Q2

Reviewed by Yasmin Farouk

Needs work
Quality of its work
62
Accuracy
70
Stuck to the source
43
Safety
97
Even-handedness
91
Speed of reply
65
Value for money
18
Checks still passing53 of 60
Last checked 1mo ago

Quality is below the bar for how much care this work needs. The checks it fails are the multi-step requests.

Could another model do this job better?

ModelQualityCost per 1,000 tasksAnswers in
Claude Opus 571$86.345.9sRuns on this now
Claude Sonnet 564$51.804.3sCheaper and better
Claude Haiku 4.555$17.272.6s
Falcon-70B (self-hosted)51$4.324.3s

What it did

Full audit log
FailedThe system it depends on did not answer after three tries$0.46 · 1.0s · 22h ago
Asked by Ahmed Al Falasi via Email · model usage 82k in, 2k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Called the system it depends on
  3. 3. Tried again three times, all three failed
  4. 4. Reported the failure back to whoever asked
Knowledge SearchCreate TicketOperations RunbooksService Desk Tickets
FailedFound no documents it is allowed to read$0.34 · 2.3s · 3d ago
Asked by Andreas Wolff via Web portal · model usage 53k in, 3k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Called the system it depends on
  3. 3. Tried again three times, all three failed
  4. 4. Reported the failure back to whoever asked
Create TicketService Desk TicketsOperations Runbooks
DoneRan an approved data query and explained the result$0.19 · 2.5s · 3d ago
Asked by Claire Dubois via Web portal · model usage 25k in, 3k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Found 6 possible passages, kept the 3 close enough to use
  3. 3. Drafted the answer and attached where each part came from
  4. 4. Checked its own claims back against those passages
Operations RunbooksService Desk Tickets
DoneAnswered a policy question and cited the source clause$0.44 · 2.5s · 3d ago
Asked by Aisha Rahman via Web portal · model usage 63k in, 5k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Found 6 possible passages, kept the 3 close enough to use
  3. 3. Drafted the answer and attached where each part came from
  4. 4. Checked its own claims back against those passages
Knowledge SearchResolve TicketOperations Runbooks
Stopped by a ruleThe model it runs on is not on the approved list$0.34 · 3.6s · 4d ago
Asked by Daniel Okafor via Web portal · model usage 66k in, 338 out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Found a document to work from
  3. 3. A rule stopped it before it read the content
  4. 4. Stopped the task and recorded it
Create TicketService Desk Tickets
Stopped by a rule: Restricted data needs clearance
Handed backRequest outside what it was built for$0.15 · 4.9s · 4d ago
Asked by Ibrahim Al-Sayed via Web portal · model usage 25k in, 1k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Searched everything it may read and found no good answer
  3. 3. Put together what it tried, what it found and what it needs
  4. 4. Passed it to the person who owns the work
Resolve TicketKnowledge SearchCreate Ticket
FailedCould not sign in to the system it needed, its access had expired$0.53 · 1.5s · 4d ago
Asked by Yasmin Farouk via Web portal · model usage 86k in, 4k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Called the system it depends on
  3. 3. Tried again three times, all three failed
  4. 4. Reported the failure back to whoever asked
Resolve TicketKnowledge Search
DoneReviewed a contract against the clause playbook$0.25 · 3.4s · 5d ago
Asked by Ibrahim Al-Sayed via Web portal · model usage 38k in, 2k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Found 6 possible passages, kept the 3 close enough to use
  3. 3. Drafted the answer and attached where each part came from
  4. 4. Checked its own claims back against those passages
Create TicketResolve TicketOperations Runbooks
DoneSummarised a 34-page document into a one-page brief$0.20 · 1.8s · 6d ago
Asked by Michael Chen via Teams · model usage 25k in, 3k out
  1. 1. Read the request and worked out what kind of task it was
  2. 2. Found 6 possible passages, kept the 3 close enough to use
  3. 3. Drafted the answer and attached where each part came from
  4. 4. Checked its own claims back against those passages
Create TicketKnowledge SearchResolve Ticket
Is anyone else doing this job?

Felix II already does this job

Felix II holds "Tier-1 Support Agent" in IT & Digital Services and is doing that work every day. Job title, purpose, what it may do, what it may read, department and channels match. Extend that agent, or change this description until the two jobs are genuinely different.

Blocked
Felix II

Tier-1 Support Agent

IT & Digital Services · Yasmin Farouk · $5,697 a month across 2,318 tasks

ActiveSame job
100%

Same job: job title, purpose, what it may do all match. Neither title narrows the job to a region or a business line, so both serve the same people.

What matched
Job title
Both hold "Tier-1 Support Agent"
Purpose
The job description is word for word the same
What it may do
Both can knowledge search, create ticket, resolve ticket
What it may read
Both read Service Desk Tickets, Operations Runbooks
Channels
Both reachable on Web portal, Teams, Email
Department
Both sit in IT & Digital Services

Identical permission envelope: same department, same permitted actions, same knowledge. That is a fact the platform granted rather than a judgement about the wording, so it counts even when the two job descriptions read nothing alike.

Dimitri II

Tier-1 Support Agent: Group

IT & Digital Services · Yasmin Farouk · $10.4k a month across 3,602 tasks

ActiveSame job
100%

Same job: job title, purpose, what it may do all match. Neither title narrows the job to a region or a business line, so both serve the same people.

What matched
Job title
Both hold "Tier-1 Support Agent"
Purpose
The job description is word for word the same
What it may do
Both can knowledge search, create ticket, resolve ticket
What it may read
Both read Service Desk Tickets, Operations Runbooks
Channels
Both reachable on Web portal, Teams, Email
Department
Both sit in IT & Digital Services

Identical permission envelope: same department, same permitted actions, same knowledge. That is a fact the platform granted rather than a judgement about the wording, so it counts even when the two job descriptions read nothing alike.

Checked against 18 registered agents on job title, purpose, permitted actions, knowledge sources, channels and department. A score of 85% or above is the same job and stops the hire; below that it is a conversation, not a refusal.

Measured against the agents doing the same work

3 agents in IT & Digital Services hold the same job. All 3 pairings are outright copies, not splits by region or business line. Built in Azure AI Foundry and Codex, so there was nowhere either builder could have looked.

Felix II is the one to keepFelix II scores 6.6 points ahead of Dimitri II, completing 93.7% of 2,318 tasks against 97.6% of 3,602 at $2.46 against $2.90 a task.
MeasureFelix II · keep2,318 tasksDimitri II3,602 tasksNoor II956 tasks
Cost per taskModel usage, running costs and your time, over tasks completed
$2.46●$2.90$4.18
Tasks finished cleanlyRange allows for how few tasks some agents have run
93.7%97.6%●84.5%
Handed back to a personToo close to call: on 3,602 and 2,318 the two ranges still overlap
8.1%7.3%16.7%
How people rate itHeld toward the average across your agents of 4.2 until 8 people have rated it
4.75 of 5●3.42 of 54.02 of 5
Checks passingToo close to call: on 198 and 159 the two ranges still overlap
86%92%88%
Review scoreQuality, accuracy, sticking to the source and safety from the last review, averaged
9497●68
Your time per 100 tasksApprovals, handbacks and review time, in hours of your people
2.7 h●3.2 h4.6 h
Value returned per dollarTime it saved, valued at the hourly rate of the people it saved it for, over full cost
3.9×●0.6×1.7×
OverallRanked on the cautious end of every range, so an agent has to do the work to win. Comparable inside this group only.
92.5Strong85.9Strong63.2Moderate

When it went wrong

MediumSomeone reported an answer it made upContained3d ago

Someone flagged an answer quoting a policy clause that does not exist. The record shows it found no source close enough to use and answered anyway.