Reviews Get Done On Time
When review season opens, the loop launches the cycle, reassigns stuck reviews, reminds reviewers before reviews fall due and compiles feedback summaries — at HR's autonomy level, every action logged.
HOW IT WORKS
How it closes a review cycle
From cycle-open to all reviews submitted on deadline — using the same templates, goals, and role scoping you already configured. It works the cycle before HR has to chase reviewers.
1. Detect
A cycle opens, a deadline approaches, a reviewer falls behind. The loop reads the cycle state before HR has to look.
2. Decide
Picks the action the cycle state calls for — open a cycle whose start date arrived, move a review off a deactivated reviewer, remind a reviewer, or compile the feedback summary for a completed review.
3. Act
Launches the cycle, reassigns the reviewer, sends the due-date reminder or compiles the summary — outright when trust is high, with HR sign-off when the level is lower. Launching a cycle always waits for HR sign-off.
4. Log
Every cycle launch, reviewer reassignment, reminder, and compiled summary is recorded as an action on the console feed.
AUTONOMY YOU CONTROL
Three levels of autonomy. You pick.
Start with it off — it surfaces suggestions but takes no action until you say so. Move to approve for a one-tap checkpoint on every action. Let it run on its own when you're ready.
Off — manual only
Nothing happens on its own — every action the loop would take becomes a suggestion in the HR hub or on the console. HR picks one — it does the rest.
Approve
It proposes the cycle launch, reassignment or reminder; HR confirms with one tap. The pending queue is your daily standup.
Auto
When it's confident, it acts. Only critical or high-impact decisions still come back to you.
Every action the loop takes is listed in "What the AI handled"
When the loop launches a review cycle, reassigns a reviewer, sends a reviewer a reminder, or compiles a feedback summary for a review, the action appears in the "What the AI handled" feed with a "Done by AI" badge and how long ago it happened. An action that failed says "AI couldn't complete" instead.
- "Done by AI" on each cycle launch, reviewer reassignment, reminder or feedback summary the loop completed.
- "AI couldn't complete" on an action that failed, so nothing fails silently.
- Reminders respect notification settings — a reviewer who turned them off, or an organization that did, is not messaged.
- Timestamp on every row — how long ago the loop acted.
One console — HR's home for review-season autopilot
The AI Performance Cycle console is the buyer-facing landing for HR and people-ops during review season. Reviews-complete % sits front and center with a per-cycle sparkline. The "What the AI handled" feed shows the most recent actions across reviewers. The "AI actions to approve" list holds the actions that wait for sign-off.
- Hero metric + trend — reviews-complete % across the active cycle with sparkline.
- "What the AI handled" feed — the most recent cycle launches, reviewer reassignments, reminders sent, and feedback summaries compiled.
- "AI actions to approve" list — actions that wait for sign-off, approved or rejected inline.
- Autonomy dial — Off, Approve or Auto, set on the console.
Why Performance Cycles Drag
AI Performance Cycle attacks the specific failures that turn review cycles into HR coordination drills — for both the manager running the review and the employee receiving it.
Goals Get Set And Then Forgotten Until Review
Q1 OKRs are set in January. Nobody opens them until June, when the mid-year review forces the conversation. By then, the goals that were going to slip already have, and the goals that were going to over-deliver are missing context.
Reviews Are Scoring Theater, Not Coaching Conversations
Managers spend the week before each review pulling together what their reports did — from their own notes, scattered Slack threads, and goal-progress reports. The actual review becomes about reconstructing the past, not coaching the future.
Peer Feedback Doesn't Get Submitted
"Can you write me peer feedback for X?" lands in your inbox during the busiest week of the quarter. Without an in-chat way to submit it grounded in your actual interactions, most peer feedback is hasty or skipped entirely.
Calibration Surfaces Risk Too Late
"Marcus is at risk" surfaces in the calibration committee meeting — three weeks before review delivery. By then, the conversation is about ratings, not about whether the relationship can be turned around with active coaching.
Reward Letters Take A Half-Day Per Person To Draft
Promotion approved, bump approved, the comp number is final — now the manager has to write the reward letter. They open last year's template, retype the new numbers, swap in personal context, and check tone. Five letters across a team is a half-day's work that should be 20 minutes per letter with a grounded draft.
Continuous Feedback Stays In DMs Instead Of On The Record
Real praise and real concerns happen all quarter long — in a Slack DM, a hallway conversation, a Friday retro. None of it makes it onto the performance record, so the review reads like a 90-day memory test for the manager, and the employee experiences ratings that aren't tied to anything they remember happening.
AI Performance Cycle At A Glance
AI Performance
Review cycles. Goal analysis. Team risk.
Inside AI Performance Cycle — The Actual Capabilities
Every block maps to a real tool the agent runs against your Performance data — covering both the IC ("what's my review status?") and the manager ("how is my team tracking?"). Every write — sending feedback, starting a review, adding reviews to a cycle, building or customizing a review template — waits for your confirmation.
Reviews And Goals — For You And Your Team
View active and past reviews, current goals, progress, and details — for yourself, and for direct reports if you're a manager. The agent surfaces what review you're in, what goals are tied to it, and where each goal stands.
view_reviews+get_review_details— past and active reviews with cycle dates, status, calibration.view_goals+get_goal_details— your Performance Management goals with status and progress (not OKR Hub objectives).view_team_reviews+view_team_goals— manager-scoped views across direct reports.analyze_okrs+score_goal— AI analysis of goal health, and an AI score for one goal; neither writes progress back.
Feedback, Templates, And Reward Letters — Gated Writes
The write actions cover the heaviest performance-cycle moves: sending feedback to a colleague, starting a review (created at once, not as a draft), adding existing reviews to a cycle, and building or customizing review templates. Each waits for explicit confirmation; template and cycle writes are admin-only. Asking for a reward letter opens the compensation screen for that employee, where the letter is written.
- Confirmed writes —
submit_feedback,start_review,add_reviews_to_cycle,ai_generate_review_template,customize_template. view_my_feedback+view_team_feedback— feedback received, item by item: named feedback shows who gave it, anonymous feedback never does.list_review_templates+get_review_template— admins browse the review templates before starting a review.- Anonymous stays anonymous — feedback given anonymously reads "Anonymous" in every answer; the agent never names its giver.
Team Risk Summary — Catch Issues Before Calibration
The most distinctive AI Performance Cycle capability for managers — a single tool that rolls up risk signals across the team: overdue and soon-due reviews, overdue goals, high-priority goals under half done, recent constructive feedback and feedback still awaiting acknowledgment. Surfaces what to coach now, not what to explain in calibration later.
view_team_risk_summary— one tool that scores each person from their reviews, goals and the last 90 days of feedback, and ranks them critical, high, medium or low.- Ranked by risk — managers ask "how's my team tracking?" and see the people with elevated risk first; low-risk people are left out unless asked for.
- Goes deeper on demand — drill into a specific report's goals, reviews, feedback received.
- Calibration prep, not calibration surprise — risk surfaces weeks ahead of committee, not in the room.
Outcomes Teams Can Measure
The agent's job is to compress performance-cycle coordination and shift manager attention from past-reconstruction to forward-coaching. Measure against your pre-agent baseline.
- Review-cycle completion rate — share of self-reviews, manager reviews, and peer feedback submitted on time.
- Goal mid-cycle check-ins — frequency of goal touchpoints between set and review (vs the "set once, look at review" pattern).
- Peer feedback submission rate — share of peer-feedback requests submitted before the deadline.
- At-risk team members caught early — share of low-rating outcomes that the team-risk-summary tool flagged 4+ weeks before calibration.
- Manager prep time — hours spent reconstructing employee context before each review, before and after the agent.
Confirmed Writes, Role-Aware Surfaces
Most of AI Performance Cycle's tools only read — reviews, goals, feedback, team views, goal analysis and scoring, review cycles, templates. Its writes — sending feedback, starting a review, adding reviews to a cycle, building or customizing a template — all wait for explicit confirmation.
- Confirmed writes —
submit_feedback,start_review,add_reviews_to_cycle,ai_generate_review_template,customize_template. - Templates are admin-only — browsing, building and customizing review templates is for business admins, HR admins and Performance Management app admins.
- Anonymous feedback stays anonymous — named feedback shows who gave it; anonymous feedback never does.
- Audit trail on every action — read or write, every tool call logs the requesting user and the tool used.
WHAT TEAMS TRY INSTEAD
The four alternatives — and why none of them respect anonymity, role gates, and cross-app context at once
Most HR and people leaders reach for one of these four. None of them stick because none of them combine anonymous feedback that stays anonymous, admin-only templates, and live performance context in one place.
ChatGPT or Claude with a feedback snippet pasted in
General-purpose AI rewriting a draft review
- Reads the live performance record — reviews, goals and feedback — not a hand-typed bullet list
- Anonymous feedback stays anonymous at the tool layer — the agent shows "Anonymous", never the giver
- Confirmation-gated writes — submit, start, and generate-reward never fire from a casual chat ask
Lattice AI, 15Five Spark AI, Culture Amp AI Perform
Vendor-trapped AI inside the performance platform
- Works the cycle, not just the form — launches cycles, moves reviews off inactive reviewers, reminds reviewers before the due date
- Admin-only templates — only business admins, HR admins and Performance Management app admins browse, build or change them
- Same chat surface the rest of the team already uses — no separate performance portal nobody opens
A custom review tool in Notion / Confluence
A review template, a Slack reminder, a manager's slog
- Already shipped — no Notion template to maintain, no convention to enforce, no Slack-bot for nudges
- Peer-feedback anonymity enforced at the tool layer — not by hoping managers don't peek at the source doc
- One set of roles and permissions across reviews, feedback, OKR, and recognition — not four separate sources of truth
The manual fallback — annual review cycle + Q4 fire drill
A 90-day cycle that nobody enjoys
- Surfaces feedback continuously — no Q4 fire drill rebuilding context from a year ago
- Compiles a summary of each completed review from its ratings, comments and goals — not from a blank page at midnight
- Standardizes review voice, length, and language so quality doesn't depend on which manager owns it
PLATFORM ADVANTAGE
AI Performance Cycle inherits everything the platform already runs
A standalone performance platform has to plumb each of these. The agent gets them for free because the platform already does.
One chat for every app
Lives in the same Ask AI as every other app's agent, so the next question — about recognition, training or a schedule — goes to that app's agent in the same chat.
Anonymous feedback stays anonymous
Feedback given anonymously reaches the model as "Anonymous" — its giver is never in the answer. Named feedback shows who gave it.
Admin-only templates
Browsing, building and customizing review templates is admin-only — HR keeps control of the review format.
Confirmation-gated writes
Sending feedback, starting a review, adding reviews to a cycle and building or changing a template all wait for explicit confirmation — no silent review state changes.
Audit trail
Every read and write is logged with the requesting user and the tool; the review itself keeps its own rating history in Performance, and the AI request log keeps detail for 30 days.
Automatic model tiering
The platform assigns each agent one model tier suited to its work — set for the agent, not picked per question.
INDUSTRY FIT
Industries where embedded performance intelligence moves the most weight
AI Performance Cycle matters most where the review cycle is structurally complex and the people-data already lives across many apps.
B2B SaaS & Tech
Reminds reviewers before reviews fall due and compiles each completed review's summary, so engineering managers spend review week coaching, not chasing forms.
Healthcare
Moves reviews off reviewers who are no longer active and reminds the rest before the due date, so clinician reviews don't stall when a charge nurse changes units.
Manufacturing
Plant supervisors ask in chat which of their people are at risk — overdue reviews, slipping goals, recent constructive feedback — instead of pulling reports between shifts.
Retail
Cuts the district manager's review-cycle effort — cycles launch on schedule, reviewers get reminders before reviews fall due, and each finished review gets a compiled summary.
Field Services
Managers in the field ask in chat for their team's review status and goals — no separate review portal to log into.
Public Sector
Agent activity is fully audit-logged. AI requests go to the platform's configured model providers.
WHY MANGOAPPS WINS
An embedded performance agent beats a chatbot, a perf-platform add-on, or a custom build on every axis
The argument finance, HR, IT, and security all share — and the one a horizontal AI or single-vendor performance AI structurally cannot answer.
Cheaper than the alternatives
No Lattice or Culture Amp SKU, no per-seat ChatGPT, no six-month custom review-tool build — and no Q4 fire-drill weekends rebuilding context.
More secure
Anonymous feedback kept anonymous, admin-only templates, and tool calls logged with the requesting user mean review data and anonymity are enforced at the tool layer.
Easier to deploy
Already deployed if Performance is enabled. Turn the agent on, point it at the templates you already authored, and it's running the same day.
Easier to use
Lives in chat next to the work — no separate review portal, no annual login dance, no "where do I start the review" hunt.
Easier to manage
Per-business role gates and confirmation rules sit in the same admin console as every other app's settings.
Easier to extend
Shares the agentic tool framework with every other MangoApps agent. New review formats, new feedback sources, and new reward types ship as tools, not rewrites.
AI is actually better
A horizontal or perf-platform AI can rewrite a review bullet. AI Performance Cycle reads the real review, goal and feedback records, ranks a manager's team by risk, and works the cycle itself.
At our head office within HR, marketing, and finance, we use ChatGPT quite a bit, and MangoApps AI Helper is a useful tool for us. As I go around to the stores, I’ve been showing them how the AI Intranet Helper can find things and answer plain language questions about our resources, and they’re quite excited about that.
Alison van Zijl
Head of Human Resources
Full House
Customer Success
MangoApps Customers
Organizations that trust MangoApps to run every workflow across their workforce.
Customer Case Studies
Customer Case Studies
Customer Case Studies
Customer Case Studies
Customer Case Studies
Customer Case Studies
Frequently Asked Questions About AI Performance Cycle
Read reviews and their details (yours and your team's), goals and their details (yours and your team's), feedback received, a team risk summary, review cycles (and their insights, for admins), and corrective actions where that module is on; analyze goals and score one goal. Writes wait for your confirmation: send feedback, start a review, add reviews to a cycle, and build or customize a review template (admins). Asking for a reward letter opens the compensation screen for that employee, where the letter is written.
submit_feedback requires explicit confirmation — the agent shows the parsed feedback (relationship, type, content, routing) and waits for the user to confirm before it's submitted. The same applies to start_review, add_reviews_to_cycle, ai_generate_review_template and customize_template.
view_my_feedback and view_team_feedback return feedback item by item, each with its type, status and a short preview, and only to people its visibility setting allows. Named feedback shows who gave it; feedback given anonymously shows "Anonymous", and the agent never identifies its giver.
view_team_risk_summary scores each person on your team from their open reviews (overdue, or due within a week), their goals (overdue, or high-priority and under half done) and the last 90 days of feedback (constructive items and items awaiting acknowledgment). It ranks the team by risk level — critical, high, medium, low — and leaves low-risk people out unless you ask for everyone, so managers can coach now instead of explain in calibration later.
Review-cycle completion rate, goal mid-cycle check-in frequency, peer feedback submission rate, at-risk team members caught early, and manager prep time. Compare against your pre-agent baseline.
Let's Talk
Every team. Every employee. Every workflow. One AI platform — built for the frontline. Since 2008, trusted by 2 million+ users with an NPS of 78.
Why Choose Us?
- Frontline AI: Governed AI for every employee and workflow.
- Top Security: HITRUST, ISO & SOC 2 certified.
- Exceptional UX: Delightful on mobile and desktop.
- Proven Results: 98% customer retention rate.
Rolled out to every employee at AutoZone (125,000), PetSmart (50,000+), A.S. Watson and Raley's (20,000) — and at larger retailers we are not permitted to name.
Prefer to explore first? Ask AI about AI Performance Cycle →