Agile Analytics: Monte Carlo forecasting, flow metrics and DORA for Azure DevOps

Agile Analytics is an Azure DevOps extension for teams on Azure DevOps Services (cloud). It adds a hub to every project with Monte Carlo forecasting, cycle time and flow metrics, aging work in progress, DORA metrics, sprint reporting and an Exec Report. No Power BI, no personal access token, and nothing to host. Azure DevOps Server (on-premises) is not supported.

Open Azure DevOps, click Agile Analytics in a project's sidebar, and see your last 6 sprints' velocity, cycle time and a delivery forecast on one screen. AGAN AI, the helper inside the hub, checks your setup and answers "what is stuck?" with the views' own rules. Your work-item content stays in your Azure DevOps tenant. Every feature is on every plan, plans start at $300/year per organisation, and a 30-day free trial starts on install with no credit card.

See it in action

The Dashboard: cycle time bands, work in progress, flow efficiency, lead time, throughput and a 30-day "How many?" forecast on one screen.

Monte Carlo forecasting: 10,000 simulations on your team's own throughput show how many items you are likely to finish by a date, at 95%, 85%, 70% and 50% confidence.

Aging work in progress: every in-flight item plotted by days in flow against your team's own p50 and p85, then ranked by urgency.

Cycle time: every completed item on a scatterplot with p50, p70, p85 and p95 lines from your own history, plus a histogram view.

What you get

Monte Carlo forecasting: when will we be done?

10,000 simulations on your team's own throughput give P50, P85 and P95 delivery estimates, and a "How many?" forecast for a chosen date, so you can answer "when will the backlog be done" with a range instead of a guess. Teams that don't run sprints can forecast from weekly throughput instead.

Cycle time and flow metrics

An interactive cycle time scatterplot with P50 and P85 percentile lines, a cycle time heat map, flow efficiency, lead time, throughput and Cumulative Flow, for Scrum and Kanban teams side by side.

Aging work in progress and blocked items

See which items are getting old, and what is stuck, before it hurts. WIP limits and Teams or Slack webhook alerts are included.

Sprint reporting without Power BI

Sprint burndown, velocity and a sprint report: Dashboard KPIs, Sprint Summary, Sprint Capacity and Next Sprint Planning, all inside Azure DevOps.

DORA metrics for Azure DevOps

Deployment frequency, lead time for changes, change failure rate and time to restore, computed from your work items and pipelines.

Dashboard widgets and the Exec Report

Companion tiles in the widget catalog under the "Agile Analytics" category. The Sprint Forecast dashboard widget is free for every organisation, no licence needed. The Exec Report is a one-page leadership summary with forecast, risks, delivery trend and DORA at a glance. Share it as a PDF, as email text to paste into your email app, or as a copied summary. It is built in your browser and nothing is sent to us.

There are 30+ analytics views in a single project hub, with full dark mode.

AGAN AI: setup help and answers inside the hub

AGAN AI checks your setup, answers quick questions about stuck and aging work with the views' own rules, finds the right view for you, and answers typed questions about the product and your setup. Its Setup advisor prepares fixes you confirm, with Undo; organisation settings still need an Agile Analytics admin. AI answers are included in every plan with a monthly allowance per organisation.

Questions it answers

Azure DevOps has powerful raw data but no built-in answers for the questions that matter in sprint planning and retros:

• "When will we be done?" Monte Carlo simulation on your actual throughput history.
• "What should we commit to next sprint?" Next Sprint Planning combines capacity, Monte Carlo on the team's last 6 sprints, cycle time per bucket, and historical commitment reliability into a single recommendation.
• "Why is this sprint struggling?" Commitment reliability, scope churn, and carryover rates in one view.
• "How long does work actually take?" A cycle time scatterplot with P50/P85 percentile lines.
• "Are we flowing or are we stuck?" Flow efficiency, lead time, and WIP trends over rolling windows.
• "Which team needs attention?" A cross-team health grid with configurable goals and coaching prompts.

Who it is for

• Scrum Masters: live sprint health, commitment reliability, scope churn, and sprint capacity planning.
• Engineering managers: team benchmarks, cycle time trends, flow efficiency, DORA metrics, and investment allocation.
• Agile coaches: configurable health thresholds, prioritised action items, and a retro snapshot you can copy or export in one click.
• Product owners: Epic and Feature completion with child story rollups, and Monte Carlo forecast dates.

Why teams choose it: no infrastructure to run (no Azure Function, no database, no service connection), WIP alerts that reach you through a Teams or Slack webhook, role-based access control with Azure DevOps directory search, and retro-ready sprint summaries in one click.

★★★★★ "The graphs and charts are easy to read and understand and can be shown to any company member without having to give any further explanation. I would recommend for any team size."

"Even for a small team (we have 5 devs and a Scrum Master) this gives us quick and easy access to progress and planning. There is a great support team who actively engage with us and respond very quickly to our feedback, suggestions and ideas." Mark Pickering, Azure DevOps Marketplace review

"Finally, a tool that's fully integrated with ADO that gives a vast array of meaningful metrics and can be configured to meet your needs." John Roby, Azure DevOps Marketplace review

Your work-item data never leaves your Azure DevOps tenant

Every analytics view runs in your browser, against the Azure DevOps REST API, using your existing ADO session. Sprint contents, work-item titles, PRs, code, comments and attachments: none of it is sent to a Baytek server.

Agile Analytics installs as an Azure DevOps Marketplace extension: it lives in your tenant, your admin controls it, and your security team can audit it through the Marketplace publisher record.

Baytek does receive a small amount of licence, usage and operational data. The main items are under "What Baytek's backend receives" below, and our Privacy Policy lists every one. None of it is work-item content.

Four read-only ADO scopes. vso.project, vso.work, vso.graph, vso.build. No vso.code. No vso.analytics. No write scopes. The extension cannot modify your work items.

What stays in your tenant: Azure DevOps work item content and any AI adoption credentials you enter. Baytek never sees any of this.

What Baytek's backend receives: licence records (organisation name, plan, expiry, activation tokens); a lightweight install/heartbeat ping (organisation name, extension version, event type, timestamp); pseudonymous usage events (which view was opened, which feature was used, configuration completion, trial-day milestones), linked to your organisation and your Azure DevOps user GUID, never your name or email, and used to count active users and enforce seat limits (an extension admin can turn them off for the whole organisation at Configuration → Privacy); and, on Team/Business plans, who holds a seat, as opaque Azure DevOps identity descriptors and role, never names or email addresses. During the trial, the organisation administrator's Azure DevOps sign-in email is recorded once per organisation, shown in a banner with a short window to undo it, and used for trial reminders and service notices, each with a one-click opt-out link. To confirm which organisation a request comes from, the extension sends a short-lived Azure DevOps access token that our server uses for one check and then discards; our licence server also records the IP address and browser of licence requests for security and rate limiting. If you opt into Industry Benchmarks, that feature's aggregate-metric data reaches Baytek too. If you type a question to AGAN AI, the question, the earlier messages of that conversation and a short list of counts and codes about your setup (including Setup advisor numbers such as your WIP limit, with item types and states sent by position, not name) reach Baytek and its AI provider, Anthropic. Never work item titles, IDs, state names or people's names. No transcript is kept; Baytek keeps only a monthly total of AI usage per organisation. An extension admin can turn AI answers off in Configuration → Access Control, and each person can use AGAN AI: On / Off in the account menu. The Exec Report is built entirely in your browser and sends nothing to Baytek.

Optional integrations: when you connect GitHub Copilot for AI Metrics or set up webhook notifications, requests go from your browser directly to that source (api.github.com for GitHub Copilot) or webhook URL (Microsoft Teams / Slack), never via Baytek.

Accessibility: WCAG 2.1 AA achieved, checked by an axe-core regression suite that runs automatically every night against the latest code.

What's New in Version 6.65

v6.65.1 to v6.65.9: meet AGAN AI, a Setup advisor, Exclude from metrics and a clearer trial-ended screen (October 2026)

Meet AGAN AI. A helper panel inside the hub, opened from the blue "AGAN AI" button in the bottom-right corner, from Ask about this view, or from an "Ask AGAN AI why" chip next to a warning. It opens with suggested questions for the view you are on, and on ten views it can explain the numbers on screen, because a typed question carries that view's own figures. Is something wrong? checks your setup and Fix it here opens the right setting. Quick questions show which items are not moving, what is blocked and what is aging past normal, using the same rules as the views. Find a feature searches every view and Configuration tab in your own words. Type a question and the answer streams in, with buttons to the right view or setting. Each organisation has a monthly AI allowance, and the panel tells you when it is nearly used up.

Setup advisor. At the top of the panel, AGAN AI prepares setup fixes such as a WIP limit or unmapped states. Nothing is saved until you click Review and apply, then Confirm, and Undo puts the old setting back.

Exclude from metrics. In Configuration › Workflow Mapping, an admin can list states such as Cancelled or Won't Do so items in those states stay out of cycle time, throughput, aging, forecasts and velocity. Your mapping itself is never changed.

Smoother first setup and admin access. During a free trial, every member of the organisation gets straight into the hub. Workflow Mapping is ready to review the moment you first open it. Organisation administrators who hold their rights through a Microsoft Entra group are recognised as admins, and if your only Agile Analytics admin has left, an organisation administrator can take over from Configuration › License.

After a trial ends. The trial-ended screen shows the end date and what is paused, puts Need more time? Ask Henry next to Buy, and lets team members press "I want Agile Analytics back" so everyone can see how many people asked. A paused dashboard widget now links to the hub.

Your data. Setup checks, quick questions, Find a feature and the Setup advisor are worked out in your browser. Only a typed question goes to Baytek and Anthropic, our AI provider, with a short list of counts and codes about your setup and the numbers shown on the view you are on (percentages, day counts, item counts), never the items behind them. Never work item titles, IDs, state names or people's names. What we send, under the box, lists every field. Admins can turn AI answers off in Configuration › Access Control, and anyone can turn AGAN AI off with AGAN AI: On / Off in the account menu.

Every earlier release is in the full release history.

Getting started in under 2 minutes

1. Install the extension into your Azure DevOps Services organisation from this listing.
2. Open any project and find Agile Analytics in the left navigation bar under your project name. If you don't see it yet, refresh the page once. It appears in every project automatically.
3. Review Workflow Mapping under Tools → Configuration to check the auto-discovered states, so cycle time and flow metrics are accurate for your process.

Requirements: Azure DevOps Services (cloud), Project Contributor access or higher, and a modern browser. Azure DevOps Server (on-premises) is not supported in this version.

Pricing. A 30-day free trial activates automatically on install, with no credit card. Annual per-seat plans: Team: $300/year (2 seats), Business: $1,200/year (6 seats), Premium: $2,000/year (unlimited seats). Each licence covers one Azure DevOps organisation, and every plan includes every feature. See pricing.

Common questions

Do I need Power BI, Analytics views or a personal access token? No. Every view reads the Azure DevOps REST API in your browser, with the session you are already signed in with.

Does it work with my process? Yes: Agile, Scrum, CMMI, Basic and custom inherited processes. First-time setup reads your project's states, and Workflow Mapping under Tools → Configuration lets you check where each state belongs.

What counts as a seat? During the trial there are no seats: everyone in your organisation has full access. On a paid plan, your extension admin assigns seats in Configuration → Access Control and can swap them at any time.

What happens when the trial ends? The trial-ended screen shows the date the trial ended and what is paused, with a field for your licence key. Nothing is deleted and your settings stay as they are; entering the key restores access at once. Admins see Buy and Need more time? Ask Henry, and team members can press "I want Agile Analytics back" so their admin sees how many people asked. The Sprint Forecast dashboard widget keeps working for free.

Is there a free option? Yes. Monte Carlo Sprint Forecast is a separate, free extension that adds a next-sprint forecast tile to your dashboard, with no account, trial or licence key.

Learn more

Full documentation, install guide and support live at ado-analytics.baytekdev.com.

Getting Started Guide  |  Support & Feature Requests  |  Privacy Policy  |  Terms of Service