Alloy release round-up: June 2026

Featuring our infinitely configurable backtesting suite and workforce analytics

Our new backtesting tools are now generally available! We’ve had testing features for different pieces of our product forever, but we set out with the lofty vision of taking a decade of workflow testing experience and remaking a fully new and improved backtesting suite. And it is now shipped to everyone!

Alongside it: a copying-and-pasting workflow piece, a new “Workforce Analytics” dashboard that turns your operations data into actionable insights, and more!

Launch: The Alloy Backtesting Suite

Policy changes carry risk, and we’ve always had feature requests for more tools and safety measures to give users confidence in setting new versions of their policies live. With Alloy’s new Backtesting Suite, clients can remove the guesswork from threshold tweaks and run some tests!

You can now replay historical events (in a highly configurable way) through a draft version of your workflow to see exactly how a rule update or threshold change would have performed before it touches production. Results are fully excluded from live queues, analytics, and billing. There's no risk in testing, so test to your heart’s content!

You can run a backtest against Ongoing workflows (Transaction Monitoring, Login Monitoring, Account Change) and Onboarding workflows (Person, Business). When you're ready to test, pick from four ways to define your sample:

• Compare to another version: the recommended starting point. Select an existing workflow version as your control; the same events that ran through it get reprocessed through your new version for a clean apples-to-apples comparison.
• Rerun a past test sample: useful when you want to reproduce earlier results or compare multiple variations against the same dataset.
• Filter to a custom date range (Events API clients): define your population by date; useful during implementation if you've backfilled historical data or want to focus on a specific period, such as a known fraud spike.
• Specify by evaluation token (Onboarding workflows): enter specific tokens to confirm that a rule fires on the exact event or entity you expect it to catch.

Results show outcome distributions for both versions side by side, with percent-change indicators per category, and a per-evaluation breakdown with rule explainability, so you can see exactly which rules fired and why.

If your workflow includes third-party services, backtesting uses cached historical responses from when those events were originally evaluated. It doesn't make new live calls, so results reflect how the workflow would have behaved at that point in time.

For teams subject to model risk management or change governance requirements, every backtest creates a reviewable record of your pre-deployment testing. This makes for a cleaner audit trail instead of toilsome manual spot-checking or spreadsheet comparisons.

Launch: Workforce Analytics

Managing a review team means constantly asking the same questions: How many cases are in the queue right now? Is my team processing them fast enough? Who’s carrying the load? Where are things getting stuck?

Traditionally, getting those answers meant submitting a request and waiting for a manual report. The process worked, but it often meant the data had aged by the time it arrived. And since it lived outside of your case management system, exploring bottlenecks required a few extra steps.

Until now!

Workforce Analytics is a built-in dashboard that gives Agents and Operations Managers a live view of their team’s performance and queue health, directly inside the platform. No ad hoc requests, no waiting.

What you’ll see:

• Top KPIs: your key numbers up front, like open cases in the queue right now, cases closed today, your team’s median completions per day (overall and per agent), and median time from assignment to close. Each KPI deep-links into the relevant detail view so you can go from headline to specifics in one click.
• Volume trends: how many cases came in and got processed by day, week, and month, with an avg/median toggle. Understand whether your team is keeping up, falling behind, or clearing the backlog, and whether your closure rate is trending in the right direction over time.
• Queue composition: break your open cases down by status, type (Fraud, Compliance, etc.), or escalation state. See how long those cases have been sitting in the queue, in hourly buckets up to 48+ hours, or daily buckets up to 8+ days, so you can spot where cases are aging and where to prioritize.
• Agent performance: a side-by-side table of every reviewer’s output, including completions today, completions this week, outcome breakdowns by decision type, and both average and median time per case. Useful for spotting coaching opportunities, balancing workload, and understanding who your fastest reviewers are and why.

The dashboard lives inside two tabs you may already know: Onboarding Analytics (for Application queues) and Ongoing Analytics (for Investigation queues). Workforce data sits right alongside the workflow metrics you’re already using.

✨ Bonus for AI Assistant customers: ✨ If you have Alloy’s AI Assistant enabled, the AI agent’s case throughput automatically appears in the Agent Performance table alongside your human reviewers. No extra setup required! It’s a direct, apples-to-apples comparison of AI and human output in the same view and one of the clearest ways to see the ROI of automation on your review team.

Workforce Analytics is now available to all customers using Alloy Journeys (Applications or Investigations). Reach out to your account team if you don’t have access yet.

Copy logic across workflows (or even across totally different accounts)

To all of our Alloy policy-writing super-users out there: you know that cloning a workflow into another account can sometimes make configuration cleanup challenging. It should always “just work.” 

Now it does! No matter how complex the copy-paste, you’ll get intuitive prompts to help reconcile any custom config, making it fast and seamless to copy policies between accounts/environments.

Also shipped in June 2026

New

• More in-depth event logs, including developer-first error exploration. A new column surfaces 4xx and 5xx responses in the Events Log (compatible with Events API-integrated clients), with filtering to zero in on what went wrong without having to dig through noise. Request and response headers and bodies are also now inspectable directly in the dashboard. We’re working on bringing this to all API requests and responses, so stay tuned.
• Fraud Attack Radar filter on the evaluations queue. Easily filter the evaluations queue to the specific group of people that triggered the anomaly alert to investigate a targeted attack pattern.
• Thomson Reuters HVIDC. A new sub-second identity verification option. Contact your account team for access.
• Plaid Layer SDK plugin. Now you can leverage Plaid’s consumer prefill solution to reduce onboarding friction with our new integration and SDK plugin.
• Equifax Canada bureau PDF. Bureau print documents from Equifax Canada are now retrievable as Alloy documents via the API.

Improvements

• ComplyAdvantage Mesh now supports multiple countries on a single request.
• NeuroID now returns a consistent top-level status field across integrations.
• Web SDK's alloyRequestFields automatically map behavioral and device plugin fields — enabling a new plugin to avoid requiring a manual field-mapping update.
• Search field filters truncate gracefully instead of overflowing.