A streaming company had more than a hundred people working in quality assurance. What it lacked was something you might carry to the supermarket: the right local payment card. Its expansion into new countries exposed a limit that another row of office desks could not easily fix. Testing a subscription meant reaching the financial habits and payment systems of people who actually lived there.
- Applause assembles and manages testers for real customer conditions.
- Its specialties include payments, accessibility, functional testing and AI.
- The lesson to borrow: recruit for the gaps in your coverage.
A hundred testers, one missing wallet
Applause’s account of that unnamed streaming customer describes a practical arrangement. In-country testers exercised payment flows twice daily ahead of launch, with one-to-two-hour reporting windows. After releases, they signed up and made small purchases. The company’s internal staff handled backend checks, including invoices and reimbursements.
The assignment expanded from one content channel to four. That progression matters: the buyer increased the scope after seeing the initial work. The missing ingredient was access to local conditions, not simply more people. A large team can still have a narrow view.
Same app.
Different test.
The name got bigger
Doron Reuveni and Roy Solomon founded the business as uTest in 2007. The proposition was straightforward: connect software teams with independent people who could find defects. In 2014, uTest became Applause, while the community kept the original name.
Reuveni later explained that the platform and community could contribute across the software lifecycle. Customers needed user feedback, payment validation and accessibility studies alongside bug reports. The broader name followed that broader job. That same year, Applause acquired Germany’s Testhub, extending its European presence.
Expansion took capital. A $43 million Series E in January 2014 was followed by a $35 million Series F in September 2016, bringing disclosed funding above $115 million. Vista Equity Partners announced its acquisition agreement in August 2017. These figures describe financing, rather than what a customer pays for testing.
The crowd needs a conductor
The word “crowd” can suggest an open invitation and a pile of incoming complaints. Applause sells a managed process. A dedicated team defines the testing objectives, selects participants, runs the work and filters findings into results engineers can use.
Its functional testing includes structured checks, exploratory testing and regression testing. Automation handles repeatable work; human testers investigate situations that a script might never visit. The company also offers UX research, customer journey testing, localization and security testing. Its platform connects testing activity to existing development workflows.
The customers are businesses whose digital experiences have many moving parts. Public examples include Audi, Microsoft, Cisco, Western Union and Peet’s Coffee. A money-transfer service needs confidence across payment routes. A collaboration product needs to work for people with different access needs. “Works on my machine” is a painfully small commercial promise.
Reported by Applause in February 2025. A recurring business metric, rather than a disclosed annual revenue statement.
That recurring revenue milestone suggests the attraction extends beyond a single launch. Products keep changing, and coverage needs to change with them. Buyers engage Applause for an organized testing capability, with commercial scope worked out through sales.
A different kind of access
Applause says its community includes more than 1.5 million independent testers and end users across more than 200 countries and territories. That is a recruiting pool, not a claim that every member tests every release. The operational question is how well a selected group matches the task.
Accessibility makes the distinction especially clear. The service combines expert assessments, feedback from people with disabilities, design reviews, fix verification and training. Someone who regularly uses assistive technology brings experience that an automated scan alone cannot supply.
In October 2026, Applause reported that Progress ShareFile had reduced critical audit bugs by 67%. The described work included earlier design reviews, assistive technology users, AI-powered code evaluation and ongoing training. It is a specific customer result, rather than a promise that every engagement produces the same reduction.
There are alternatives. Testlio also provides managed crowdtesting. Internal QA, traditional outsourcers and automated tools cover overlapping needs. Applause’s case rests on its combination of managed execution, community reach and specialist services. Access earns its keep when it reveals something useful.
Now the software takes action
The same logic has carried Applause into AI evaluation. A chatbot can produce a convincing answer that is wrong. An agent can go further and execute the wrong action. Checking the wording alone leaves part of the product untested.
Applause’s agent testing examines tool use, workflow completion, retrieval quality and correctness across multiple steps. Its broader AI services include expert review, automated evaluation, red teaming and data collection. The company’s current approach combines human judgment with model-based scoring.
“AI is fundamentally changing how software is built, but it hasn’t changed the standard for quality.”
Aatish Salvi · CEO, Applause · July 2026
Salvi became CEO in July 2026, with Tacita Morway promoted to CTO. In October, Applause announced plans to demonstrate its AI evaluation services at the Dubai AI Festival. The business is adapting its original question to a new product category: what happens when this reaches somebody outside the team?
Borrow the recruitment brief
A product team can copy that question without recruiting a million people. Choose one important journey. List the conditions your colleagues cannot reproduce. Recruit someone who has them. Ask for evidence of what happened, then fix and retest.
This approach needs boundaries. A poorly chosen crowd can repeat the team’s blind spots. Unclear instructions produce findings that are difficult to interpret. Restricted systems need suitable access arrangements. Engineers still need time to repair defects. Human exploration complements code-level checks and automation; a sampled test cannot prove an AI system safe in every situation.
Start with the person your current process leaves out. They may have the missing card.