How Wonderlic cut response time from 18 hours to under a minute.

Wonderlic is a US-based assessment platform that has been in market for decades, with a product line covering cognitive ability assessments, personality testing, and pre-employment screening. The customer base is enterprise HR teams and recruiting operations. The product is the kind of thing that gets used in production hiring workflows under time pressure, which means the support load is real. A candidate halfway through an assessment hits a technical issue at 11pm on a Sunday. A recruiting team needs to verify a result before a 9am Monday interview. The window between a ticket arriving and a response landing is the window where the customer experience either holds or breaks.

The structural problem at Wonderlic was the after-hours coverage gap. The support team ran a US business-hours model. Tickets that came in after 5pm or over the weekend sat in the queue until the next business morning. The average after-hours response time was 18 hours. That number was an average. The 90th percentile was worse. Customers in the Pacific time zone or international markets felt the latency the most. Internal sentiment among the support team was that the queue was a constant source of weekend dread, with nobody happy about working a weekend and nobody happy about leaving the queue closed.

The deeper problem was the cost shape of the fix. Hiring a 24/7 support team to close the coverage gap would have meant staffing overnight and weekend shifts across at least two time zones. That is four to six additional hires loaded against a function that was already operating at the upper end of what the unit economics would support. The alternative was outsourcing the overnight queue to a BPO that did not know the product, did not have access to the historical knowledge base, and would have been answering tier-1 questions with scripts written six months out of date. Both fixes were structurally bad.

Wonderlic came to EOI because the assets the support team had built over a decade were sitting unused most of the time. The internal knowledge base was deep. The help-desk archive contained answers to almost every tier-1 question the product had ever produced. The macros library was extensive. The problem was that those assets only got applied during business hours when a human was on the queue to pull them. Every tier-1 ticket that came in at 11pm was being answered the next morning by a human who had to search the same KB the previous tickets had already been answered against. The labor was not the bottleneck. The coverage hours were the bottleneck.

The shape of the fix that worked was an AI Support Department trained on the full historical knowledge base, with the operating authority to answer tier-1 tickets directly during the hours no human was on the queue. The agents handle the password reset, the assessment access issue, the report download problem, the basic configuration question, in seconds. The human reps see the queue every morning already worked, with the tickets that needed a person escalated cleanly. The deeper case for this kind of support consolidation is in What is a Fractional AI Department and the matching department breakdown sits at AI Support Department.

The other reason the fit was clean was the cost shape. The fractional model is one monthly retainer regardless of queue volume. The overnight queue is the same cost on the invoice as the daytime queue. The cost of closing the 18-hour coverage gap collapsed to a fraction of what hiring or BPO would have cost. The economics finally matched the customer experience the product needed.

The headline result is the one the testimonial captures. After-hours response time dropped from 18 hours to under a minute. That metric is the one the customer feels. A candidate hitting an assessment issue at 11pm gets the answer before they close the laptop. A recruiting team checking on a result over the weekend gets the answer before the Monday meeting. The customer experience that used to break at 5pm and stay broken until 9am the next business day stopped breaking. The CSAT improvement followed almost immediately.

The internal result is the one the support team feels. The morning queue stopped being a wall of accumulated overnight tickets. The human reps now arrive to a queue that is already worked, with the escalation tickets surfaced and ready to handle. The work that humans do during business hours is the work that humans should be doing. The empathy cases, the billing edge cases, the customer success conversations that build retention. The tier-1 grind that was consuming most of the team day moved to the agent layer.

The institutional result is the one that matters for the long-term operating model. The decade of help-desk knowledge that Wonderlic had built up over the years finally became a strategic asset instead of a static document. The KB is alive now. Every resolved ticket either matches a KB entry or generates a new one. The institutional knowledge compounds week over week instead of getting stuck at whatever state the previous senior reps left it in. New product launches get their KB built into the agent layer the same week they ship. The support function moved from a static cost center to an asset that compounds.

The deeper structural lesson is what the testimonial frames. The cost of running 24/7 multilingual tier-1 support used to be the cost of staffing it. Hiring four to six additional reps to close the coverage gap. Or outsourcing to a BPO that did not know the product. Both were structurally bad. The fractional AI Support Department fixed the coverage gap on a single monthly retainer smaller than a single full-time hire would have cost, with the institutional knowledge applied to every ticket because the agents work the same KB the senior reps wrote. The same pattern works at every SaaS and platform business we work with, which we have detailed at 24/7 AI Customer Support and the matching AI Support Department page.