IntouchCX says Superpunch Roleplay improved first-month CSAT by 7%, raised QA scores by 4%, and helped agents reach AHT targets by week 11, and a 2026 AI Excellence Award adds independent recognition to the story, but the public case still leaves out cohort sizes, baseline scores, and program costs buyers would need for a firm ROI call.
The 7% CSAT improvement published by IntouchCX deserves attention, but the number also needs context. The May 2026 case study describes a major US retailer replacing static scripts with AI customer personas that could respond differently as each conversation unfolded. Superpunch Roleplay has also already picked up a 2026 AI Excellence Award from the Business Intelligence Group, giving the IntouchCX AI customer experience story a tidy awards hook.
That recognition came with named executive commentary worth weighing alongside the raw CSAT figures. Russ Fordyce, Chief Recognition Officer at the Business Intelligence Group, said in IntouchCX’s March 24, 2026 award announcement, framed the win around the maturity of the category itself:
“AI has arrived! 2026 is about execution, accountability, and results.”
One caution: the award quote is boilerplate. The Business Intelligence Group used the exact same wording in its 2026 AI Excellence Award announcements for at least six unrelated winners, including Wolters Kluwer, ActivTrak, Cytora, SIB, and Sagility. It tells buyers very little about Superpunch and doesn’t independently verify the CSAT or QA numbers. The trophy is nice. What changed for agents is the part that matters.
TL;DR: What Buyers Need to Know
- Product fit: Roleplay is an agent-training capability inside the broader IntouchCX Digital CX story. Its published outcome evidence is voice-centered, not proof of end-to-end omnichannel performance.
- Evidence: IntouchCX reports better CSAT, QA, customer effort, and AHT across two customer programs, but customer names, cohorts, costs, and statistical testing remain undisclosed.
- Verdict: Run a 60- to 90-day pilot with baselines, usage targets, live QA comparisons, and a cost-per-proficient-agent calculation.
How Does Superpunch Roleplay Fit Into IntouchCX Digital CX?
Superpunch Roleplay handles the practice part of IntouchCX’s contact center offer. Agents rehearse difficult voice calls, get scored right away, and take another run before speaking with customers. IntouchCX Digital CX covers the wider operation, including voice, chat, social, SMS, email, self-service, agent assistance, analytics, and quality management.
Sidd Spark and Laivly support live agents, Catapult analyzes performance, and Vision combines CRM, Catapult, and Superpunch data. Together, they can create a loop: practice, production, analysis, coaching, and another attempt.
That loop addresses a real problem. Zendesk reported in 2026 that 72% of CX leaders believed they had provided adequate generative AI training, while 55% of agents said they had received none. Jason Rosser, EVP of Solutions and Operations Strategy at IntouchCX, said in IntouchCX’s March 24, 2026 AI Excellence Award announcement Roleplay lets agents “practice, get real-time feedback, and improve” before customer contact.
Cresta raised an awkward question for IntouchCX in July 2026. Its simulator can turn real customer calls into practice scenarios and mark them against live QA rules. Buyers should find out whether Roleplay can do the same with their own policies, accents, customer behavior, call types, and scorecards.
Key Takeaways
- Roleplay addresses agent readiness; Digital CX covers the wider contact center and channel environment.
- The public evidence tests voice training, not performance across the full omnichannel journey.
What Outcome Does IntouchCX Claim For Superpunch Roleplay?
IntouchCX reports that Superpunch Roleplay users generally performed better on customer satisfaction and quality measures, while the handle-time advantage varied by deployment. The public evidence covers two customer programs and three writeups, but missing cohort sizes, baselines, selection rules, costs, and statistical testing prevent buyers from treating the results as a universal ROI claim.
| Public evidence | Reported result | What remains unclear |
|---|---|---|
| Retailer case, January 2026 | Week two: 18.9% higher CSAT and two-minute lower AHT. Across eight weeks: more than 3% higher CSAT and a 53-second average AHT advantage. | Customer name, cohort size, selection rules, starting scores, statistical testing, and program cost. |
| Home technology case, March 2026 | 85% customer effort score for users versus 80% for peers. AHT improved by 216 seconds for users versus 208 seconds for non-users. Usage rose from 9% to 18%. | Cohort size, frequency of use, scenario count, cost, and whether motivated agents were more likely to participate. |
| Retailer summary, May 2026 | 7% higher first-month CSAT, 4% higher QA, and AHT targets reached by week 11. | The comparison baseline, control-group timing, sample size, and how this summary relates to the January retailer writeup. |
The January results provide the most convincing evidence because they show the advantage narrowing across eight weeks rather than presenting one dramatic headline figure. However, IntouchCX doesn’t explain how those numbers connect with the May summary.
The 2026 AI Excellence Award and IntouchCX’s eight Asia-Pacific Stevie Awards add external recognition. They show that judges found the submitted work credible, but they don’t independently validate the customer data.
A separate awards submission complicates the picture further. IntouchCX’s TITAN Business Awards entry for Superpunch reports a different set of figures for what appears to be the same platform: 55% faster time-to-proficiency, 18% higher CSAT among AI-trained agents, tNPS increases up to +27%, and a 4.2% First Contact Resolution improvement. None of those numbers match the 7% CSAT, 4% QA, or week-11 AHT figures in the case study this piece is built on, and IntouchCX doesn’t reconcile the two sets anywhere public.
Buyers can reasonably conclude that Superpunch Roleplay deserves a controlled pilot. They can’t assume every deployment will produce a 7% CSAT increase or the same AHT improvement.
Key Takeaways
- Two deployments point toward better customer and quality results, but the size of the improvement changes considerably.
- Missing cohort, baseline, cost, and selection details prevent a firm ROI conclusion.
How Does Adoption Affect AI Training ROI?
Adoption decides whether AI training changes production performance or becomes a demo. Buyers should measure practice frequency, score improvement between attempts, and whether those gains appear in live QA, CSAT, customer effort, transfers, repeat contacts, and handle time. One login is not meaningful adoption.
The March 2026 pilot doubled usage from 9% to 18%, but 82% of eligible agents still didn’t use Roleplay. There’s also a self-selection problem: the people who used it most may already have been the keenest agents.
Salesforce’s May 2026 survey of 3,075 service professionals found that 70% of organizations using AI service agents measured value within 60 days. It isn’t a Roleplay comparison, but it does support a short test window. Add licensing, setup, scenario writing, trainer time, calibration, security checks, and management work, then divide the total by the number of agents who meet the agreed production standard.
Key Takeaways
- Usage doubled from 9% to 18%, but most eligible agents still didn’t participate and the results may favor those who were already more engaged.
- ROI must connect repeated practice to live outcomes and the full cost of agent readiness.
What Should Buyers Test Before Buying Superpunch?
Buyers should test whether Superpunch Roleplay reflects customer journeys, connects with production quality data, and fits the commercial model. A voice simulation is useful, but omnichannel buyers need evidence that training can cover the behaviors, policies, and handoffs found in chat, email, messaging, social, and self-service escalations.
- Channel realism: Can scenarios use the buyer’s own voice, chat, email, social, and messaging interactions?
- Production alignment: Are simulations graded with the same criteria used in Catapult, Vision, or the existing quality platform?
- Governance: Who approves scenarios, calibrates AI scoring, corrects bad feedback, and controls transcript and agent data?
- Operational fit: Which languages and accents are tested, how quickly can policies change, and can managers export results?
- Commercial fit: Is Roleplay available separately, what is the minimum deployment size, and what are setup, maintenance, and usage costs?
Cresta now claims it can build simulations from actual customer calls and grade them using the contact center’s current QA rules. The feature isn’t proof of better performance, but it gives buyers something specific to test Roleplay against.
Key Takeaways
- The main omnichannel gap is evidence that training gains transfer beyond voice.
- Data, calibration, integrations, pricing, deployment size, and scenario maintenance belong in the pilot plan.
Has IntouchCX Proven the Value of Roleplay for Contact Center AI Training?
Superpunch Roleplay has enough evidence to justify a test, especially for teams struggling with inconsistent onboarding or limited trainer capacity. Two 2026 deployments point to better customer outcomes among users, though the handling-time benefit changes sharply by case and the public record still lacks the cost and methodology needed for a firm ROI call.
The product idea is sound. New agents need somewhere safe to make mistakes, recover, and try again before a real customer pays the price. The strongest part of this contact center AI training model is the feedback loop: practice, scoring, another attempt, then comparison with live QA. That gives managers something more useful than a polished demo built around a perfectly behaved synthetic caller.
Still, the commercial evidence is limited. IntouchCX hasn’t named the customers, published cohort sizes, or shown what the contact center training simulation costs to configure and run. It also hasn’t demonstrated a clear small-to-medium deployment.
For Intouch Digital CX buyers, the sensible next move is a time-boxed pilot with agreed baselines, adoption targets, and a cost-per-proficient-agent calculation.
FAQs
What is Superpunch Roleplay?
Superpunch Roleplay is an AI agent training tool that lets contact center employees rehearse live-style conversations with simulated customers. The platform changes its responses as the exchange develops, then scores the agent against company-specific communication and process rules. It gives managers a more repeatable way to test readiness before new hires enter the queue.
Does Superpunch Roleplay improve contact center performance?
The published IntouchCX cases show stronger customer and quality outcomes among users, which makes the product worth testing. They don’t establish a universal result. The effect changes between deployments, and IntouchCX hasn’t released enough information about group size, starting performance, selection rules, or cost to support a guaranteed return from contact center AI training.
How should buyers measure AI training ROI?
Buyers should connect simulation activity with live performance. Track repeat usage, score improvement, time to proficiency, customer satisfaction, quality, transfers, repeat contacts, and trainer hours. Then calculate cost per proficient agent. A contact center training simulation earns its budget when better practice leads to better production results, rather than higher scores inside the training tool alone.
Does Superpunch Roleplay support omnichannel contact center training?
The published Superpunch Roleplay evidence focuses mainly on simulated voice conversations, so buyers shouldn’t assume the same results apply across chat, email, SMS, social media, or self-service. Intouch Digital CX supports a wider omnichannel customer experience portfolio, but buyers should ask which Roleplay channels are available and whether each one has been tested against live performance data.
What should buyers include in a Superpunch Roleplay pilot?
A Superpunch Roleplay pilot should start with defined scenarios, agreed performance baselines, minimum usage targets, and a comparison group. Buyers should also test AI scores against experienced trainers and connect simulation results with live CSAT, QA, AHT, transfers, and repeat contacts. The final calculation should include configuration, supervision, maintenance, and trainer time, not only software costs.
