Customer experience is entering a new phase—where simply analyzing interactions is no longer enough. Organizations are under pressure to translate insights into measurable improvements, especially as AI reshapes how customer conversations are handled. The challenge lies in bridging the persistent “insight-to-action gap” across contact centers, sales, and support environments.
As a Co-Founder and CEO at Solidroad, Mark Hughes operates at the intersection of AI, performance optimization, and customer experience. His work focuses on enabling organizations to continuously improve both human agents and AI systems. That by using automation to scale coaching, quality assurance, and training.
What is Fundamentally Missing?
Q1. Many organizations have invested heavily in CX analytics, yet struggle to see meaningful improvement. From your perspective, what is fundamentally missing in how CX is approached today?
MH: What’s missing in CX today is the proper infrastructure. Most companies have loads of data. ut they’re missing the piece they need to actually see what’s going on at scale.
Companies are only looking at 1-2% of their customer conversations. And are getting insights and basing decisions on this small subset. With most conversations never being evaluated, coaching and quality improvement insights aren’t fully informed.
Also a lot of companies are only evaluating agents on metrics like response time or ticket counts. These measurements can’t be used to improve the customer experience though. For improved customer satisfaction companies need to look at quality measures. But our recent State of CX survey showed that around 81% of support conversations are never reviewed for quality. This is leaving the vast majority of customer interactions in a black box. Without consistent feedback, agents rarely improve, customers receive inconsistent support, and the quality gap worsens over time.
High-performance CX
Q2. How do you define “high-performance CX” in an era where both human agents and AI systems are interacting with customers?
MH: High-performance CX is really about consistency and quality. Every single conversation, whether it’s handled by a human or AI, needs to reflect the brand accurately and actually resolve the customer’s issue.
What makes that especially difficult right now is that AI agents can be handling thousands of conversations at the same time, so if something’s going wrong, it’s happening at scale. On top of that most customer interactions aren’t even being evaluated for quality, so these issues can go undetected for a while.
We recognize how important quality oversight is, which is why we’ve just raised a $25 million Series A to help companies with this. Our goal is to give them full visibility into every interaction and turn insights into action for both their human and AI agents, so they can scale their customer support capabilities while still maintaining quality.
Move from Siloed CX Initiatives
Q3. What does it take for organizations to move from siloed CX initiatives to a truly integrated, performance-driven CX strategy?
MH: Teams might be reviewing a small sample of conversations, tracking metrics in dashboards, and running coaching programs separately, but there’s often no connective layer tying those pieces together. Insights live in one place, coaching happens in another, and neither consistently feeds into how teams actually improve. That makes it very difficult to drive consistent, measurable progress in performance.
What it takes to move beyond that is shifting from disconnected workflows to a system where insight and action are directly linked. That means having a complete, objective view of customer interactions across both human and AI agents, and more importantly, ensuring those insights immediately translate into coaching, training, and process improvements.
Operationalize CX Insights
Q4. In your experience, where do organizations typically encounter friction when trying to operationalize CX insights across teams?
MH: The biggest friction point is the gap between surfacing an insight and knowing what to do with it. A team might identify that human agents are underperforming or that an AI agent is making errors, but without a direct path to fixing it, that insight ends up just living in a report somewhere.
The other major friction point is scale. Even when teams know what needs to change, traditional coaching and training processes aren’t built to move fast enough or cover enough people. Reviews are manual, feedback is delayed, and fixes aren’t applied fast enough. By the time insights are translated into action, hundreds or thousands of similar interactions have already happened, making it nearly impossible to keep quality consistent as volume grows.
Bridging Insights with Action
Q5. Solidroad focuses on bridging insights with action. How does AI enable this shift from passive analytics to active performance improvement?
MH: When we first started Solidroad, we noticed a gap in how customer experience teams traditionally operated. We’ve seen a lot of effort go into identifying issues in customer interactions, but scoring and flagging a problem is only half the job. The most important step is what happens next, after you’ve identified the issue.
At Solidroad, we automatically generate customized simulations so that agents can practice and improve before the mistake gets repeated in live customer interactions. The same goes for AI agents. We turn those insights into process updates that lead to better resolution rates. Our AI systems make it possible to do that across 100% of conversations, rather than just the small sample a team could manually get through.
It also addresses a common challenge: our recent survey showed over half of agents say applying training to real customer situations is the hardest part of their onboarding. By connecting insights directly to practice, we make sure lessons actually translate to improved performance on the job.
Balance Between Human Agents and Automation
Q6. As AI becomes more embedded in customer interactions, how should organizations rethink the balance between human agents and automation?
MH: I’m a big believer that the future is hybrid. There’s a lot of talk right now about AI replacing humans entirely in customer support, but I think it should actually be more of a division of labor.
AI agents perform very well with high-volume, simple, administrative requests like password resets and refund questions. The best ones are resolving most conversations on their own, which is pretty remarkable. But what that means, is the conversations that do reach humans are the ones that require more complex, empathetic responses.
That’s why it’s so important that AI agents are trained well to reliably handle those high-volume, simple requests at scale. When AI agents can get those right consistently, it frees up human agents to focus on the more complex, nuanced interactions where they add the most value. The future of CX is hybrid, and companies that intentionally design for that are the ones who are going to come out ahead.
Embedding Learning and Improvement
Q7. Continuous coaching is often talked about but rarely scaled effectively. How can organizations build a culture where learning and improvement are embedded into everyday workflows?
MH: Our State of CX survey shows 51% of agents believe 1:1 coaching is the most effective form of feedback, but it’s also the least scalable and most resource-intensive. Because teams can’t deliver this type of coaching consistently, agents are left without the guidance they need to actually improve.
However, what works is making feedback objective, specific, and continuous. When every conversation is scored against a consistent rubric, and agents get coaching tied to real examples from their own transcripts instead of just a number, it stops feeling like a performance review and instead feels actionable and easier to implement. When the feedback is consistent and constructive it makes it easy to integrate into a team’s everyday workflow.
Performance Tools Empower Employees
Q8. How do you ensure that AI-driven performance tools empower employees rather than create a sense of surveillance or pressure?
MH: The key here is for AI tools to meet employees where they are and actually help them get better at their jobs. That means when something is flagged as an error or issue, it’s immediately paired with specific guidance and coaching, so employees know exactly how to improve.
The tools also need to work alongside employees in their day-to-day workflows, rather than as a separate system. The most effective systems surface feedback in the moment, within the tools agents are already using, so improvement becomes part of the work itself rather than an extra step.
When employees see the tool as something that helps them succeed, rather than something they’re being judged by, it drives adoption instead of resistance.
Beyond Traditional Metrics
Q9. Beyond traditional metrics like CSAT and NPS, what indicators should CX leaders focus on to truly measure performance improvement?
MH: Leaders should pay close attention to things like quality of interactions, response consistency across agents, and if soft skills like empathy and active listening are improving over time based on situational feedback.
These measures matter because they get closer to what actually drives customer outcomes. Traditional metrics like customer satisfaction scores are important, but they tell you how a customer felt after the interaction, not what caused that outcome. By focusing on interaction quality and consistency, leaders can identify and fix issues earlier, before they show up in customer sentiment.
They also create a clearer path to improvement. When you can pinpoint what “good” looks like in a conversation, you can coach it, replicate it, and scale it, ensuring that both humans and AI can maintain and deliver consistent, high-quality experiences.
Closing the Insight-to-action Gap
Q10. Can you share an example of how closing the insight-to-action gap translates into tangible business outcomes?
MH: Across all of our customers, we typically see a 20x increase in quality assurance coverage and 90% reduction in manual review time.
To put this into the context of our customers, Crypto.com cut the time agents spend resolving issues by 18%, and they’re now getting full visibility into over 800,000 conversations a month. With Podium, new agents are hitting their performance benchmarks in 60 days instead of 90, and resolving customer issues 33% faster once they started handling live customer support requests in real time. Ryanair also used our simulations for candidate screening, and have since cut recruiter time by 50%, saving 38 recruiter hours per 100 candidates.
Shaping Next-generation CX Strategies
Q11. As we look ahead, how do you see the role of AI evolving in shaping next-generation customer experience strategies?
MH: As AI continues to evolve, it’s going to take on more and more conversation volume. But what’s currently underappreciated is the oversight portion. A lot of companies are moving fast to deploy AI agents without building the infrastructure necessary to know how they’re actually performing. There’s a real conflict of interest if the same system handling the conversation is also the one telling you how it went. That’s essentially AI marking its own homework.
In the next generation of customer experience, independent quality oversight is going to become standard. In the same way that security certifications became a must-have once companies moved data to the cloud, quality certification will become essential as companies move customer conversations to AI.
CX Leadership and Laggards
Q12. What will differentiate CX leaders from laggards over the next 3–5 years?
MH: I think it will come down to who builds the right oversight layer early on. The companies that are going to fall behind are the ones that deployed AI to cut costs, assumed everything was fine because the metrics looked okay, and then had a big problem on their hands before they even realized.
The leaders will be the ones who treat quality as infrastructure, not an afterthought, and who figured out the hybrid model intentionally. For most companies, this will look like AI handling the high-volume conversations, humans handling the complex ones, and an independent layer sitting across both telling them actually working and what isn’t.
Insight-to-Action Gap: Ability to Understand Customers
In today’s experience economy, the ability to understand customers is no longer a differentiator—acting on that understanding is.
In this CXQuest conversation, Mark Hughes of Solidroad explores how organizations can bridge the long-standing gap between insight and execution. Many enterprises have invested in analytics platforms, dashboards, and voice-of-customer tools. he real challenge lies in translating those insights into consistent, scalable improvements in customer interactions.
Mark highlights a fundamental shift underway: CX is moving from a reactive, measurement-driven function to a proactive, performance-driven discipline. At the heart of this transformation is AI. Not just as an analytical tool, but as an enabler of continuous coaching, automated quality assurance, and real-time skill development.
Insight-to-Action Gap: Convergence of Human and AI Performance Management
A key insight from the conversation is the convergence of human and AI performance management. AI agents are becoming more prevalent. Organizations must rethink CX as a unified system. A system where both humaln and digital agents are continuously trained, evaluated, and optimized.
Equally important is the cultural dimension. Embedding continuous learning into daily workflows requires more than technology. It demands leadership alignment, trust, and a shift in how performance is perceived and managed.
Ultimately, this discussion underscores a critical truth:
The future of CX will belong to organizations that can close the loop between insight, action, and measurable impact—at scale.
Insight-to-Action Gap: Key CX Leadership Insights
Insight alone does not drive CX transformation—execution at scale does
AI is evolving from an analytics tool to a performance improvement engine
The future of CX lies in continuous, automated coaching systems
Organizations must manage human and AI agents as a unified workforce
Culture and trust are critical in adopting AI-driven performance tools
Insight-to-Action Gap: Editorial Reflection
This conversation reflects a broader industry inflection point. As CX matures, the emphasis is shifting from listening to customers toward systematically improving every interaction. Leaders who embrace this shift—combining AI, strategy, and culture—will define the next generation of customer-centric organizations.
