Support Experience

Support Experience

By Krishna Raj RajaBusinessTechnology
Download on the App Store

All Support Experience episodes

  • Quantifying GenAI Confidence in Customer Support: Judge LLMs and Automated Scoring Loops

    In this episode, we explore how the SupportLogic Engineering Team is transforming generative AI summarization from a risky, black-box experiment into a trustworthy, enterprise-grade system. Moving GenAI into real-world production requires more than just a good underlying model—it demands measurable confidence. We break down SupportLogic's innovative evaluation framework, which relies on "Judge LLMs" to automatically assess AI-generated summaries across six critical dimensions: faithfulness, instruction adherence, hallucination risk, topic coverage, clarity, and persona usability.


    Listen in as we discuss how this continuous, automated scoring loop enables data-driven prompt tuning and dynamic model routing. We also dive into their latest benchmark data, comparing the quality and cost-efficiency of top-tier models like Claude 4 Sonnet, Gemini 1.5 Pro, and GPT-4o Mini. Whether you are balancing high-stakes accuracy with latency-sensitive workflows or simply trying to eliminate hallucinations in customer-facing summaries, this episode provides a strategic roadmap for deploying GenAI with quantifiable, reliable results.

    21 min
  • The Billion-Dollar Generative AI Illusion

    In this episode, we strip away the hype surrounding modern artificial intelligence to explore the reality of AI in Customer Experience (CX). We discuss why AI should be viewed as "intelligent automation" rather than magic, and examine the fundamental shift from deterministic to probabilistic computing. Discover why a massive, billion-dollar Large Language Model can fail at basic math while a pocket calculator from the 1970s succeeds, and what this means for enterprise technology.

    We dive deep into why a staggering number of generative AI projects fail, tracing the root causes to unrealistic expectations and a lack of proper infrastructure. Listeners will learn about the "last mile" problem in automation and how modern organizations are held back by four massive enterprise silos: data, context, signals, and AI itself.


    To overcome these hurdles, we explore the rise of highly composable "ambient AI agents" that run continuously in the background to extract valuable customer signals, resolve issues, and provide critical contextual memory. Emphasizing that AI is like fire or nuclear power, we highlight why continuous human oversight and monitoring are foundational to safely taming AI's capabilities.

    Finally, we challenge the invisible constraints holding the industry back. We urge business leaders to shift their mindset away from using AI purely for cost-cutting and back-office efficiency, and instead use it to spark a "cognitive revolution" that creates entirely new value, personalized services, and revenue opportunities for the future of CX

    24 min
  • Combating Combinatorial Complexity of Case Routing

    Are your engineering and support teams stuck building unmaintainable trees of exceptions? Traditional case routing relies on deterministic logic—static "if/then" rules that completely fail at the combinatorial complexity of enterprise scale.

    In this episode, we dive deep into the architecture of SupportLogic’s Intelligent Case Assignment (ICA) to understand how it shifts routing from a rigid classification task to a continuous, multi-dimensional optimization problem. We unpack the technical mechanics behind ICA's five-pillar ML scoring engine, which evaluates time overlap, agent bandwidth, customer history, and skills simultaneously in real-time.


    Listen in for a technical breakdown of how ICA replaces simple boolean skill flags with LLM-driven logical inference and TFIDF weighting, preventing high-volume cases from unfairly dominating an agent's skill signal. We also explore the system's strict architectural separation between "soft" ML recommendations and "hard" availability limits—such as CRM omnichannel presence, assignment hours, and active backlog caps.


    If you want to understand the mathematical models and queue architecture required to route global enterprise cases seamlessly without writing another static rule, this technical deep-dive is for you

    50 min
  • Crawl, Walk, Run: Enterprise AI Sidecar Playbook

    We are wasting 14 billion support hours annually—time that, with the right AI strategy, can be reclaimed and redirected toward value creation.

    But rushing to adopt AI without a clear plan risks chaos. This episode reveals a pragmatic, step-by-step framework that enables enterprises to harness generative AI safely and effectively, transforming support operations while avoiding costly pitfalls.

    We break down how the AI hype cycle is misguiding many, and why the real opportunity lies in incremental, phased adoption—moving from simple wrappers around public models to deploying custom, private LLMs tailored to your company's unique data.

    Discover how retrieval-augmented generation (RAG) is revolutionizing enterprise workflows, grounding AI in proprietary knowledge, and drastically reducing errors and hallucinations. Learn why a ‘sidecar’ approach—integrating AI alongside existing systems—is the smartest way to stay agile amid rapid tech evolution.

    This episode explore concrete use cases like persona-based support summaries, language translation tools that eliminate communication barriers, and intelligent escalation prediction. These innovations cut resolution times by shifting human roles from reactive firefighting to strategic oversight—managing AI systems, tuning models, and focusing on complex issues only humans can handle. Importantly, you'll understand the critical guardrails needed to prevent financial, legal, and reputational risks, like data privacy safeguards and understanding hallucination dangers.This episode provides the clarity you need as a leader or practitioner to act decisively, turning chaos into competitive advantage.

    The key message: whether you're in customer support, operations, or product development, AI is not a distant future but a current sidecar attachment—ready to accelerate your business, if implemented thoughtfully, quickly, and responsibly. Don’t wait for tech to settle—embrace it now and shape your organization into a future-ready powerhouse.

    Perfect for executives, AI strategists, and product teams aiming to turn disruption into opportunity. This is your blueprint to move fast, stay safe, and lead the AI revolution from the front.

    2 min
  • The Issue With Customer Surveys And How AI Fixes It

    In this episode, we unpack the "Big Three" customer feedback metrics used by organizations worldwide: Customer Satisfaction (CSAT), Net Promoter Score (NPS), and Customer Effort Score (CES). We'll explore the heated debates surrounding these metrics—from experts claiming NPS calculations are flawed and "fake science" to university studies showing CES might be the weakest predictor of all

    The conversation delves into the evolution of customer feedback surveys, the shift to a subscription economy, the controversy surrounding the Net Promoter Score (NPS), the flaws of traditional customer feedback metrics, survey fatigue and response rate biases, the frustration of the creator of NPS, the need for change, the solution of implicit feedback and NLP, and the philosophical implications of AI-driven feedback.

    Takeaways

    • Customer feedback surveys are evolving from traditional explicit surveys to AI-driven implicit feedback.
    • The future of customer feedback relies on natural language processing (NLP) to analyze customer sentiment and eliminate survey biases.

    Chapters

    • 00:00 The Evolution of Customer Feedback Surveys
    • 05:59 The Controversy of Net Promoter Score (NPS)
    • 13:01 Survey Fatigue and Response Rate Biases
    • 18:02 The Solution: Implicit Feedback and NLP
    23 min
  • The USB-C for AI: Shattering Support Silos Using MCP

    In today’s enterprise landscape, over 95% of organizations report near-zero measurable returns on AI investments because critical data remains trapped in fragmented "AI silos". In this episode, we deep dive into the SupportLogic MCP Server, a secure, real-time bridge designed to connect SupportLogic’s deep intelligence directly to your preferred AI assistants and agentic frameworks, including Claude Desktop, ChatGPT, and VS Code.


    We explore why industry leaders are calling the Model Context Protocol (MCP) the "USB-C for AI"—a universal integration layer that replaces brittle, bespoke code with a standardized, enterprise-grade context.


    In this episode, you’ll learn about:

    • The Three AI Primitives: How the SupportLogic MCP Server uses Tools, Resources, and Prompts to move beyond traditional REST APIs and enable AI to perform complex, autonomous actions.
    • Enterprise-Grade Security: A look at the Zero-Trust architecture and the MCP Gateway, which ensures every AI request is authenticated, authorized, and policy-checked before execution.
    • Operational Grounding: How the server ensures AI outputs are "grounded" in real-time signals like sentiment, escalation risk, and account health.
    • Real-World Agentic Workflows: We break down five transformative use cases where AI agents autonomously orchestrate workflows, including:
      • Generating Executive Escalation Briefings without manual intervention.
      • Achieving 100% QA Coverage and automated coaching notes.
      • SLA Breach Prevention through persistent, "always-on" monitoring.
      • Detecting Cross-Account Trends to catch emerging product issues before they escalate.


    Who should listen: Support leaders, AI engineers, and enterprise architects looking to transform their support data into a competitive advantage by building scalable, intelligent AI workflows

    22 min
  • Braze, Coupa, and the "Build vs. Buy" AI Dilemma

    In this episode, we cut through the hype to explore exactly companies are successfully integrating AI in their support workflows.

    We are discussing the panel featuring Erika Semtei, VP of Customer Support at Braze, and Declan Fanning from Coupa, as they share their firsthand experiences partnering with SupportLogic to drive tangible business results.

    Whether you are weighing the "build vs. buy" dilemma or trying to figure out which AI use case to tackle first, this conversation delivers actionable insights for CX leaders. Erica and Declan break down their pragmatic, step-by-step approaches to AI adoption, proving that the best AI strategies often start behind the scenes rather than directly in front of the customer.


    In this episode, we cover:

    • Prioritizing AI Use Cases: Why both Braze and Coupa chose to focus on internal tools—like escalation management, case sentiment analysis, and intelligent routing—before rolling out customer-facing virtual assistants.
    • The "Build vs. Buy" Debate: Why purchasing a specialized AI solution often provides better agility, scale, and time-to-value compared to building in-house, and how hybrid models might shape the future.
    • Data as the Backbone of AI: Why you must clean up and unify your knowledge base and existing data streams to avoid the "garbage in, garbage out" trap.
    • Change Management & Employee Buy-in: Strategies for training your team, creating internal champions, and empowering top performers to use AI as a co-pilot rather than fearing it as a replacement.
    • The Irreplaceable Human Element: Fascinating data revealing that a support agent's soft skills and communication consistency—not just their technical product knowledge—are the most critical traits that AI cannot replace.


    Tune in to learn how to strategically deploy AI to reduce resolution times, lower escalation rates, and boost both customer and employee retention.

    22 min
  • Empty Chairs and Giant Screens: The Support Experience Cultures of Amazon, Freshworks, & Sunbasket

    In this episode, we explore why having the right technology and processes isn't enough to deliver a world-class Support Experience (SX) without the right company culture to back it up. Drawing on insights from Chapter 10 of Krishna Raj Raja's book, we discuss why resilient companies like Adobe constantly refine their core values to survive and thrive through decades of rapid market changes.

    Tune in as we unpack the three foundational pillars of a true SX culture:

    • Driven by the Voice of the Customer: Discover why executives at Amazon, Apple, and Airbnb refuse to outsource customer listening. We also explore the concept of "dogfooding" and how companies like Sunbasket train their employees to become "Customer Zero" to build deep empathy for the customer journey.
    • Busting Silos Through Radical Transparency: Uncover the dangers of the "yellow state," where small problems fester between departmental silos. We discuss how Salesforce uses their public "V2MOM" strategy documents to align 50,000 employees and how FreshWorks publicly broadcasts positive customer feedback to motivate their teams.
    • Valuing 1% Improvements Over Short-Term Fixes: Why the engineering urge to "rip and replace" broken systems is a trap. Learn how adopting an "outside-in" approach and a coaching mindset can turn negative customer interactions into long-term brand loyalty.

    Whether you are a frontline support engineer or a CEO, this episode will show you how to embed genuine customer obsession into the very fabric of your organization's daily operations

    19 min
  • Firefighting to Forecasting: Deep Dive Into Escalation Prediction Engine

    In many organizations, customer escalations are treated like sudden, destructive lightning strikes. We challenge that notion, exploring the hard truth that escalations are not random, but the logical conclusion of detectable patterns.
    This podcast provides the blueprint for shifting your team from "firefighting" to forecasting through architectural intent and predictive intelligence.

    Join us as we deep dive into:

    • The Science of Prediction: Deep dives into Likely to Escalate (LTE) models and why focusing on recall is more critical for business impact than simple accuracy or precision.
    • The Dual-Signal Engine: How to leverage patented AI to analyze both structured signals (like case patterns) and unstructured "silent signals" such as sentiment trajectory and emotional intensity.
    • Proactive Architecture: Why traditional CRM tracking fails and how to build a dedicated escalation custom object to capture immutable timestamps and sentiment shifts.
    • MLaaS and Innovation: Exploring how Machine Learning as a Service (MLaaS) allows organizations to deploy models rapidly and monitor for data drift to ensure peak performance.

    Featuring insights from SupportLogic engineering experts and CX industry leaders, we tell the stories of how world-class brands are preventing escalations.

    This episode is for Engineering leaders, CTOs, ML architects and Engineers

    15 min
  • Turning Support Tickets Into Revenue - Lessons from Hubspot, And Ernst & Young

    In this episode, we discuss the panel discussion between the host Ryan Nichols (Partner at DYDX Capital and former EVP of PM at Salesforce Service Cloud), Sowmya (VP of Revenue Operations at HubSpot) and Vijay (Partner for AI and Data at EY) to tackle a critical business challenge: transforming everyday customer support data into strategic executive insight.

    While customer support is a gold mine of data—often accounting for the vast majority of a company's customer interactions—there is a fundamental disconnect in how it is viewed by leadership. Support is traditionally measured in lagging, operational metrics like SLAs and ticket deflection, which fail to capture the attention of a C-suite focused on revenue, growth, and customer lifetime value.

    Listen in as our expert panel discusses how to break down data silos and translate support "noise" into powerful business signals that executives simply cannot ignore.

    What You'll Learn in This Episode:

    • Overcoming the Leadership Language Barrier: Discover why presenting support issues as "product roadmap requests" or "drivers of growth" is far more effective than just reporting on bugs and ticket volumes.
    • The Power of Hidden Metrics: Learn why the customer "Effort Score" and HubSpot's unique "Pain Value Index" are unsung heroes that serve as critical leading indicators for churn risk.
    • Driving Expansion and Revenue: Hear how HubSpot utilizes support interactions to uncover massive expansion opportunities, including using ambient AI agents to automatically generate Sales Qualified Leads (SQLs) from unstructured support conversations.
    • Actionable Advice for Getting Started: Find out why starting small, focusing on a specific problem like upcoming renewals, and initially "doing things that don't scale" is the proven formula for building a business case and winning executive buy-in.

    Featured Guests:

    • Sowmya, VP of Revenue Operations at HubSpot.
    • Vijay, Partner for AI and Data at Ernst & Young (EY)
    20 min