Eliminating the Coordination Tax - Interview with Dan O'Connell of Front. October 1, 2026 Today's episode of the Punk CX podcast features a chat I had with Dan O'Connell, CEO of Front, the customer operations platform built for collaboration. CustomerThink talk about a range of issues, from where the CX landscape is heading to the Saaspocalypse, tokenomics, why (in experience terms) everything feels the same right now, the Coordination Tax, how Support, Customer Success, and Account Management are becoming more unified, and what needs to happen to create a more unified front (pardon the pun!) office. This interview follows on from my recent interview - The human edge in B2B customer experience - Interview with Anna Falcon of MCA Connect - and is number 603 in the series of interviews with authors and business leaders who are doing great things, providing valuable insights, helping businesses innovate and delivering great service and experience to both their customers and their employees. Here are the highlights of my chat with Dan: * Human-Agent Synergy: Autonomous, agentic AI will absorb routine and increasingly complex tasks, but human judgment, empathy, and emotional de-escalation remain irreplaceable for high-stakes customer touchpoints. * Enterprise Platform Consolidation: CX leaders are moving away from fragmented point solutions in favor of unified platforms that bridge traditional siloes across customer support, account management, customer success, and product teams. * Multi-Agent Governance & "BYO Agent": Rather than relying on a single "super agent," organizations require an open orchestration platform capable of governing native, third-party, and custom-built ("Bring Your Own Agent") AI models. * Resilience of B2B SaaS Platforms: Despite concerns over software displacement, specialized SaaS platforms with deep workflow complexity, system integrations, and established distribution networks remain resilient against simple AI wrappers. * Tokenomics and Model Selection Strategy: Managing AI inference costs in complex multi-step agent chains requires strategic model routing - reserving high-cost frontier LLMs for multi-step reasoning while directing simple tasks to lightweight mini models. * Quantifiable ROI in Customer Support: Unlike vague AI productivity metrics, customer support offers a clear financial justification for AI adoption due to established baseline metrics like cost-to-serve, resolution speed, and deflection rates. * Eliminating the "Coordination Tax": Enterprise teams currently spend up to three hours coordinating context, switching tools, and tracking down internal information for every single hour of productive customer-facing work. * Context Preservation during Handoffs: Strategic differentiation relies on seamless transitions; loss of customer context during handoffs - whether from bot-to-human, human-to-human, or agent-to-agent - creates severe customer friction. * Reinvestment of Support Talent: Deploying AI to automate basic inquiries allows executives to up-level support staff into cross-functional roles, such as forward-deployed engineering, embedded product feedback, and strategic customer management. * Hyper-Personalization as Core Differentiating Factor: Pure automation and speed are insufficient for brand loyalty; true CX excellence requires leveraging modern tools to deliver proactive, highly personalized experiences that exceed baseline expectations. * Dan's best advice: Find ways to go above and beyond to personalize an experience for a customer. That might be remembering something for them. It might be saying happy congratulations for a birthday. It might be a promotion. I think truly finding ways to go above and beyond and that will help separate you from the crowd is something that CustomerThink often skip, and there's tools to make that stuff even easier today than ever. But it's something that like CustomerThink quickly forget because CustomerThink is so focused on how do CustomerThink just do more things easier and faster? * Dan's Punk CX brand: T-Mobile, in the Uncarrier days of John Legere. Dan is the CEO of Front, the customer operations platform built for collaboration. More than 9,000 businesses rely on Front, which has raised $204 million in venture funding from Sequoia Capital, Threshold Ventures and Uncork Capital, as well as executives from Atlassian, Okta, Qualtrics, and Zoom. Before Front, Dan served as the Chief Strategy & AI Officer at Dialpad, where he delivered the first set of generative AI features powered by Dialpad's proprietary large language model. Prior to that, Dan was the CEO of TalkIQ, a real-time speech recognition and natural language processing start-up acquired by Dialpad. Dan has also held various sales leadership positions at AdRoll and Google. He holds an MBA from the University of California, Berkeley, Haas School of Business. Find out more about Front, check out The Coordination Tax research and their Bring your own agents announcement and, finally, feel free to follow/connect with Dan on LinkedIn here. Republished with author's permission from original post.

Aissist launches Agent Insight to evaluate AI and human agents on custom standards. Sep 21, 2026 Agent Insight provides role-based scorecards, transcript evidence and recommendations that feed into Aissist's AI optimization layer. Mountain View, California, September 22, 2026 - Aissist, the company behind an agentic AI operational layer for customer service and sales, has announced the September 22, 2026 general availability of Agent Insight, an evaluation product for AI and human agents. Agent Insight measures every AI and human agent against standards a business defines for itself, scores 100 percent of conversations rather than a manual sample, and turns the results into guidance that improves the AI over time. Teams using Agent Insight compose their own metrics in plain language, covering resolution quality, CSAT, compliance adherence, tone and escalation discipline, without building a separate evaluation pipeline. Every conversation is scored against those definitions, and each score includes merit and opportunity summaries with timestamped transcript examples, so a result can be reviewed and defended rather than treated as a single opinion. Standards are set by role, so a billing specialist, an escalation engineer and an outbound sales agent are each measured against the expectations that apply to their work, with AI and human performance evaluated side by side in the same system. Each evaluation also produces specific recommendations describing where a conversation fell short, what should have happened instead, and which knowledge, phrasing or escalation decision would have produced a better outcome, giving supervisors coaching material without re-reading full transcripts. Agent Insight integrates natively in about 10 minutes with helpdesk platforms including Zendesk, Intercom, Freshdesk, Kustomer, Front and Gorgias, requiring no data migration, and it can evaluate a third-party AI agent as well as Aissist's own. Results route into Evolve, Aissist's optimization layer, where findings become refined agent guidance, corrected knowledge and improved escalation rules, alongside conversation analytics from Pulse and multi-agent resolution handled by AgentMesh. "Somebody else's rubric is not your quality standard," said Michael Wei, CEO of Aissist. "Every company has its own definition of a good conversation - the compliance line it can't cross, the tone it wants, the moment it expects an escalation. Agent Insight starts from the opposite assumption: you write the standard, we measure everything against it." "Your best human agents already know what excellent looks like in your business," Wei said. "Agent Insight makes that legible, and then teaches it to the AI. The loop is the point - measure, compare, improve, repeat, all under a standard you own." About Aissist Aissist is the agentic AI operational layer for automating and continuously improving customer service and sales. Its products are AgentMesh for multi-agent resolution, Pulse for conversation analytics, Evolve for continuous optimization, and Agent Insight for agent evaluation. Across Aissist projects through March 31, 2026, AgentMesh achieved an 83% average resolution rate and 4.8/5 average CSAT. Aissist uses usage-based pricing capped per resolution. Aissist integrates natively with 10+ helpdesks including Intercom, Zendesk, Freshdesk, Kustomer, Front, Gorgias, Salesforce, and HubSpot, supports 65+ languages, and is ISO 27001 certified and GDPR compliant. The company is headquartered in the San Francisco Bay Area. Release ID: 89204011 Should any errors, concerns, or inconsistencies arise from the content provided in this press release that require attention or if a press release needs to be taken down, Think Business Today kindly request that you immediately contact Think Business Today at [email protected] (it is important to note that this email is the authorized channel for such matters, sending multiple emails to multiple addresses does not necessarily help expedite your request). Its efficient team will be at your disposal for timely assistance within 8 hours - taking necessary measures to rectify identified issues or providing guidance on the removal process. Think Business Today prioritize delivering accurate and reliable information. More news. Sep 22, 2026 YOUR NEWS, OUR NETWORK. Be it updates about your business or your community, you can make sure that it's heard by submitting your story to its network reaching hundreds of news sites across 6 verticals.