ClearChoice Cuts Dental Call Times to Eight Minutes With Cresta AI Agents
Ai4
Automating after-hours intake and routine confirmations with custom AI models maintains high patient engagement while improving call center operational efficiency.
## Background
This YouTube video was published by Ai4 on August 20, 2026. It features a fireside chat between Craig Kierst, SVP of Sales Strategy and Growth at ClearChoice (The Aspen Group), and Parthh Pathak, Field CTO for Regulated Industries at Cresta. The session details how ClearChoice deploys conversational AI agents across voice, SMS, and chat to engage patients and transform the dental implant customer journey.
## TL;DR
ClearChoice partnered with Cresta to deploy 'Hope,' an AI agent designed to address patient engagement barriers and handle after-hours inbound calls for dental implant consultations. Facing lengthy onboarding times and high turnover in traditional call centers, ClearChoice used AI to provide consistent, patient, and empathetic interactions without rushing prospective patients. By adopting a 'Human + AI' model, repetitive confirmation calls and routine intake tasks are automated, allowing human staff to dedicate more time to anxious patients who require direct care. Through continuous hands-on testing and iterative fine-tuning, ClearChoice optimized call durations to 8–9 minutes while successfully expanding AI deployment into daytime and outbound channels.
## Atmosphere & style
The presentation carries an insightful, collaborative, and pragmatic tone as two industry experts reflect on real-world AI implementation. The dynamic between Parthh Pathak and Craig Kierst feels candid and conversational, blending technical strategy with practical business advice. Watching the session conveys a grounded sense of innovation, highlighting the delicate balance between technical optimization and genuine human empathy when serving patients facing sensitive healthcare decisions.
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=4s) ClearChoice Dental Implant Centers operates over 100 specialized clinics across the US, focusing on full-mouth dental rehabilitation and prosthodontic procedures.
- [00:04](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=4s) Because dental care is paid out of pocket, patients frequently delay treatment, making consultative education and engagement essential prior to booking.
- [00:42](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=42s) ClearChoice launched its initial AI use case with an after-hours inbound agent named "Hope" to overcome regulatory time limits on outbound calls after 8:00 p.m.
- [00:42](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=42s) Only about 20% of after-hours callers request human escalation, with most becoming comfortable with the AI agent after a minute or two.
- [00:06](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=6s) Traditional call center staff face high turnover and require 2 to 3 months of onboarding, whereas AI agents provide consistent, patient, and non-judgmental interactions.
- [00:09](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=9s) Craig Kierst evaluated over 20 vendor platforms before selecting Cresta for its LLM flexibility and native support across voice, SMS, and chat.
- [00:11](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=11s) While inbound scheduling automated successfully, outbound AI calls required fine-tuning to handle smartphone call-screening and mid-voicemail pickups.
- [00:13](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=13s) ClearChoice follows a "Human + AI" model, transferring transactional tasks like appointment confirmations to AI so human staff can focus on anxious, high-need patients.
- [00:13](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=13s) To prevent implementation delays, ClearChoice started its rollout in September 2023 focused strictly on 3 initial target use cases.
- [17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) After finding that initial voice interactions ran too long, ClearChoice optimized the AI agent to balance patient education with an efficient 8 to 9 minute call time.
- [17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) Leadership advises organizations to dive into AI adoption immediately, iteratively testing and calling their own system to refine the experience.
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=4s) Introduction and the ClearChoice Patient Journey
Parthh Pathak, Field CTO for Regulated Industries at Cresta, opens the fireside chat at Ai4 by polling the audience on how many organizations run AI agents in production and are familiar with Cresta's conversational AI platform. Joining him is Craig Kierst, SVP of Sales Strategy and Growth for ClearChoice Dental Implant Centers (part of The Aspen Group). ClearChoice operates over 100 specialized clinics across the United States focused on full-mouth dental rehabilitation and restoring smiles.
In dentistry, patients receive baby teeth and adult teeth for free, but must pay out of pocket for any subsequent dental work. Consequently, oral care is often the first health priority people abandon when facing financial or personal hurdles. Delaying necessary treatment leads to severe dental issues where individuals lose all their teeth and require full-mouth restoration. ClearChoice bridges this gap by reaching out to patients through targeted advertising and meeting them at their point of need, connecting them with prosthodontists and oral surgeons to rebuild their smiles.
- ClearChoice operates over 100 specialized dental implant clinics across the United States.
- The average patient conversation with the "Connected Care Team" lasts between 8 to 10 minutes rather than operating like a standard high-volume scheduling call center.
- Initial patient touchpoints focus heavily on education, guiding prospects from initial awareness and TV ad exposure into full consultations.
[00:42](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=42s) Launching the Initial AI Agent Use Case
To address engagement barriers without disrupting core daytime operations, ClearChoice identified after-hours phone calls as the optimal starting point for its AI deployment. Regulatory constraints restrict outbound marketing calls past 8:00 p.m., leaving late-night inbound calls from prospective patients underserved. Previously, ClearChoice relied on a third-party call center after hours, but this setup failed to capture detailed patient nuances or support effective, timely follow-up.
Partnering with Cresta, ClearChoice created "Hope," a 24/7 female AI agent trained on the company's language models and patient interaction frameworks. Operating after hours offered a low-risk environment to prove the technology. Contrary to expectations that patients would demand a human, ClearChoice discovered that only about 20% requested human escalation. Most patients became comfortable within a minute or two, engaging in deep dialogues, sharing personal concerns, and addressing the AI agent respectfully as "ma'am."
- Outbound calls face strict regulatory time limits at 8:00 p.m., making inbound automation essential for night-time leads.
- Only approximately 20% of after-hours callers request human assistance once they experience a minute or two of natural dialogue with the AI agent.
- Hope provides continuous 24/7/365 coverage for inbound after-hours patient inquiry and education.
[00:06](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=6s) Overcoming Call Center Friction and Scaling Agent Consistency
Traditional call centers face severe turnover and long onboarding cycles. In specialized dental care, newly hired human agents require two to three months of training to reach full productivity and deliver empathetic interactions to anxious or self-conscious patients. Patients facing severe dental issues often feel embarrassed about their smiles and frustrated by their situation, which can make calls challenging for human representatives to navigate.
The AI agent addresses this consistency challenge by remaining patient, listening attentively, and taking time to clarify information without showing frustration or hanging up. Unlike human staff who require months of retraining or ramp-up when turnover occurs, the AI model maintains performance standards continuously and refines its performance over time. Following success with after-hours calls, ClearChoice began expanding AI usage into daytime operations and outbound workflows.
- Human call center onboarding historically takes 2 to 3 months to reach baseline performance and consistency.
- AI agents handle patient hesitation and speech difficulties patiently, ensuring patients never feel rushed during complex calls.
- ClearChoice expanded its AI agent implementation from initial after-hours inbound reception into daytime interactions and outbound campaigns.
[00:09](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=9s) Selecting Cresta and Evaluating Multimodal Capabilities
When evaluating prospective AI partners, Craig Kierst personally reviewed over 20 vendor demonstrations across the conversational AI landscape. ClearChoice chose Cresta due to its native large language model (LLM) flexibility and customizability. Because ClearChoice engages in 10- to 15-minute qualification calls to ensure both patient suitability and treatment fit, rigid "next-best-action" scripts proved insufficient.
ClearChoice required an omnichannel solution capable of expanding beyond voice into SMS text messaging and chat workflows. Meeting patients on their preferred channel required a unified platform rather than disparate point solutions.
- Craig Kierst evaluated over 20 different AI platforms during a comprehensive RFP process before selecting Cresta.
- Consultative qualification interactions with patients range between 10 to 15 minutes per session.
- Cresta's platform supports voice, SMS, and web chat to provide a flexible, multi-channel strategy.
[00:11](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=11s) Operational Insights and Initial Technical Fine-Tuning
Deploying AI agents exposed distinct performance characteristics between call workflows. Inbound scheduling and basic inquiry containment performed reliably out of the box, matching human agent conversion benchmarks. However, outbound automated calls presented technical challenges due to variable answer rates and voicemail interactions.
Modern smartphones and operating systems introduce features such as real-time voicemail transcription where users intercept calls mid-message, requiring precise timing for voicemail detection and tone triggers. Identifying these edge cases allowed ClearChoice and Cresta to refine tone-detection protocols and technical parameters while keeping core inbound operations moving smoothly.
- Inbound scheduling and intake automation matched human agent benchmarks early in deployment.
- Outbound automated calls required technical calibration to handle smartphone call-screening and mid-voicemail pickups effectively.
- Integrated analytics allow real-time monitoring of agent performance metrics rolled up across business functions.
[00:13](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=13s) The "Human Plus AI" Vision and Internal Buy-In
ClearChoice maintains a guiding "North Star" where AI and human agents work in tandem rather than replacing one another. Highly repetitive, transactional tasks—such as appointment confirmations, routine reminders, and quick inquiries—are transitioned to AI agents. Transitioning these tasks reduces burnout among human staff, who previously spent hours executing repetitive confirmation calls.
By offloading transactional tasks to AI agents, ClearChoice frees human staff to spend meaningful time with uncertain or anxious patients who need dedicated support. When pitching this strategy internally to corporate leadership and IT partners, Kierst emphasized simplicity to avoid momentum loss. Rather than rolling out dozens of complex features at once, ClearChoice launched with just three target use cases to demonstrate tangible ROI and establish a solid foundation for growth.
- ClearChoice's core strategy centers on a "Human + AI" model that automates repetitive tasks to expand human capacity.
- Prior to AI implementation, human teams spent significant hours manually placing repetitive confirmation calls, contributing to high staff turnover.
- To avoid implementation delays ("complexity kills momentum"), ClearChoice launched its rollout focused strictly on 3 initial use cases.
- The rollout initiative began in September 2023, building a year of operational experience prior to scaling.
[17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) Iterative Learning, Optimizing Voice Agent Duration, and Practical Advice
The journey of implementing AI agents is an evolving, iterative process that requires continuous learning and willingness to pivot. Reflecting on the past 12 months, Craig Kierst explains that "you really don't know what you don't know" until you actively go through the process. Throughout the deployment, ClearChoice evaluated what was working as anticipated, what required doubling down, and where things were moving slower than expected. Patient interactions provided a wealth of insights, forcing the team to constantly adjust their strategy based on real-time back-and-forth engagement.
A prime example of this iteration involved their original AI voice agent. While the agent delivered lengthy, highly educational interactions that drove strong downfunnel engagement, the long call durations began hurting overall operational efficiency. ClearChoice realized that in their quest for patient education, they were actually slowing themselves down. Through fine-tuning and adjustments made in partnership with Cresta, they found a sweet spot that preserved high-quality patient engagement and education while bringing average call times back down to an efficient 8 to 9 minutes.
Ultimately, organizations cannot afford to wait for perfection before diving into AI deployment. Craig's advice to leaders is to jump in with conviction and actively immerse themselves in the experience, even if it feels daunting initially. He admits to testing the system constantly—joking that his team probably hates that he acts like "the worst patient" by calling in every single day to test scenarios. By directly living the patient's journey, leaders can understand the exact questions being asked, identify friction points, and work with tech partners to make swift edits without requiring drastic retraining. Parthh Pathak echoes this sentiment, concluding that businesses must "start now and stay at it" as part of an ongoing journey.
- Successful AI deployment requires a continuous cycle of testing, learning, and pivoting over time.
- ClearChoice fine-tuned their AI voice agent to balance thorough patient education with operational efficiency, bringing average call durations down to 8 to 9 minutes.
- Hands-on leadership and testing—such as regularly calling into the AI system as a secret shopper—helps teams understand real patient experiences and make rapid prompt adjustments.
## Video description
Craig Kierst - SVP Sales Strategy and Growth, ClearChoice at The Aspen Group
Parthh Pathak - Field CTO, Regulated Industries, Cresta
Getting a patient to show up for a dental consultation isn't just a scheduling problem, it's an engagement problem. In this fireside chat, Craig Kierst, SVP of Sales Strategy & Growth at The Aspen Group, joins Cresta Field CTO Parthh Pathak to walk through how ClearChoice is deploying AI Agents across voice, chat, and SMS to guide patients during the Dental implant customer journey. Craig will share how AI-powered two-way texting took over post-confirmation communication, what it actually looks like to have an AI Agent handle the back-and-forth with a patient at scale, and what results ClearChoice has seen from embedding agentic AI into the full patient journey. A practical, real-world look at what it takes to make AI Agents work in a high-stakes, consumer-facing environment.
ClearChoice Dental Implant Centers integrated Cresta's conversational AI agents to automate after-hours intake, routine scheduling, and preliminary patient qualification, addressing high call center turnover while maintaining empathetic support for complex dental cases. According to ClearChoice executive Craig Kierst and Cresta's Parthh Pathak, adopting a focused 'Human plus AI' framework allowed the healthcare provider to improve operational capacity without replacing specialized human care.
Targeting After-Hours Inbound Calls as a Low-Risk Entry (0:42)
Because regulatory restrictions prohibit outbound telemarketing calls after 8:00 p.m., ClearChoice faced a persistent coverage gap for prospective patients inquiring late at night. The organization previously relied on a third-party call center that struggled to capture patient nuances or support effective follow-up. To resolve this without disrupting daytime workflows, ClearChoice deployed an AI agent named 'Hope' to provide continuous, 24/7 coverage for inbound inquiries.
While leadership initially anticipated consumer resistance, Kierst reports that only approximately 20% of after-hours callers requested transfer to a human representative. Most prospective patients adapted to the AI within one to two minutes, engaging in deep dialogues regarding full-mouth restorations, sharing personal vulnerabilities, and speaking respectfully to the automated system.
Standardizing Patient Care and Mitigating Contact Center Churn (0:06)
Consultations for extensive dental rehabilitation require high emotional sensitivity, as patients often feel self-conscious, embarrassed, or frustrated regarding severe tooth loss. In traditional call centers, onboarding new representatives requires two to three months of training to reach consistent productivity and empathy, a hurdle exacerbated by severe industry turnover. Repetitive manual outbound tasks, such as reminder and confirmation calls, further contributed to staff burnout.
The automated agent counteracts these operational frictions by listening patiently, accommodating speech hesitations, and avoiding frustration during lengthy interactions. By offloading routine, transactional touchpoints to conversational AI, ClearChoice freed human staff to concentrate on uncertain or emotionally distressed patients who require dedicated human guidance.
Omnichannel Selection and Overcoming Outbound Technical Friction (0:09)
Evaluating over 20 conversational AI vendors, ClearChoice selected Cresta to accommodate 10- to 15-minute qualification discussions that rigid, pre-scripted decision trees could not navigate. The deployment required native large language model flexibility alongside omnichannel capabilities spanning voice, SMS, and web chat to follow patient preferences across touchpoints.
While inbound scheduling workflows met human conversion benchmarks early in the rollout, outbound automated calling introduced unexpected technical friction. Mobile operating system features, such as automated call-screening and real-time voicemail transcription where recipients intercept calls mid-message, required ClearChoice and Cresta to recalibrate tone-detection timing and voicemail triggers to avoid dropped connections.
Calibrating Educational Depth Against Operational Call Efficiency (17:34)
Deploying conversational AI required continuous operational trade-offs between depth of care and contact center throughput. Early iterations of the voice agent provided extensive patient education that drove strong downfunnel conversion but significantly increased average handle times, creating operational bottlenecks. Through iterative prompt adjustments, ClearChoice compressed average call durations to an optimal range of 8 to 9 minutes while preserving consultative quality.
Kierst advises business leaders against postponing AI initiatives in search of operational perfection, recommending instead a narrow launch focused on a few high-impact use cases. To accelerate learning, Kierst conducted daily hands-on testing as a 'secret shopper,' allowing the team to uncover conversational edge cases and execute prompt refinements rapidly without extensive engineering overhauls.
ClearChoice's experience demonstrates that conversational AI can manage nuanced, emotionally sensitive healthcare inquiries when deployed as an operational support rather than a total human replacement. However, long-term efficiency requires ongoing technical calibration to manage mobile call-screening barriers and balance educational call length against volume capacity.
## Background
该视频由 Ai4 于 2026 年 8 月 20 日发布在 YouTube 平台。内容为 ClearChoice（隶属于 The Aspen Group）销售策略与增长高级副总裁 Craig Kierst 与 Cresta 监管行业现场 CTO Parthh Pathak 的炉边对话，主要介绍了 ClearChoice 如何利用 Cresta 的对话式 AI 智能体覆盖语音、短信和在线聊天等多渠道，重塑牙科种植患者的全旅程沟通与预约转化。
## TL;DR
ClearChoice 分享了在决策链长、情感复杂的牙齿种植医疗咨询场景中落地 AI 智能体的实战经验。团队以非工作时间为低风险切入点推出名为“Hope”的 AI 语音智能体，有效承接夜间长达 8 至 10 分钟的深度科普与资格确认，将转人工率降至约 20%，同时彻底解决了呼叫中心人员流失与服务一致性难题。随后，ClearChoice 将 AI 拓展至全渠道与外呼场景，确立了“AI 处理标准化事务、人工负责深度共情关怀”的人机协同模式，并通过持续迭代在患者科普深度与通话效率之间找到最佳平衡。
## Atmosphere & style
整场对话务实、严谨且富有感染力，两位嘉宾配合默契，从齿科医疗的真实痛点与技术落地细节层层拆解。Craig 的分享生动风趣又充满同理心，他以“第三副牙齿”的幽默比喻和每日亲自拨打测试电话的实操经历，展现了对患者体验的极致追求。现场探讨不避讳外呼打磨与通话时长控制等真实挑战，兼具战略远见与可落地的实操经验，整体氛围真诚而鼓舞人心。
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=4s) 全口牙齿重建等专业齿科医疗决策极具情感与医疗复杂性，患者沟通并非简单挂号，而是平均耗时8到10分钟的深度资质评估与信任建立。
- [02:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=134s) 受限于晚间外呼法规限制及患者夜间咨询习惯，ClearChoice选择非工作时间（After-hours）作为AI智能体切入点，以低风险填补全天候服务空白。
- [02:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=134s) 非工作时间AI智能体“Hope”试点中仅约20%患者转人工，展现出自然的情感安抚与专业答疑能力，有效瓦解了患者的防备心理。
- [07:06](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=426s) 传统呼叫中心受制于2至3个月的培训爬坡期和高流失率，而AI智能体具备永不疲倦、情绪稳定及免培训即时上线的服务一致性优势。
- [09:18](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=558s) 在评估超过20家AI厂商后，ClearChoice选中Cresta，核心看重其原生LLM底座对长达10至15分钟深度双向医疗咨询的支持与全渠道多模态扩展能力。
- [11:01](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=661s) AI在常规排期和进线拦截承接上表现出色，但外呼场景中仍需克服语音信箱精准判定与手机实时留言（Live Voicemail）中断切换等技术挑战。
- [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) ClearChoice的AI北极星愿景是“人机协同”（Human + AI），由AI全权接管预约确认等重复性事务，释放人工坐席专注于高价值情感共情与复杂病情沟通。
- [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) 内部推进AI项目需警惕“复杂度扼杀势头”，应聚焦初期少数核心场景跑通明确ROI，并强调尽早启动以建立先发优势。
- [17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) 团队在迭代中通过精细化测试与微调，成功平衡了AI语音代理深度科普与通话效率之间的冲突，将平均通话时长控制在理想的8到9分钟区间。
- [18:39](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1119s) 业务负责人亲自每日拨打测试电话以患者视角实测流程，依托敏捷配置架构实现快速迭代，强调企业应消除恐惧、立即行动并持之以恒深耕AI应用。
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=4s) 开场介绍与嘉宾背景
Cresta 负责监管行业（Regulated Industries）的现场首席技术官（Field CTO）Parthh Pathak 主持了本场对谈。在正式开始前，Parthh 向现场观众抛出了两个互动问题：首先是有多少人在生产环境中运行 AI 智能体（AI Agent），现场举起了不少手；接着他又询问有多少人了解 Cresta，台下也有数位观众示意。随后，Parthh 介绍了 Cresta 的定位——这是一家提供对话式 AI（Conversational AI）的企业级 SaaS 平台，专注于在高度监管的行业中跨语种、跨区域、跨业务场景地大规模部署 AI 解决方案。
ClearChoice（隶属于 The Aspen Group）的销售策略与增长高级副总裁（SVP of Sales Strategy and Growth）Craig Kierst 随后介绍了 ClearChoice 的业务背景。ClearChoice 专注于全口牙齿重建与种植等专业齿科医疗服务，在全美拥有超过 100 家医疗中心与诊所。Craig 强调，他们的业务是高度以患者（Patient）为中心的，整个患者旅程充满了情感与医疗决策的复杂性，这也是他们探索 AI 智能体应用的核心出发点。
---
### [02:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=134s) 齿科患者旅程的痛点与非工作时间 AI 智能体破局
Craig 详细拆解了 ClearChoice 面对的独特患者旅程。他幽默而写实地打了个比方：“每个人一生中都能免费得到两副牙齿——乳牙和恒牙；但第三副牙齿你就必须自己掏钱包了。”对许多患者而言，口腔健康往往是最先被忽视的环节。随着口腔状况逐渐恶化、未得到及时治疗，最终可能演变为严重的创伤性后果，导致全口牙齿脱落，必须接受全口重建（Full-mouth Rehabilitation）。当患者看到电视广告或搜索信息产生意识并联系 ClearChoice 时，面对的是极度复杂且不透明的信息。
ClearChoice 的客服团队并非普通的呼叫中心，而是被称为“互联护理团队”（Connected Care Team）。每一次患者沟通的平均时长高达 8 到 10 分钟，甚至长达 15 分钟，需要深入教育、建立信任并确认患者的治疗资格，绝非简单的挂号排期。然而，传统运营面临两大瓶颈：一方面是严格的法规限制（例如晚上 8 点后不得进行外呼呼叫）；另一方面是患者常常在晚上 9 点、10 点结束一天忙碌后才有时间打进电话咨询。过去使用第三方呼叫中心接听夜间电话时，不仅无法深入理解患者需求，更无法在患者最渴望交流的黄金时刻提供深度解答。
```
[ 患者意识阶段 ] --> 看到电视广告/主动线上搜索
|
v
[ 深度沟通阶段 ] --> ClearChoice 互联护理团队 (平均单通 8-10 分钟)
|
+-------------------+-------------------+
| (工作时间内) | (非工作时间 20:00+)
v v
[ 人工坐席深度沟通 ] [ Cresta AI 智能体 "Hope" ]
- 7x24x365 无间断守护
- 深度解答与情感安抚
- 仅约 20% 患者转人工
```
为此，ClearChoice 选择将非工作时间（After-hours）作为引入 Cresta AI 智能体的首个切入点。这个场景风险极低，既不干扰核心日间业务，又填补了 7x24x365 全天候服务的空白。AI 智能体被赋予了女性身份与名字“Hope”。Craig 分享了一个生动的案例：许多用户初进电话时通常会习惯性地说“我想找人工客服”，但在最初的试点中，实际要求转人工的比例仅有约 20%。因为仅仅交流一两分钟后，患者便发现 Hope 能够自然对答并解答疑惑，心中的防备彻底瓦解；他们开始礼貌地称呼 Hope 为“女士”（Ma'am），询问她的名字，并像与真人交谈一样敞开心扉。
---
### [07:06](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=426s) 呼叫中心人员流失痛点与 AI 智能体的高一致性
呼叫中心行业长期受到高流失率（High Turnover）的困扰。Craig 指出，培养一名新的人工坐席需要漫长的爬坡期（Ramp-up Time），而一旦有坐席离职，重新招募并培训到具备同等业务水平通常需要 2 到 3 个月的时间。加之齿科医疗涉及患者对自身笑容的羞耻感与自卑心理，患者在沟通中可能情绪暴躁、口齿不清或充满焦虑，这对坐席的情感耐受力和沟通一致性提出了极高要求。
AI 智能体彻底解决了这种服务一致性的难题：
- **永不疲态与情绪稳定**：AI 智能体不会沮丧，绝不主动挂断电话，面对情绪激动或发音含混的患者能够耐心倾听、反复澄清与确认，绝不让患者感到被催促。
- **即时上线与免培训**：AI 借助底层大模型具备自我进化与稳定输出能力，无需经历数月的培训周期，上线即具备最高水平的业务知识。
- **全渠道扩展性**：在非工作时间取得良好成效后，ClearChoice 开始逐步将 AI 智能体推广至日间服务及外呼（Outbound）场景，并从纯语音向短信（SMS）、在线聊天（Chat）等多模态全渠道（Omnichannel）延伸。
---
### [09:18](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=558s) 供应商选型历程：为何在 20 多个 Demo 中选择 Cresta
在确定引入 AI 方案时，ClearChoice 进行了严格的 RFP（方案征集）选型流程。Craig 透露，他个人亲自看过了超过 20 家不同 AI 厂商的产品演示（Demos）。市场上虽然有很多开箱即用的工具，但大多数只能处理呆板的“下一步最佳动作”（Next Best Action）流程，根本无法胜任长达 10 到 15 分钟的深度医疗咨询与患者双向资质审核（Qualification）。
ClearChoice 最终选择与 Cresta 合作，核心原因在于 Cresta 具备灵活的原生大语言模型（Native LLM）底座与深度的定制能力。齿科种植并非单纯促成预约，而是要双向评估：确认患者是否真正需要且适合这项治疗，以及 ClearChoice 是否是其最合适的医疗机构。此外，Cresta 在语音、短信、聊天等多模态（Multimodality）全渠道上的扩展弹性，也完全契合了 ClearChoice “在患者所在的渠道与其相遇”的长期战略。
---
### [11:01](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=661s) 落地实操中的亮点与外呼技术挑战
在实际落地过程中，ClearChoice 快速摸索出 AI 与人工在能力分工上的差异。过去企业往往要求人工坐席“全能化”，导致培训极为繁琐；而在实际运行中，AI 在常规排期（Scheduling）、基础咨询以及进线（Inbound）承接与拦截（Containment）方面的表现极其惊人，完全达到了优秀人工坐席的水准。
然而在外呼（Outbound）场景中，团队也遇到了一些前期的技术磨合挑战：
- **接通率与语音信箱判定**：外呼面临接通率参差不齐的问题，AI 需要精准识别语音信箱的提示音（Beep）并在恰当时机留下一段完美的留言。
- **手机系统新特性的干扰**：例如最新款 iPhone 允许用户在 AI 播报语音信箱中途随时接起电话（Live Voicemail），这要求底层技术必须具备极高的实时中断与状态切换能力。
尽管外呼存在一定技术复杂性，但进线场景的巨大成功为 ClearChoice 注入了坚定的信心，推动其继续解锁更多业务用例。同时，Cresta 提供的实时数据挖掘与分析看板，使管理层能够随时查看每位智能体的即时表现，并向上汇总以评估对整体业务指标的支撑效果。
---
### [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) “人机协同”愿景与内部立项推动策略
谈及 ClearChoice 的未来愿景，Craig 强调他们的“北极星”并非用 AI 取代人类，而是追求“AI 赋能人类”（AI and Human / Human + AI）的深度协同模式。一年前，ClearChoice 的人工团队需要日复一日地拨打枯燥的预约确认电话（Confirmation Calls），这种高度重复的事务性劳动极大地消耗了员工精力，也是导致人员流失的重要诱因。
```
[ 事务性/标准化交互 ] ------------> [ AI 智能体全权接管 ]
- 预约确认电话 (Confirmation) - 降低坐席流失率
- 快速跟进 (Follow-up) - 释放企业运营产能
- 常见简短问答 (FAQ)
[ 复杂/情感依赖交互 ] ------------> [ 人工坐席深度投入 ]
- 迷茫/不知所措的患者 - 建立深度信任与共情
- 复杂病情方案教育 - 提供高价值医疗关怀
```
在向内部领导层和董事会推销这一愿景时，Craig 总结了务实的推进经验：
1. **与 IT 及管理层紧密配合**：确保 API 接口与现有业务系统打通时的简洁性，Craig 直言“复杂度会扼杀一切势头”（Complexity kills all momentum）。
2. **拒绝盲目铺大，坚持小步快跑**：很多企业一开始就铺开数十个场景导致项目停滞，而 ClearChoice 严格将初期试点聚焦在 3 个具体场景中，通过清晰、可衡量的投资回报率（ROI）建立信任。
3. **立刻行动，抢占先机**：对于仍在观望是否引入 AI 智能体的企业，Craig 给出的唯一建议就是“立即启动”（Start now）。ClearChoice 正是在一年前的 9 月果断迈出了第一步，才得以在行业中建立起显著的先行优势。
### [17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) 过去一年的实战迭代：平衡患者教育深度与通话效率
在回顾过去一年的落地过程时，Craig Kierst 感慨道，团队在这个过程中学到了极其庞大的知识量，因为在一开始你真的“不知道自己不知道什么”（you really don't know what you don't know）。在过去的 12 个月里，ClearChoice 不断调整方向：某些环节表现完全符合预期就加倍投入，某些环节进展偏慢就主动收缩并深入排查原因。通过观察患者与 AI 交互的海量洞察，团队进行了无数次反复沟通与微调迭代。
其中一个非常具有代表性的挑战出现在最初上线的 AI 语音代理（AI voice agent）上。该语音代理在与患者沟通时聊得非常深入详尽，患者因此获得了极为充分的科普与教育，这也确实有效推动了更深层转化漏斗（downfunnel）的参与度。然而，团队很快发现通话时间被拉得太长了，反而牺牲了最初追求的运营效率，导致整体流程变慢。为此，团队展开了密集的测试与微调，最终成功找到了平衡两者的“最佳契合点”（sweet spot）——既保留了高质量的患者教育与强互动，又成功将平均通话时长控制在理想的 8 到 9 分钟区间。
- **动态调整策略**：过去 12 个月中根据实际数据持续 pivot，对符合预期的链路加码，对低效链路收缩排查。
- **效率与深度的平衡**：早期 AI 语音代理科普详尽但耗时过长，经持续调优后通话时长稳定在 8 至 9 分钟，兼顾了转化率与处理效率。
---
### [18:39](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1119s) 亲测患者旅程与行动号召：立即起步，持续深耕
谈及给同行的建议时，Craig Kierst 强调，想要达到理想的业务状态，唯一的途径就是“今天立刻开始”。尽管面对未知难免会感到害怕，但大家起初都是两眼一抹黑，唯有抱定信念大胆尝试并全力推进，AI 才能真正产生价值，因为“成功绝不会凭空降临或随机发生”。Craig 分享了自己的亲身体验，他半开玩笑地自嘲是团队眼里“最讨厌的患者”——因为他几乎每天都亲自拨打测试电话，从头到尾体验患者所经历的每一个环节、每一次互动以及系统提出的具体问题。只有真正代入患者的视角生活在他们的场景中，才能精准捕捉到系统需要做哪些微调。
Craig 进一步指出，在 Cresta 这样合适的技术伙伴支持下，AI 系统的优化变得非常敏捷，通常只需要进行快速的配置与编辑，而不需要经历大动干戈的繁重重新训练，这能为企业迅速积累起强大的变革势头。Parthh Pathak 对此深表赞同，并呼应了前一日交流中的核心观点：拥抱 AI 是一场持续学习的旅程，大多数企业的业务模式都会随着 AI 的深入应用而不断演变升级。他最后总结道，关键就在于“立即行动，并且持之以恒”（start now and stay at it），随后两人在掌声中结束了整场炉边谈话。
- **沉浸式亲测**：高管亲自每天拨打电话实测患者旅程，从真实交互中发现痛点并指导精准调优。
- **敏捷迭代优势**：依托成熟的 AI 架构支持轻量级快速编辑与优化，免去繁重的全量重训成本。
- **核心倡议**：AI 转型是一场动态旅程，企业必须消除恐惧、尽早入局并保持持续迭代的定力。
## Notable quotes
> [02:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=134s) 每个人一生中都能免费得到两副牙齿——乳牙和恒牙；但第三副牙齿你就必须自己掏钱包了。 — Craig Kierst
> [02:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=134s) 我想找人工客服
> [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) 复杂度会扼杀一切势头 — Craig Kierst
> [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) Complexity kills all momentum — Craig Kierst
> [13:14](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=794s) Start now — Craig Kierst
> [17:34](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1054s) you really don't know what you don't know — Craig Kierst
> [18:39](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1119s) 成功绝不会凭空降临或随机发生 — Craig Kierst
> [18:39](https://www.youtube.com/watch?v=QVCdWNWNeaY&t=1119s) start now and stay at it — Parthh Pathak
## Mentioned
Craig Kierst, Parthh Pathak, ClearChoice, The Aspen Group, Cresta, Hope, iPhone
## Video description
Craig Kierst - SVP Sales Strategy and Growth, ClearChoice at The Aspen Group
Parthh Pathak - Field CTO, Regulated Industries, Cresta
Getting a patient to show up for a dental consultation isn't just a scheduling problem, it's an engagement problem. In this fireside chat, Craig Kierst, SVP of Sales Strategy & Growth at The Aspen Group, joins Cresta Field CTO Parthh Pathak to walk through how ClearChoice is deploying AI Agents across voice, chat, and SMS to guide patients during the Dental implant customer journey. Craig will share how AI-powered two-way texting took over post-confirmation communication, what it actually looks like to have an AI Agent handle the back-and-forth with a patient at scale, and what results ClearChoice has seen from embedding agentic AI into the full patient journey. A practical, real-world look at what it takes to make AI Agents work in a high-stakes, consumer-facing environment.
[Source](https://www.youtube.com/watch?v=QVCdWNWNeaY)
Source quotes
[2:14] Craig Kierst: 每个人一生中都能免费得到两副牙齿——乳牙和恒牙；但第三副牙齿你就必须自己掏钱包了。
[2:14] 我想找人工客服
[13:14] Craig Kierst: 复杂度会扼杀一切势头
[13:14] Craig Kierst: Complexity kills all momentum
[13:14] Craig Kierst: Start now
[17:34] Craig Kierst: you really don't know what you don't know
[18:39] Craig Kierst: 成功绝不会凭空降临或随机发生
[18:39] Parthh Pathak: start now and stay at it
Source quotes
[13:00] Craig Kierst: complexity kills momentum
[17:34] Craig Kierst: you really don't know what you don't know
[19:34] Parthh Pathak: start now and stay at it
Original transcript
[0:04] All right, thank you. So I'll kick it off.
Uh before I start, I have two questions.
[0:09] The first question is that how many of you
are running AI agents in production? Okay,
[0:16] quite a few hands. Of course you. And the second
question is how many of you know what Cresta does?
[0:26] A few hands. Yep. So with that I'll just kick it
off. I'm part of APEC. I'm the field CTO. I cover
[0:32] regulated industries and in that role I work with
Craig in many different companies to implement the
[0:40] Cresta platform for our customers. Uh with that
over to Craig. Sure. Uh hello everyone. Welcome.
[0:47] Uh Craig here. So I'm the SVP for sales strategy
and growth for Clear Choice. Um Clear Choice is a
[0:53] specialized healthcare focused on dental. Um
we have over a 100 centers or clinics around
[0:59] the US. Um and we focus primarily on patients
helping them get back to their smile fullmouth
[1:05] rehabilitation. Um and so we're very customer
in our world patient focused and that's what I
[1:11] think the conversation today is going to be about.
Excellent. Excellent. So let me kick this thing
[1:16] off. Uh first of all I think Craig when they were
starting their journey you know in the world of
[1:21] AI we have larger and larger LLM models trained on
billions and billions of parameter when Craig and
[1:28] his team were looking for like hey which AI agent
we can use that's what that's when ka came in and
[1:36] I think the problem they're trying to solve for
is that well how can I improve my patient journey
[1:41] keeping consistent keeping humanly and you know
deploy at scale So and that's what Cresta does.
[1:48] We are a conversational AI company, a SAS platform
and we work with many regulated industry to deploy
[1:55] our platform irrespective of what industry you
are in, irrespective of what use case you have,
[2:01] irrespective of what your language your customers
speak and now we are global. So irrespective of
[2:05] what region of the world you're deployed and
that's what Cresta does. So with that I'll kick
[2:10] it off with the first question is that Craig
why don't you kind of walk us through what is
[2:14] the patient journey looks like for the clear
choice uh customers. Sure. So uh many of our
[2:20] patients uh have struggled you know um everybody
gets two sets of teeth for free. Your baby teeth
[2:26] then your adult teeth. The third set you have to
pay for and so it becomes very challenging. Um,
[2:32] and many many patients that we interact with,
it's the first thing they stop taking care of,
[2:38] which is their oral health. And so, as
their oral health kind of fades over time,
[2:43] um, and they don't get their treatment when
they need it, it can lead to really traumatic
[2:49] situations where patients are losing all of
their teeth or they need full remouth uh,
[2:53] uh, full mouth rehabilitation. So this is where
clear choice then inserts itself and we we market,
[2:59] we advertise and we try to connect those patients
that have that need, meet them where they are and
[3:05] get them into our centers with worldclass oral
surgeons and prostadonists to create that new
[3:10] smile and get them back to living. So there's a
lot of noise out there. There's a lot of dentists.
[3:15] There's a lot of people. And one of the biggest
things that we try to do is meet patients where
[3:19] they are, the time they're ready, um, where
they're where they're at in their process. Um,
[3:25] and we really try to focus on education. A lot of
people don't know what we do, don't know what we
[3:31] offer, they don't understand the entire process
from A to Z. So we have call centers. We have
[3:36] a lot of different opportunities to engage with
these patients to educate them along the journey.
[3:42] And the journey starts at awareness. They start
to understand, they search, they might see some
[3:46] TV ads, um they might start to get curious,
and then they land at my team. And my team is,
[3:52] we call it our connected care team because we
don't just call it a call center because our
[3:57] average conversations are around 8 to 10 minutes
with these patients. So, it's deep conversations,
[4:03] not a standard scheduling center. uh we go a
little deeper, we educate, we inform, we make
[4:08] sure that the patient feels really comfortable to
move forward because we want to help them in their
[4:13] journey and we're pretty much the starting point.
So as we started to say how do we scale, how do we
[4:18] leverage this? One of the first use cases that we
started when we started to look for a partner was
[4:24] after hours. So uh there's a lot of regulations on
outbound engagement with patients or consumers. So
[4:31] at 8:00 p.m. we can't be outbound calling. we
have some limitations. So what do we do? Well,
[4:37] we have all this great marketing and awareness and
we want to be able to meet those patients where
[4:41] they are. And so we said, how do we improve our
after hours engagement with patients when they're
[4:46] calling in at, you know, 9:00, 10:00 at night when
they're done with their day? How do we make sure
[4:51] that we're we're not missing that opportunity
to help them? And so this is where one of the
[4:55] first use cases was we were using a thirdparty
call center. We weren't really capturing the
[4:59] patient understanding. we were following up with
those patients but not when they were ready to
[5:03] have the conversation. And so this is where Cresta
helped us as our partner um for our AI agentic uh
[5:10] approach to make sure that we can have somebody
that's there trained in the same uh language,
[5:16] the same models, the same understanding of how
we engage with patients, but we don't need to
[5:21] staff humans. So we have 247 365 availability with
our AI agents to engage with these patients with
[5:28] their needs. And so this was the first use case.
Low risk because it's after hours. So we're not
[5:34] disrupting our core business, but it's a great
way and a great place to start so that we can
[5:38] better understand the needs of the patients and
follow them through this journey. And we've seen
[5:43] some strong engagement um and I just shared with
my CEO uh a story which was uh really interesting.
[5:50] Patient called in, engaged with our agent, was
going through the process and said I want to speak
[5:56] with a human. Okay. How many people have said,
"I want to speak with a human." Yeah, everybody.
[6:01] We're actually seeing in our initial pilots and
engagement, only about 20% of the patients that
[6:06] are calling in are asking for that human presence
because after about a minute or two, it starts
[6:12] to go right into a dialogue and the patients open
up and they start to feel comfortable and they're
[6:18] getting that that correspondence back from the
AI agent. And so, it's really interesting to see
[6:22] the walls break down. They start calling it ma'am
because our agent's female. Her name is Hope. And
[6:28] so they start calling her by her name. They ask
what her name is. They start engaging like it's
[6:31] a real human dialogue. So this is really breaking
down the walls, allowing the patients to feel more
[6:36] comfortable, ask the questions that they need to
feel that they're now educated and ready to move
[6:41] forward. So it was a great place to start and
it's been it's been working really well for us.
[6:45] Excellent. Excellent. I think great summarization.
So Craig I think next question I have is that what
[6:50] has been maybe two questions it's like what was
the first step you took in starting that journey
[6:55] and maybe kind of you piggyback on that well that
first step like what were the friction point it
[7:01] resolved what are the challenges it immediately
resolved yeah I think the the first is if anybody
[7:07] has interactions with call centers it it's a high
turnover business so we have to constantly train
[7:14] our human agents they're new It takes a ramp up
time. Then when you're in the health care segment
[7:20] and you start to talk about patients, there's
a sensitivity to it. A lot of these patients,
[7:24] as I mentioned, they're embarrassed with their
smile. They're not comfortable. So, you really
[7:29] need to be able to break these walls down. Um,
and so one of the challenges we had is getting
[7:34] our agents up to a consistent level that we can
replicate and make sure that we're not missing the
[7:40] opportunity to help these patients. That's where
the first use case in some of these these uh areas
[7:44] we started was how do we get consistency and the
AI agent never gets frustrated. The AI agent never
[7:51] hangs up on you. The AI agent really understands
everybody where they're at. So as patients can be
[7:58] a little bit more irritated as they're struggling
as maybe because they haven't taken care of their
[8:03] oral health, you know, it's sometimes it could
be difficult to understand what they're saying.
[8:07] The the AI really listens well. It understands,
it clarifies, it takes its time. So the patients
[8:14] are never feeling rushed. They're really feeling
educated and an understanding of how they are
[8:18] going to move forward in this treatment. So
it opens us up to really consistency um of
[8:24] our agents. So, you know, we don't need to train
the AI because the AI trains itself, right? And
[8:29] so we're seeing over time we want to consistently
get better. This is what we're trying to work on.
[8:34] We're still trying to find ways to improve it uh
at a quicker speed. Um but you can see that if a
[8:40] agent a human agent left today the replacement of
that agent is going to take us 2 to 3 months to
[8:46] get to a consistency level where with the AI agent
we have the same consistency level every day. So
[8:53] we've started to now expand out from our initial
use cases um and use it a little bit more within
[8:57] the days. Um we've started to do some additional
use cases on outbound. So we're starting to unlock
[9:03] the opportunities now that we have a baseline
that we feel comfortable with which is really
[9:07] important as you move forward and start this
process. Excellent answer. As model matures it
[9:13] seamlessly scale. So let me kind of pivot back to
my next questions a little bit about Cresta. How
[9:18] was your experience working with the Cresta team?
Any feedback on the engagement and the growth we
[9:23] have you know achieved collectively? Yeah. Um yeah
so we did an RFP process. uh we we've explored and
[9:29] I personally went through uh over 20 different
uh demos working with all the different AI
[9:36] organizations out there and there's a lot of good
AI. Um but one of the things that led us to Cresta
[9:43] and the partnership that we've had is um that that
native LLM the the flexibility because for us we
[9:49] really wanted a custom message for our patients.
Again, it's not a next best action workflow for
[9:55] us. We are engaging with 10, sometimes up to 15
minutes with patients to go through this back
[10:01] and forth process. And it's not easy. It's not
trying to get you to a schedule. It's trying to
[10:05] make sure that you're an actual the right patient.
So for us, qualification is really important to
[10:11] make sure that both they're the right patient,
they are looking for this treatment, and that
[10:15] we're the right provider for them. And so this is
where I think it's it's trying to make sure that
[10:20] it's it's a little bit of all sides. And when you
look at other uh AIs out there, there's some that
[10:26] are right off the box. Um there's some that will
accelerate in certain things. We also wanted to
[10:31] make sure that we gave ourselves room for more of
an omni channel approach because when we say we
[10:36] want to meet patients where they are, it doesn't
just mean voice. So we've also looked at text,
[10:41] SMS, chat. So different functionality to make sure
that we can communicate with patients in the ways
[10:47] that they want to be communicated with. And so
this is why that was one of the the main reasons
[10:52] that we wanted to make sure even if we start with
voice we want to have the flexibility to grow. No
[10:57] I think the multimodality of the platform is one
of the key attributes of crest. Uh I guess I think
[11:02] let's segue into the my favorite question is that
well you know when when AI is deployed it gives a
[11:08] surprising you know results initially some of
the results you were like wow and some of the
[11:13] results you kind of fine-tune. So maybe kind of
share some of your experience about going through
[11:18] that journey finding out certain functions work
really really well and when the other functions
[11:23] we need to refine and optimize the model. Yeah,
I think it's um one of the things that we started
[11:30] learning quickly is uh skilling, right? And so as
we started to look at our agents, it's what are
[11:36] they skilled to do? And a lot of our human agents,
we we realized that we train our human agents to
[11:42] do everything, right? So that's why it takes some
time to to get them up and running. So we saw that
[11:47] uh a lot of the scheduling and some of the
basic functionality was really strong. It was
[11:52] in line with our human human agent performance.
we're able to scale and get that up quickly. Um,
[11:57] we're able to address and contain a lot of our
inbound uh calling and and that that kind of touch
[12:02] point. Um, but some of the challenges that we had
early on was a little bit on the outbound. So,
[12:07] a lot of our outbound, if you guys are doing
any outbound engagement, um, answer rates vary,
[12:13] right? And so, it's a lot of voicemails. So,
the technicals, you know, making sure that the
[12:18] voicemail is being led, the beep. Now with some
of the technology with your iPhone, you can pick
[12:22] up the voicemail in the middle of it, you know,
so there's a lot of technical capabilities that
[12:26] you need to be aware of to make sure that you're
still addressing and getting to those patients or
[12:31] getting to them in the journey in the right
way. So we saw some technical feasibility
[12:36] um but overall the inbound approach which was
our primary focus um we saw some quick results on
[12:42] scheduling on it working in a very similar fashion
to our human agents which again gives us gives us
[12:48] the confidence to continue to push add additional
use cases and build out our our kind of AI agents
[12:53] into the marketplace. No, certainly certainly I
think one of the strong points of the platform
[12:57] is that you you know we provide the data mining
analytics on top of our platform. So you'll be
[13:03] able to see it instantaneously live how each agent
is performing and you can roll up all that report
[13:08] to see what business function it supports. So
let me kind of segue into more of a journey and a
[13:15] process perspective. I think you touched upon it a
little bit. You know AI adoption is not a project,
[13:20] it's a journey. So what lies ahead for clear
choice now that you know you embarked on this
[13:26] outbound patient like what lies ahead? What's
your vision? Yeah, we're we're trying. So our
[13:31] north star is to have AI be where AI needs to
be and have humans be where humans need to be.
[13:38] Um we don't see a world where it's AI or human.
We see a world where it's AI and human. And so
[13:44] uh Ping the CEO has I've seen presentations where
he's presented that way. But it's really aligned
[13:50] to I think clear choices approach as well. I mean
at the end of the day we want to spend time with
[13:54] patients and make them feel very comfortable
and educated to move forward. So all of the
[13:58] transactional work confirmations quick follow-ups
you know I have a quick question. We don't want
[14:04] to use human time and resources there. So this
is where if we can start to deploy more of those
[14:09] agents in that functionality, it actually frees
up our human capability and our human capacity to
[14:16] spend more time with the patients that don't know
where to go, that don't know where they're at in
[14:20] their journey. And so for us, this is where we're
trying to actually create more human capacity by
[14:25] investing in and starting to build out some
additional capacity. We we we a year ago,
[14:30] we had humans just making confirmation call
after confirmation call after confirmation
[14:34] call. It gets taxing. it gets it's you know the
team then starts to turn over right so if we can
[14:40] actually make it more dynamic I think we'll also
have less turnover on the human side because the
[14:45] AI will just consistently perform on a lot of
the transactional work that needs to be done on
[14:49] a day-to-day basis excellent and you touched upon
an excellent point I think's foundational thesis
[14:55] is that you know we believe that in the future of
AI is human plus AI where AI augments the human
[15:02] function and it takes over those repeated monitor
functions to do the AI. So, excellent point. So,
[15:09] maybe we kind of pivot back to some of the
challenges I think you faced selling this
[15:13] vision internally because I think there's a lot
of aspirations and then going to meet a lot of
[15:17] meetings, a lot of boardrooms across America. This
uh question has been discussed well are we ready
[15:23] for the AI? So, what was your journey look like
convincing internal leadership embarking on this
[15:30] AI journey? Yeah. Um the good news I mean Clear
Choice is very receptive to innovation. So we
[15:37] want to consistently be an innovative leader
in in the space. Um so there's a willingness
[15:42] to engage in these types of uh these topics,
these conversations, but we we want to invest
[15:49] as well. Um then it comes down to what's the right
investment. You can't invest everywhere. So where
[15:54] do you lay your bets and what's going to make the
most of it for the business and the ROI? So for
[15:58] me within clear choice I was kind of the business
lead for this and then you really need to partner
[16:03] with uh both your leadership teams as well as
it there is a lot of technical connectivity
[16:09] um we use different platforms and and bases
so API plugs in m making sure that it's simple
[16:15] um because complexity kills all momentum and so as
long as you have something that is simple easy you
[16:22] can stand it up um so with our journey uh I know a
lot of companies I have friends that I talked to,
[16:28] they go wide with their use cases and they start
with many use cases which will slow down the
[16:33] process. So we started with three use cases. Um
so it wasn't a heavy investment. Um but we really
[16:38] wanted to test have a P make sure that this works
and then build off of that success. And so for us,
[16:45] I think if you can present it in a way that
is very measurable, very tangible, and if
[16:50] the audience or your leadership has the right
mindset to be ready to have that conversation,
[16:56] I think you can move it forward. Excellent. Start
small. And I love the way you said, you know,
[17:01] complexity kills curiosity. So, excellent point.
Maybe last question is that I think one message
[17:08] for those who are on the edge they're doing
a pilot uh for their representative use cases
[17:13] of course it makes sense to pick something small
tangible measurable deploy that monitor scale but
[17:19] I mean one message for anybody who's like on the
fence maybe your message to them uh start now I
[17:27] I think uh we started this process uh it's almost
a year now so last September we started to really
[17:34] uh go through our process and in a year you
learn so much. Um and you really don't know
[17:40] what you don't know. And so uh along the the last
12 months uh we've pivoted. We've said maybe we we
[17:47] pull back here. We doubled down here. Um this is
working exactly how we anticipated it. This may be
[17:53] a little slower. We want to figure out what this
is. Um sometimes uh the way patients will interact
[17:59] with this again it's a lot of insights. It's a
lot of back and forth and iteration. Um, you know,
[18:05] we saw some uh our original AI voice agent
um was having very lengthy conversations. So,
[18:12] the patients were extremely educated and it was
good because it was driving better downfunnel
[18:16] engagement. But we started to see it's just
taking too long, right? And so now we're not
[18:21] being efficient. So the efficiency we're trying
to gain, we started to slow ourselves down. And so
[18:25] this is where we go back and forth, the learning,
the tweaks, the iteration. and now we found a
[18:29] good sweet spot where we're still having great
education, great engagement with our patients. Um,
[18:35] but now we're back down to that 8 to 9 minutes.
So, this is that fine-tuning, but you're never
[18:39] going to get to that place you want to get to
unless you start today. So, for me, I think if you
[18:44] you have to jump in, um, even though it's scary
and you don't know, nobody knows. I didn't know.
[18:50] You just have to have some conviction and give it
a try and really make sure that you pull through.
[18:55] That would be the other maybe piece of advice is
this isn't going to win by accident or randomly.
[19:00] I'm in it. I'm calling constantly. I know the
team hates that I do this, but I'm, you know,
[19:05] I'm the worst patient cuz I'm calling every day
and I'm testing and I'm seeing because I want to
[19:09] put myself in the patient's journey. I want to see
and understand what they're going through and what
[19:13] the engagement looks like and what questions are
being asked. And so when you live in the day of
[19:19] the patient, you really can then start to iterate
and better understand what what tweaks need to be
[19:23] made, how you want to adjust it, and then if with
the right partnership, you know, you can make some
[19:27] quick edits with AI, and you don't have to retrain
drastically. So um you can build good momentum,
[19:33] but you have to start now. Excellent, excellent
point. I think it resonates very well with the
[19:38] fireside ch session my colleague had
yesterday. Start now. It's a journey.
[19:44] Most of the business will pivot with the AI
because there's a lot of learning. So start
[19:48] now and stay at it. All righty. That's it. Thanks.
Uh thank you everyone. Thank you. Appreciate it.
