Voice Assistant In Customer Experience

The difference becomes much clearer when you put it into a real product. Take ElevenLabs’ voice AI as an example. 𝟏. 𝐓𝐡𝐞 𝐛𝐚𝐬𝐞 𝐥𝐚𝐲𝐞𝐫: 𝐠𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐯𝐞 𝐀𝐈 𝐜𝐚𝐩𝐚𝐛𝐢𝐥𝐢𝐭𝐲 At the first layer, ElevenLabs can turn text, scripts, voice references, or multilingual content into natural speech. For many products, this appears as a generative AI feature: AI narration in an education platform automatic voiceover in a video tool multilingual dubbing for content natural voice response in a support system Here, the value is mainly output quality. The system is generating voice, but it is not necessarily running a workflow. 𝟐. 𝐓𝐡𝐞 𝐦𝐢𝐝𝐝𝐥𝐞 𝐥𝐚𝐲𝐞𝐫: 𝐀𝐈 𝐯𝐨𝐢𝐜𝐞 𝐚𝐠𝐞𝐧𝐭 The next layer is when voice becomes interactive. A generated voice is not an agent. But a voice interface that can listen, understand intent, respond in context, ask follow-up questions, and manage a conversation starts to look much closer to one. This is where voice AI becomes more than audio generation. It becomes an interaction layer. The user is not just listening to generated speech. They are talking to a system that can handle a role inside a conversation. 𝟑. 𝐓𝐡𝐞 𝐡𝐢𝐠𝐡𝐞𝐫 𝐥𝐚𝐲𝐞𝐫: 𝐚𝐠𝐞𝐧𝐭𝐢𝐜 𝐀𝐈 𝐬𝐲𝐬𝐭𝐞𝐦 The more interesting layer appears when the voice agent is connected to real company systems. CRM. Support tickets. Calendars. Order databases. Knowledge bases. Payment tools. Internal APIs. Telephony stacks. Workflow automation tools. At that point, the system can do more than speak naturally. It can check an order, update a customer record, create a ticket, schedule a demo, trigger a follow-up, escalate to a human, or write the result of the conversation back into the system. In short: Generative AI creates the voice. An AI agent uses voice to interact. An agentic system connects that interaction to tools, data, permissions, and workflows. Explore more here https://lnkd.in/g57BYwHz *The chart is simplified, but it gives us a useful starting point to map these ideas to an actual product.

🔮 Design Guidelines For Voice UX. Guidelines and Figma toolkits to design better voice UX for products that support or rely on audio input ↓ 🤔 People avoid voice UIs in public spaces, or for sensitive data. ✅ But do use them with audio assistants, learning apps, in-car UIs. ✅ Good conversations always move forward, not backwards. 🤔 The way humans speak is different from the way we write. 🤔 What people say isn’t always what they mean by saying it. ✅ First, define relevant user stories for your product. ✅ Sketch key use cases, then add detours, then edge cases. ✅ Design VUI personas: tone of voice, words, sentence structure. ✅ Listen to related human conversations, transcribe them. ✅ Write conversation flows for happy and unhappy paths. ✅ Add markers (Finally, Now, Next) to structure the dialogue. ✅ Accessibility: support shaky voices and speech impediments. ✅ Allow users to slow down or speed up output, or rephrase. ✅ Adjust speech patterns, e.g. speaking to children differently. 🚫 There are no errors or “wrong input” in human interactions. 🤔 Give people time to think: 8–10s is a good time to respond. ✅ Design for long silences, thick accents, slang and contradictions. Keep in mind that many people have been “burnt” with horrible, poorly designed automated phone systems. If your voice UX will come across even nearly as bad, don’t be surprised by a very low usage rate. You can’t replicate a long scrollable list in audio, so keep answers short, with max 3 options at a time. Instead of listing more options, ask one direct question and then branch out. Re-prompt or reframe when certainty is low. People choose their voice assistant based on the personality it conveys, and the friendliness it projects. So be deliberate in how you shape the tone, word choice and the melody of the voice. Don’t broadcast personality for repetitive tasks, but let is shine in a conversation. And: if you don’t assign a personality to your product, users will do it for you. So study how your customers speak. How exactly they explain the tasks your product must perform. The closer you get to a personal human interaction, the easier it will be to earn people’s trust. Useful resources: Voice Principles, by Ben Sauer https://lnkd.in/dQACgwue Voice UI Design System, by Orange https://lnkd.in/ezP-9QUu Designing A Voice Persona, by James Walsh https://lnkd.in/e3WXaxEC Voice UI Kit (Figma), by Shadiah Garwell https://lnkd.in/eGjJCWf7 Conversational UIs (Figma), by ServiceNow https://lnkd.in/enHVSEWP Voice UI Guide, by Lars Mäder https://vui.guide/ #ux #design

Customer support hasn't changed in 30+ years because nobody solved the math. AI just did. You call your telecom provider because the internet's down. You wait 15 mins. An agent picks up, puts you on hold again, comes back with “Have you tried unplugging it?..” You're now 27 mins in and even angrier than when you started. This is the reality for millions and millions of customers every day. And here's the brutal part: it's not about a lack of care. A company like Deutsche Telekom or Deliveroo handling millions of customers simply can't staff enough humans to answer every billing question, process every return, handle every outage notification in real time. So the system breaks, and everyone loses. Turns out, the fix was a voice. Or voice agents, to be precise. These agents handle order tracking, billing disputes, password resets, identity verification for account recovery (the stuff that makes up 60 percent of call volume). They respond in sub-second time. They work 24/7. They speak 70-plus languages. And when they hit smth complex… they hand off to a human with full context already loaded into your CRM. ElevenLabs' ElevenAgents platform is built specifically for this. You upload your SOPs, your knowledge base, your policies. The agent learns from that. E-commerce companies use it to generate return labels and suggest alternatives based on purchase history. Financial services use it for identity verification and dispute escalation with full audit trails, etc. Right now, they're processing over 250 million (!!) conversations every month. Revolut, Meesho, Immobiliare, Cisco – these aren't startups. These are companies handling MASSIVE volume. If you want to see what this actually looks like, their agents platform is worth a proper look: https://lnkd.in/gK-h2Xks Uncomfortable question: would you sacrifice your own time so a support agent could keep their job?

Voice AI ordering in the drive-thru. Every time I have posted about this in the past it gets a wave of contrary thoughts. Will it be ubiquitous in QSR? Will consumers rebel against it? Our data from mystery shopping every year remains a work in progress because there's a., not quite enough case studies relative to the field just yet to capture a full picture over time, and b., you have a lot of different tech fueling it across that landscape (as you'd expect). So, like any innovation, we're in that phase where you're not going to yield uniform feedback. All that said, I think everything I said will change soon enough. Taco Bell now has it in nearly 900 stores. No surprise to see the brand on the front lines. According to Omilia (the company working with Taco Bell), restaurants using Voice AI have reported higher employee retention than comparable locations without tit. Transaction times have remained on par with—or in some cases faster than—traditional order-taking methods. Guest satisfaction has also remained at or above restaurants not using AI. Omilia’s system uses proprietary small language models to recognize speech despite road noise, regional accents, menu modifications, and changing restaurant conditions. It also adapts automatically to individual restaurant menus, limited-time offers, and real-time inventory without requiring manual retraining. More here, and we'll see where this all goes next: https://lnkd.in/en6rnQ45

Dear Brand Leaders, Voice AI agents are here. Great news for customers, but making decisions for your business can be a complex process. Identifying the RIGHT partner isn't always easy. Here's what I recommend... Based on my experience as a buyer of emerging technologies, speaking at many events, meeting vendors, and hearing their roadmaps, ask solution providers the following questions: ⁣ ①Can the technology instantly access a customer's full history across all systems, so our team has the necessary information, and customers never need to repeat themselves? ② Does it track tone, sentiment, and emotion during every interaction, so our teams can detect frustration early and adjust in real time? ③ Can we measure the impact on resolution time, customer effort, and employee experience to improve results? I’m sharing 10 more must-ask questions in my upcoming blog article. (👇link in comments) ⁣ Most importantly....No silo decision making! To create exceptional Customer eXperiences, bring cross-functional teams together from the start and define the requirements collaboratively. ✓Your IT friends know what the systems can and can’t do. ✓Your operations team knows where the process breaks. ✓Your marketing team protects the brand promise. ✓Your CX and frontline teams know where customers get angry. Approaching this non-negotiable is crucial; otherwise, you'll spend more in the long run and risk losing the very people you aim to serve. ⁣ That's Doing CX Right®. ❤️Stacy⁣ Doing CX Right®‬ #VoiceAI #CustomerExperience #DigitalTransformation⁣

Many asked us for this update: How popular AI services handle data protection, confidentiality and professional secrecy. Now also with #Anthropic (#Claude) and #DeepL in addition to #OpenAI, #Microsoft and #Google and some Swiss providers. In short: ➡️ The overall situation has not really changed since last year; ➡️ Consumer versions are a "no go" for corporate users; ➡️Microsoft has still not managed to get its data protection matters in order with #Copilot & #Azure OpenAI Services concerning web grounding (i.e. the Internet search) – we do not understand why they cannot or do not want to offer a DPA and clear confidentiality undertakings re web search, as Google does (some say OpenAI has the same issue, but does not disclose it); many companies really have an issue with this; ➡️ Anthropic and OpenAI are in our view no option for users bound by professional/official secrecy (and the two providers do not seem to bother); if you want to use their models for such use cases, you need to use the hyperscalers Google, AWS or Microsoft (who host copies of their models and are anyhow more mature in our view); ➡️ There is still no fully satisfactory offering on the market from Hyperscalers for professional/official secrecy for small and medium-sized outfits – either the process of entering the prof sec addendum is not automated (Google) or abuse monitoring is not offered for these customers (Microsoft) – we hope this will change and we continue to press, as this is a hurdle in the use of AI (note: going through SaaS providers is not necessarily better, as they have to do their homework, too, which we often see is not the case); the alternative are on-prem solutions or national providers; ➡️The market is still not really transparent – it is a pain to compile and go through all the fine print and the various different offerings; no wonder people are unsure about what they may use; ➡️The contract is only one element of compliance, the correct configuration of the service is another; ➡️DeepL Pro, at least the standard offerings we see, are in our view contrary to a broad perception not ready for professional secrecy. These are our views and they relate to the #GDPR, the Swiss #DPA and Swiss professional and official secrecy. Thanks to my colleagues Lucian Hunger and Jonas Baeriswyl for their tireless effort to create and update the table. Yes, we could include other providers, but this is what our clients at VISCHER usually ask us about. Get the overview here in English https://lnkd.in/eGeT2Uz4 and here in German https://lnkd.in/euVPQ5hS. Share it. Our detailed blog explains our findings concerning the latest update, which is available in English at https://lnkd.in/eY9qijG6 and in German at https://lnkd.in/eRM8S-XG. More legal/risk management info re AI in the corporate field? See our previous blog posts at https://vischer.com/ai and subscribe to our them at the bottom of the page.

✴ AI Governance Blueprint via ISO Standards – The 4-Legged Stool✴ ➡ ISO42001: The Foundation for Responsible AI #ISO42001 is dedicated to AI governance, guiding organizations in managing AI-specific risks like bias, transparency, and accountability. Focus areas include: ✅Risk Management: Defines processes for identifying and mitigating AI risks, ensuring systems are fair, robust, and ethically aligned. ✅Ethics and Transparency: Promotes policies that encourage transparency in AI operations, data usage, and decision-making. ✅Continuous Monitoring: Emphasizes ongoing improvement, adapting AI practices to address new risks and regulatory updates. ➡#ISO27001: Securing the Data Backbone AI relies heavily on data, making ISO27001’s information security framework essential. It protects data integrity through: ✅Data Confidentiality and Integrity: Ensures data protection, crucial for trustworthy AI operations. ✅Security Risk Management: Provides a systematic approach to managing security risks and preparing for potential breaches. ✅Business Continuity: Offers guidelines for incident response, ensuring AI systems remain reliable. ➡ISO27701: Privacy Assurance in AI #ISO27701 builds on ISO27001, adding a layer of privacy controls to protect personally identifiable information (PII) that AI systems may process. Key areas include: ✅Privacy Governance: Ensures AI systems handle PII responsibly, in compliance with privacy laws like GDPR. ✅Data Minimization and Protection: Establishes guidelines for minimizing PII exposure and enhancing privacy through data protection measures. ✅Transparency in Data Processing: Promotes clear communication about data collection, use, and consent, building trust in AI-driven services. ➡ISO37301: Building a Culture of Compliance #ISO37301 cultivates a compliance-focused culture, supporting AI’s ethical and legal responsibilities. Contributions include: ✅Compliance Obligations: Helps organizations meet current and future regulatory standards for AI. ✅Transparency and Accountability: Reinforces transparent reporting and adherence to ethical standards, building stakeholder trust. ✅Compliance Risk Assessment: Identifies legal or reputational risks AI systems might pose, enabling proactive mitigation. ➡Why This Quartet? Combining these standards establishes a comprehensive compliance framework: 🥇1. Unified Risk and Privacy Management: Integrates AI-specific risk (ISO42001), data security (ISO27001), and privacy (ISO27701) with compliance (ISO37301), creating a holistic approach to risk mitigation. 🥈 2. Cross-Functional Alignment: Encourages collaboration across AI, IT, and compliance teams, fostering a unified response to AI risks and privacy concerns. 🥉 3. Continuous Improvement: ISO42001’s ongoing improvement cycle, supported by ISO27001’s security measures, ISO27701’s privacy protocols, and ISO37301’s compliance adaptability, ensures the framework remains resilient and adaptable to emerging challenges.

90% of voice AI deployments fail in 90 days. The demos look perfect. Then reality hits: legacy systems, compliance requirements, millions of concurrent calls. Most vendors can't bridge this gap: Why? Because demos run on perfect test data with a predictable load. Production means angry customers, decades-old systems, and security teams who've seen every vendor promise before. The gap between "impressive demo" and "trusted infrastructure" is where voice AI goes to die. Here's what separates demos from production: Enterprise voice AI needs 5 non-negotiables. Pillar 1: Reliability at massive scale. Contact centers can't afford to tell customers "the system is down, call back later." When your voice AI crashes during peak hours, every minute means frustrated customers and lost revenue. Voice.ai uses distributed edge inference with autoscaling. Architecture built for the worst-case scenario, not the demo. But uptime means nothing if you can't prove security. Pillar 2: Security and compliance. Security teams kill more voice AI projects than technical teams. One data breach, one GDPR violation, and you're explaining to the board why customer conversations leaked. Voice.ai encrypts in transit and at rest with ephemeral tokens. SOC 2, ISO 27001, GDPR, CCPA aren't checkboxes - they're architecture. Pillar 3: Integration capability. Enterprise buyers have seen this movie before. Vendor promises "seamless integration," then you discover it takes 6 months of custom development to connect to your legacy CRM. Voice.ai uses modular connectors built for Salesforce, Zendesk, ServiceNow, SAP, Oracle. Voice agents trigger refunds and password resets in real-time without custom code. Integration solves the technical challenge. But enterprises need control. Pillar 4: Customization and governance. Enterprises need brand tone, approved scripts, compliance guardrails. Voice.ai supports prompt templates with monitoring built in. Every interaction stays on-brand. But customization creates a new problem: how do you prove it's working? Pillar 5: Analytics and continuous improvement. Track handle time, resolution rate, sentiment, abandonment. Voice.ai logs every exchange for retraining and QA. Data drives improvement. These 5 pillars work as an integrated system. Voice.ai sits behind your phone number, replacing IVR logic. Compatible with Twilio Flex, Genesys Cloud, Avaya OneCloud. Systems-level architecture designed for enterprise scale. Most vendors build for demos, then panic when enterprises ask about SOC 2 compliance or what happens when call volume spikes 10x. If you're evaluating voice AI for contact centers at scale, this framework is your checklist. We can walk through how Voice.ai maps to your infrastructure and compliance needs.

The AI notetaker on your last call? It might already be a compliance problem. Here's what's happening right now, and what you need to know. Regulators in the EU, UK, and US are all moving at the same time on AI meeting tools. And the pace is faster than most teams realise. In the EU, Article 50 of the AI Act requires organisations to clearly label AI-generated content so people know they're not reading something a human wrote. That enforcement window opens around August 2026. Which sounds far away until you realise your organisation's templates, workflows, and vendor agreements haven't changed yet. In the UK, the ICO is clear. If your transcription tool identifies who's speaking, it's processing biometric data. That means you need a lawful basis. "We didn't think about it" is not a lawful basis. In the US, it's messier. One-party consent covers most states, but California and Illinois require everyone's agreement before you hit record. And there's a wave of class action litigation building around AI tools that identify speakers, arguing those voiceprints are biometric data under Illinois' BIPA law. These cases are not hypothetical. They're in court now. What does this mean for you as an assistant? You are often the person who sets up the meeting. Books the room. Sends the calendar invite. Turns on the recording. Which means you're also the person who needs to understand what that recording does once it exists. You don't need to be a lawyer to manage this risk. You need to know the seven steps, and you need to brief your principal before they find out about this the hard way. The tools haven't changed. The regulations around them have. Know the rules. Protect your executive. That's what strategic looks like.

You probably saw the "Amara" Dental clinic Inbound call agent I built some weeks ago. Now, this is different because I used a natural American accent, PLUS, I added a CRM to take all the necessary details from the inbound call the AI Voice Agent was attending to. Don't forget,Open AI,n8n,google Calendar and Vapi were used. So, meet Lily, who will be handling: The Patient Experience: When a patient calls (like "Joy Daniel"), the agent handles the entire booking process. It captures the patient's name, phone number, email, and insurance details, then checks the calendar for availability in real-time. The Technical Workflow: - n8n Integration: I used n8n to connect the agent to a Google Calendar and Airtable.  - Smart Data Extraction: Using a JavaScript node, the system automatically assigns appointment durations based on the service—30 minutes for a routine checkup versus 90 minutes for a root canal. - Knowledge Base: I integrated a PDF containing the clinic's pricing, insurance info, and opening hours. This allows the agent to answer specific questions about the business during the call. - Airtable CRM: All caller details, the call transcript, and the recording URL are populated into Airtable.  - Instant Notifications: The system sends confirmation emails to the patient and a follow-up alert to the company. Challenge? During the build, I encountered a glitch where records were being overwritten because data was firing too quickly. I resolved this by adding a 3-second wait period in the workflow to ensure the Airtable records update correctly. I used a voice from ElevenLabs that fits the client's needs—clear, professional, and easy for patients to understand. Automation like this makes the process more efficient for the clinic and more convenient for the patient. What do you think about this AI voice agent for appointment-based booking? #AI #Automation #n8n #Vapi #Airtable #DentalTech #CustomerExperience #ElevenLabs