Customer Experience

Most Retrieval-Augmented Generation (RAG) pipelines today stop at a single task — retrieve, generate, and respond. That model works, but it’s 𝗻𝗼𝘁 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁. It doesn’t adapt, retain memory, or coordinate reasoning across multiple tools. That’s where 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗥𝗔𝗚 changes the game. 𝗔 𝗦𝗺𝗮𝗿𝘁𝗲𝗿 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 𝗳𝗼𝗿 𝗔𝗱𝗮𝗽𝘁𝗶𝘃𝗲 𝗥𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 In a traditional RAG setup, the LLM acts as a passive generator. In an 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗥𝗔𝗚 system, it becomes an 𝗮𝗰𝘁𝗶𝘃𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺-𝘀𝗼𝗹𝘃𝗲𝗿 — supported by a network of specialized components that collaborate like an intelligent team. Here’s how it works: 𝗔𝗴𝗲𝗻𝘁 𝗢𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗼𝗿 — The decision-maker that interprets user intent and routes requests to the right tools or agents. It’s the core logic layer that turns a static flow into an adaptive system. 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗿 — Maintains awareness across turns, retaining relevant context and passing it to the LLM. This eliminates “context resets” and improves answer consistency over time. 𝗠𝗲𝗺𝗼𝗿𝘆 𝗟𝗮𝘆𝗲𝗿 — Divided into Short-Term (session-based) and Long-Term (persistent or vector-based) memory, it allows the system to 𝗹𝗲𝗮𝗿𝗻 𝗳𝗿𝗼𝗺 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲. Every interaction strengthens the model’s knowledge base. 𝗞𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗟𝗮𝘆𝗲𝗿 — The foundation. It combines similarity search, embeddings, and multi-granular document segmentation (sentence, paragraph, recursive) for precision retrieval. 𝗧𝗼𝗼𝗹 𝗟𝗮𝘆𝗲𝗿 — Includes the Search Tool, Vector Store Tool, and Code Interpreter Tool — each acting as a functional agent that executes specialized tasks and returns structured outputs. 𝗙𝗲𝗲𝗱𝗯𝗮𝗰𝗸 𝗟𝗼𝗼𝗽 — Every user response feeds insights back into the vector store, creating a continuous learning and improvement cycle. 𝗪𝗵𝘆 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 Agentic RAG transforms an LLM from a passive responder into a 𝗰𝗼𝗴𝗻𝗶𝘁𝗶𝘃𝗲 𝗲𝗻𝗴𝗶𝗻𝗲 capable of reasoning, memory, and self-optimization. This shift isn’t just technical — it’s strategic It defines how AI systems will evolve inside organizations: from one-off assistants to adaptive agents that understand context, learn continuously, and execute with autonomy.

Google's new AI Agent demo just blew me away. This is the future of customer service. Here's what it can do: - Identifies plants from video you share on the spot - Recommends products based on your location & needs - Swaps items in your cart for better alternatives - Schedules services with access to live calendars - Negotiates discounts (with manager approval) - Sends personalised care instructions to your phone The multi-modal approach changes everything: → Show the agent what you're looking at via video → Talk about products without knowing their names → Get recommendations tailored to your garden's conditions → Complete your entire shopping journey in one conversation What impressed me most is how it maintains context: 1. Remembers your purchase history 2. Understands your preferences 3. Adapts to your location-specific needs 4. Picks up conversations where you left off It can even generate QR codes for loyal customers and instantly update your information across multiple backend systems. My takeaway: I believe customer service is evolving in two different directions. Some will embrace AI for speed and efficiency. Others will seek out human connection even at a premium. The most successful companies won't choose one or the other. They'll understand when a customer needs empathy versus when they just need their problem solved quickly. It's about giving users a choice depending on their needs in the moment. The real winners will master this balance. Follow me Alex Banks for daily AI highlights and insights. P.S. If you want to start leveraging AI today, subscribe to my newsletter: https://lnkd.in/e9UcAxTd

🗺️ AirBnB Customer Journey Blueprint, a wonderful practical example of how to visualize the entire customer experience for 2 personas, across 8 touch points, with user policies, UI screens and all interactions with the customer service — all on one single page. AirBnB Customer Journey (Google Drive): https://lnkd.in/eKsTjrp4 Spotify Customer Journey (High-res): https://lnkd.in/eX3NBWbJ Now, unlike AirBnB, your product might not need a mapping against user policies. However, it might need other lanes that would be more relevant for your team. E.g. include relevant findings and recommendations from UX research. List key actions needed for next stage. Add relevant UX metrics and unsuccessful touchpoints. That last bit is often missing. Yet customer journeys are often non-linear, with unpredictable entry points, and integrations way beyond the final stage of a customer journey map. It’s in those moments when things leave a perfect path that a product’s UX is actually stress tested. So consider mapping unsuccessful touchpoints as well — failures, error messages, conflicts, incompatibilities, warnings, connectivity issues, eventual lock-outs and frequent log-outs, authentication issues, outages and urgent support inquiries. Even further than that: each team could be able to zoom into specific touch points and attach links to quotes, photos, videos, prototypes, design system docs and Figma files. Perhaps even highlight the desired future state. Technical challenges and pain points. Those unsuccessful states. Now, that would be a remarkable reference to use in the beginning of every design sprint. Such mappings are often overlooked, but they can be very impactful. Not only is it a very tangible way to visualize UX, but it’s also easy to understand, remember and relate to daily — potentially for all teams in the entire organization. And that's something only few artefacts can do. Useful resources: Free Template: Customer Journey Mapping, by Taras Bakusevych https://lnkd.in/e-emkh5A Free Template: End-To-End User Experience Map (Figma), by Justin Tan https://lnkd.in/eir9jg7J Customer Journey Map Template (Figma), by Ed Biden https://lnkd.in/evaUP4kz Free Figma/Miro User Journey Maps Templates https://lnkd.in/etSB7VqB User Journey Maps vs. Service Blueprints (+ Templates) https://lnkd.in/e-JSYtwW UX Mapping Methods (+ Miro/Figma Templates) https://lnkd.in/en3Vje4t #ux #design

If I was the Head of Events at a $100M ARR SaaS, and had a $1,000,000 event budget, here’s the exact playbook I’d run (with budget): BACKGROUND: Replicating SaaS is only getting easier. Building moats is not. The best moat you can build is your community. That should be the #1 focus of every GTM team. Here’s the event program: 1. Flagship Event 60% of budget is going here. Pair on the back of a major product announcement. Use sponsorship and ticket sales to generate another $500k - $1m Attendance: 50% customers, 20% BoFu, 10% partners, 10% MoFu Invest in niche influencers. Make your event the “it” event. 2. Field Marketing Target 15-20 cities Bring in 1-3 partners. Total cost per city should be < $10k including travel Attendance: 20% Customers, 20% BoFu, 40% MoFu, 20% ToFu Get your SDR team onboard. Watch response rates go from <1% for cold outbound to >18% with dinner invites 3. Webinars / Virtual Full time role + $1,000 per event for promotion & speaker gifts 3 objectives here Build relationships with speakers Generate content You can’t be in every city every month. Use this to maintain mindshare throughout the year Attendance: 10% Customers, 10% BoFu, 40% MoFu, 40% ToFu (I'd use Accelevents to manage 1 through 3) 4. 3rd Party Events Only invest in the top 3-5 industry events Spend $50k - $100k per event Host a micro event at each You can’t build a moat from 3rd party events so I’d focus on our owned event program. 5. Content distribution Any remaining budget goes to content distribution. You’re building a brand around your events. Allocate 90% of budget to creating and distributing short form video. Not lengthy sessions. Look, it’s a lot of work. But it can define your brand. And your brand will be the only thing that matters when products get commoditized. P.S. Your CEO and CMO need to believe in events. What would you change? How would you allocate your budget? One platform can run all your owned events. Check out Accelevents --> https://hubs.la/Q03d3MZ70

Introducing the web's first market map of the Product Analytics Market: I was floored when I couldn't find one of these online. Surely, Gartner or CBInsights or A16Z would have created one? It turns out not. So I spent the past 3 months: • Talking with 25 buyers • Researching the space myself • Interviewing 5 product leaders at key players This is what I learned about the most significant players in each space: (that PMs and product people need to know) 1. Core Product Analytics Platforms     The foundational tools for tracking user behavior and product performance Amplitude : The leader, an all-in-one platform for PMs to master their data Mixpanel : The leader in easy UX and pioneer in event-based analytics Heap | by Contentsquare: The automatic event tracking and real-time insights leader 2. A/B Testing & Experimentation     Platforms for analysis Optimizely : The premier tool for sophisticated A/B and multivariate testing VWO : The best for combining A/B testing with heatmaps and session recordings AB Tasty: The all-in-one solution for testing, personalization, and AI-driven insights 3. Feedback & Session Recording     Capture qualitative insights and visualize user interactions Medallia: The top choice for comprehensive experience management Hotjar | by Contentsquare: The go-to for visual feedback and user behavior insights Fullstory: The best for detailed session replay and user interaction analysis 4. Open-Source Solutions     Customizable, free analytics platforms for data sovereignty Matomo: The robust, privacy-focused open-source analytics platform Plausible Analytics: The lightweight, privacy-first analytics solution PostHog: The versatile, open source product analytics tool 5. Mobile & App Analytics     Specialized tools for mobile and app performance analysis UXCam: The best for in-depth mobile user interaction insights Localytics: The leader in user engagement and lifecycle management Flurry Analytics: The comprehensive, free mobile analytics platform 6. Data Collection & Integration     Gather and unify data across platforms Segment: The top choice for effortless customer data unification Informatica: The enterprise-grade solution for data integration and governance Talend: The flexible, open-source data integration tool 7. General BI & Data Viz     Non-product specific tools for data analysis and visualization Tableau: The leader in interactive, rich data visualization Power BI: The best for deep integration with Microsoft tools Looker: The modern BI tool for customizable, real-time insights 8. Decision Automation & AI     Systems for automated insights and decisions Databricks: The unified platform for data and AI collaboration DataRobot: The leader in automated machine learning and AI Alteryx: The comprehensive solution for analytics automation Check out the full infographic to see where your favorite tools fit and discover new platforms to enhance your product analytics stack.

Last week, I described four design patterns for AI agentic workflows that I believe will drive significant progress: Reflection, Tool use, Planning and Multi-agent collaboration. Instead of having an LLM generate its final output directly, an agentic workflow prompts the LLM multiple times, giving it opportunities to build step by step to higher-quality output. Here, I'd like to discuss Reflection. It's relatively quick to implement, and I've seen it lead to surprising performance gains. You may have had the experience of prompting ChatGPT/Claude/Gemini, receiving unsatisfactory output, delivering critical feedback to help the LLM improve its response, and then getting a better response. What if you automate the step of delivering critical feedback, so the model automatically criticizes its own output and improves its response? This is the crux of Reflection. Take the task of asking an LLM to write code. We can prompt it to generate the desired code directly to carry out some task X. Then, we can prompt it to reflect on its own output, perhaps as follows: Here’s code intended for task X: [previously generated code] Check the code carefully for correctness, style, and efficiency, and give constructive criticism for how to improve it. Sometimes this causes the LLM to spot problems and come up with constructive suggestions. Next, we can prompt the LLM with context including (i) the previously generated code and (ii) the constructive feedback, and ask it to use the feedback to rewrite the code. This can lead to a better response. Repeating the criticism/rewrite process might yield further improvements. This self-reflection process allows the LLM to spot gaps and improve its output on a variety of tasks including producing code, writing text, and answering questions. And we can go beyond self-reflection by giving the LLM tools that help evaluate its output; for example, running its code through a few unit tests to check whether it generates correct results on test cases or searching the web to double-check text output. Then it can reflect on any errors it found and come up with ideas for improvement. Further, we can implement Reflection using a multi-agent framework. I've found it convenient to create two agents, one prompted to generate good outputs and the other prompted to give constructive criticism of the first agent's output. The resulting discussion between the two agents leads to improved responses. Reflection is a relatively basic type of agentic workflow, but I've been delighted by how much it improved my applications’ results. If you’re interested in learning more about reflection, I recommend: - Self-Refine: Iterative Refinement with Self-Feedback, by Madaan et al. (2023) - Reflexion: Language Agents with Verbal Reinforcement Learning, by Shinn et al. (2023) - CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing, by Gou et al. (2024) [Original text: https://lnkd.in/g4bTuWtU ]

Last week, I shared how Gen AI is moving us from the age of information to the age of intelligence. Technology is changing rapidly and the way customers shop and buy is changing, too. We need to understand how the customer journey is evolving in order to drive customer connection today. That is our bread and butter at HubSpot - we’re deeply curious about customer behavior! So I want to share one important shift we’re seeing and what go-to-market teams can do to adapt. Traditionally, when a customer wants to learn more about your product or service, what have they done? They go to your website and explore. They click on different pages, filter for information that’s relevant to them, and sort through pages to find what they need. But today, even if your website is user-friendly and beautiful, all that clicking is becoming too much work. We now live in the era of ChatGPT, where customers can find exactly what they need without ever having to leave a simple chat box. Plus, they can use natural language to easily have a conversation. It's no surprise that 55% of businesses predict that by 2024, most people will turn to chatbots over search engines for answers (HubSpot Research). That’s why now, when customers land on your website, they don’t want to click, filter, and sort. They want to have an easy, 1:1, helpful conversation. That means as customers consider new products they are moving from clicks to conversations. So, what should you do? It's time to embrace bots. To get started, experiment with a marketing bot for your website. Train your bot on all of your website content and whitepapers so it can quickly answer questions about products, pricing, and case studies—specific to your customer's needs. At HubSpot, we introduced a Gen AI-powered chatbot to our website earlier this year and the results have been promising: 78% of chatters' questions have been fully answered by our bot, and these customers have higher satisfaction scores. Once you have your marketing bot in place, consider adding a support bot. The goal is to answer repetitive questions and connect customers with knowledge base content automatically. A bot will not only free up your support reps to focus on more complex problems, but it will delight your customers to get fast, personalized help. In the age of AI, customers don’t want to convert on your website, they want to converse with you. How has your GTM team experimented with chatbots? What are you learning? #ConversationalAI #HubSpot #HubSpotAI

3 out of 5 product pages I work on have bounce rates above 70%. Most often, this is due to - 1. Low-quality ad traffic 2. Poor UX on the PDP  3. Basic, non-engaging product images  4. Lack of information / USPs about the product  5. First fold of the PDP not being optimized If you're sure that your targeting is on-point and that good quality traffic lands on your PDPs, then reducing your bounce rate should be the #1 priority. As, a lower bounce rate equals better conversions and higher revenue. In this example, using Mothercare PLC's PDP, I’ve implemented changes that can reduce the bounce rate by building trust in the brand and the product. Below are the 6 changes I recommend a/b testing - 1. Moving the product name in the first fold along with other details like reviews, price. In this case, I've also changed the product name a little, adding 'Pack of 3' which creates value for the amount they're paying. 2. Add a lifestyle image of your product being worn or used. More important for fashion brands where size is a common concern. 3. Add key USPs about the product. Especially in this case where the parent wants to know whether it's a good material and easy to change. 4. Add info on what size the baby is wearing, enabling the shopper to be more confident about their sizing decision. More important here as baby clothing is often bought by relatives and friends and gifted at baby showers, birthdays. 5. Replacing the wishlist with a 'Buy Now' CTA as that can help the user checkout immediately. 6. Optimizing the area around the add to cart by adding information on the shipping timeline, free shipping, and returns. Add a prominent 'View offers' option which can motivate users to complete their purchase. Other than that, I've implemented the following UX changes to improve the shopping experience: 1. Improved the image browsing experience by adding thumbnails and a sneak peek of the next image. 2. Made the size selection easier and more intuitive. Changing the section's title to 'Select Size', and showing a prominent, selected size by default. 3. Increased the CTA size for Add to Cart and Buy Now. Found this useful? Let me know in the comments! P.S. Being able to identify these minor details is a skill worth developing. The best way to do this is by looking at competitor websites. Carefully observing what elements they use on their PDPs can help you understand the industry and its requirements better.

Voice AI is more than just plugging in an LLM. It's an orchestration challenge involving complex AI coordination across STT, TTS and LLMs, low-latency processing, and context & integration with external systems and tools. Let's start with the basics: ---- Real-time Transcription (STT) Low-latency transcription (<200ms) from providers like Deepgram ensures real-time responsiveness. ---- Voice Activity Detection (VAD) Essential for handling human interruptions smoothly, with tools such as WebRTC VAD or LiveKit Turn Detection ---- Language Model Integration (LLM) Select your reasoning engine carefully—GPT-4 for reliability, Claude for nuanced conversations, or Llama 3 for flexibility and open-source options. ---- Real-Time Text-to-Speech (TTS) Natural-sounding speech from providers like Eleven Labs, Cartesia or Play.ht enhances user experience. ---- Contextual Noise Filtering Implement custom noise-cancellation models to effectively isolate speech from real-world background noise (TV, traffic, family chatter). ---- Infrastructure & Scalability Deploy on infrastructure designed for low-latency, real-time scaling (WebSockets, Kubernetes, cloud infrastructure from AWS/Azure/GCP). ---- Observability & Iterative Improvement Continuous improvement through monitoring tools like Prometheus, Grafana, and OpenTelemetry ensures stable and reliable voice agents. 📍You can assemble this stack yourself or streamline the entire process using integrated API-first platforms like Vapi. Check it out here ➡️https://bit.ly/4bOgYLh What do you think? How will voice AI tech stacks evolve from here?

I've coached 400+ CEOs. The best ones don't communicate better. They communicate differently. While average leaders wing it, great ones use proven methods that turn conversations into opportunities. After 20+ years studying top performers, I've identified 7 communication systems that separate good from great. (Save this. You'll need it for your next big meeting.) 1. The 3 Levels of Listening Stop listening to reply. Start listening to understand. Level 1: You're thinking about your response Level 2: You're focused on their words Level 3: You're reading the room—energy, tone, silence One CEO used this to uncover why his top performer was really leaving. Saved a $10M account. 2. What? So What? Now What? Transform rambling updates into decisive action. What = The facts (30 seconds max) So What = Why it matters to the business Now What = The specific decision needed Cut meeting time by 40%. 3. PREP Method Never fumble another investor question. Point: Your answer in one sentence Reason: Why you believe it Example: Proof from your business Point: Reinforce your answer Practice this for 5 minutes daily. Sound prepared always. 4. RACI Matrix Kill confusion before it starts. Responsible: Who does the work Accountable: Who owns success/failure (only ONE person) Consulted: Who gives input Informed: Who needs updates Projects with clear RACI are 3x more likely to succeed. 5. Story of Self/Us/Now Move hearts, not just minds. Story of Self: Why YOU care (personal conviction) Story of Us: Our shared challenge Story of Now: The urgent choice we face This framework has helped politicians win. It'll help you raise capital or inspire your team to meet a big goal. 6. The Pyramid Principle Get board approval in half the time. Start with your recommendation Give 3 supporting arguments (max) Order by impact (strongest first) Data goes last, not first McKinsey consultants swear by this. So should you. 7. COIN Feedback Model Make tough conversations productive. Context: When and where it happened Observation: What you saw (facts only) Impact: The business consequence Next: Agreed action steps No more avoided conversations. No more resentment. Your next funding round, key hire, or major deal doesn't depend on working harder. It depends on communicating better. Because in the end, leadership isn't about having all the answers. It's about asking better questions, listening deeper, and communicating with precision. Your team is waiting for you to lead like this. P.S. Want a PDF of my Leadership Communication Cheat Sheet? Get it free: https://lnkd.in/dbaSN9fJ ♻️ Repost to help a founder level up their communication. Follow Eric Partaker for more leadership tools.