PRNEWS.IO Cuts Onboarding Time 90% with Veriff IDV Solution

We auto-sent AI-drafted replies to "simple" support tickets — password resets, basic how-to questions, feature lookups with a direct doc match. Narrow scope, 90% confidence threshold, real guardrails from day one. It worked for two weeks. Response time dropped from a several-hour queue wait to under a minute, with satisfaction scores holding steady. Then a ticket came in titled "password reset." The classifier matched it correctly — right category, high confidence — and auto-sent the standard reset instructions. What it didn't catch: the body described someone locked out of an account tied to an active, time-sensitive business process, written with real urgency. The reply was factually correct and completely tone-deaf. The customer wrote back angrier than before. The category classifier was doing exactly what it was built to do — matching the requested action, not reading the emotional register of the message. "Password reset" as a category says nothing about whether the situation around it is actually simple. The fix wasn't more categories or a higher confidence bar. It was a second, independent check: a single yes/no question asking whether the message shows urgency or frustration beyond a routine request, run before anything auto-sends. About 1 in 8 tickets that matched an eligible category tripped this check and got routed to a human instead — which tells you how much the category-only filter was quietly letting through. The lesson that's carried into every automation since: don't build a generic checklist of guardrails. Identify the one specific way this automation is most likely to embarrass you, and build the one check that catches exactly that. Full writeup, including the quieter "template fatigue" problem we found on repeat tickets and what we monitor now: https://lnkd.in/eVbkESvZ #CustomerSupport #AIAutomation #AITools #CX

Today's big move towards an estate agency Client Portal for an excellent customer experience. 👇 The 6-digit verification code. No one wants to remember a password, and clients of estate agents certainly shouldn't have to. To sign in to the JACOB Client Portal, future clients simply input their email address. The verification code is then immediately sent to their email. A fast and secure customer experience for the client to log onto the portal to view the latest listing statistics, feedback, and an analysis of the next best steps to get their home sold. The Client Portal doesn't take away the weekly updates from the agent, and instant viewing feedback over e.g. WhatsApp. Instead, the Client Portal supports the client on an anxious Saturday night. The sherry on top? Compliant onboarding is also handled directly in the client portal: ID, ML, and a full PIQ form, connected to the JACOB Operations system. This is part of a big operational software fully vibe coded by myself on Lovable since April. 💜 Jacob x

Customer due diligence doesn't end at onboarding. Businesses evolve. Ownership changes. New relationships are formed. What looked like a low-risk customer a year ago may present a very different risk profile today. Periodic reviews help institutions move beyond a point-in-time assessment and identify changes that could impact customer risk. They aren't just about meeting regulatory expectations they're about making better-informed decisions and strengthening risk management. The infographic highlights why looking beyond the initial onboarding process is essential for maintaining an accurate and up-to-date view of customer risk. Explore further: If you'd like to learn more about customer due diligence, periodic reviews and ongoing monitoring, explore our KYC Fundamentals course at https://lnkd.in/dnnCZ_GX

Customer Service Without Humans Isn’t Service at All Companies everywhere are racing to position AI as an all‑inclusive customer service solution, but the cracks in that strategy are becoming increasingly visible. My experience today inside the Shop app which was created by Shopify is a perfect example of why AI cannot yet replace human judgment. A fraudulent store called MyStore pushed a fake sale through the platform, triggering a phishing attempt designed to get customers to call a number and hand over personal information, ultimately giving scammers access to financial accounts. I recognized the pattern immediately and reported it, not for myself, but for the people who don’t understand how these scams work and trust the platform to protect them. Unfortunately, the AI chatbot inside Shop was completely ineffective. It offered no escalation path, no fraud‑reporting workflow, no human option, and no meaningful guidance. It simply looped through generic responses, making it painfully clear that the system wasn’t built to solve problems it was built to deflect them. This is the deeper issue with companies attempting to replace customer service with AI. AI is a powerful tool, but it is still only a tool. It lacks the human judgment needed to interpret nuance, the flexibility to handle edge cases, and the accountability required when real harm is at stake. When platforms remove human escalation entirely, they send an unintended but unmistakable message…. efficiency matters more than customer safety. That is not innovation it is negligence disguised as automation! AI excels at triaging simple questions, automating repetitive tasks, and enhancing human agents, but it is not ready to replace humans in situations requiring fraud detection, emotional intelligence, ethical decision‑making, or relational trust. Customer service is not just transactional, it is deeply human. My experience with Shop wasn’t just a frustrating interaction, it was a case study in what happens when companies chase automation without considering the human cost. AI is extraordinary, but it is not a substitute for people. Not yet. And until companies recognize that, customers will continue to fall through the gaps created by systems that prioritize scale over safety, and automation over accountability.

New customer onboarding call. Migration plans are being set. A business stakeholder asked why a monitor was set to high sensitivity on one table and low on twelve nearly identical ones next to it. Nobody in the room could answer. The person who'd set it had left the team a year earlier. The commit history said what changed, not why. Best guess, reconstructed secondhand over a call: that one table had caused a bad week once, an incident nobody wanted repeated, so someone turned the dial up and moved on. The other twelve tables never had their bad week, so they stayed on the default and nobody thought twice about it. That's not a documentation failure exactly. Nobody sits down and writes a memo explaining a threshold they changed in five seconds during an incident. It's just how configuration actually accumulates in a real team: reactively, under pressure, by whoever was on call that day, encoded into a setting instead of a sentence anywhere a new hire could find it. We ended up helping the customer rebuild the reasoning from scratch, table by table, because reconstructing intent after the fact takes longer than it would have taken to write one line down in the moment. Every threshold in your stack has a story behind it. Most of those stories only exist in someone's head, and heads leave companies.

Your transfer log says success. Your customer just got asked for their account number again. A customer explains a billing error twice before the call reaches a specialist. The specialist opens a blank case and starts from zero. Chat completed its escalation rule. Telephony accepted the transfer. The agent followed the script. Every system reported success. A real handoff review checks four things: did the trigger match, did the conversation reach the right agent, did the context arrive, did ownership actually change. Miss one and the log still reads green while the customer starts over. The fix isn't a better prompt. It's treating chat, voice, and ticketing as one operating layer instead of three systems that each declare victory on their own. Customers don't track how many tools touched their issue. They track whether anyone remembered it. What's the last handoff that made a customer repeat themselves? #CustomerExperience #ContactCenter https://lnkd.in/emNnge59

For Remitly’s Kevin Duarte, the formula for effective customer onboarding is simple: verify every legitimate customer on the first attempt. In our recent webinar, he broke down why: → First impressions last. Verifying a customer immediately builds trust, and that trust is the foundation of a long-term partnership. → Seconds matter. Every additional second during onboarding can erode that trust, making speed and accuracy a key focus for Kevin’s team. → Small solutions have a big impact. When AI handles a bounded task like address formatting, cleaner inputs lead to more reliable decisions and a smoother experience when it matters most. For the full conversation, request the recording at the link in comments.

Your customer onboarding is part of your product think about what happens after someone decides to buy from you. Inquiry → Payment → Welcome → Setup → Delivery → Follow-up On paper, that's a simple process but customer confidence can be lost at every single handoff. The customer asks a question, nobody responds quickly. Confidence drops. They make payment, nobody confirms what happens next. Confidence drops. They're told someone will contact them, nobody does. Confidence drops. They've provided information to one person, then another employee asks for the same information again. Confidence drops. The service is delivered, but nobody checks whether everything went well. Confidence drops. None of these problems necessarily started with the customer facing employee, they started behind the scenes. A missed handoff, an unclear owner, a delayed notification a disconnected system, a team member who doesn't know the next step. That's why I don't separate operations from customer experience, your customer experience is the visible output of your operations. A strong onboarding process should make the customer feel: They know what they're doing, I know what happens next, someone is paying attention, I don't have to chase them. That's customer confidence and confidence is what turns a transaction into trust. When I look at an onboarding journey, I ask: Inquiry → How quickly is the customer acknowledged? Payment → Do they immediately know what happens next? Welcome → Does the customer feel expected? Setup → Is responsibility clear? Delivery → Are expectations being managed? Follow-up → Does anyone close the loop? Because customers don't see your internal processes, they experience the consequences of them. If your customer experience feels inconsistent, let's talk about what's happening behind the scenes. #CustomerExperience #CustomerSuccess #Operations #BusinessOperations #OperationalExcellence #SystemsThinking

The trend of customer SUPPORT talking over you is a fascinating addition over the last decade. I just repeated "please listen to me" five times before finally hanging up. If anyone in a support role reporting to me disregarded a customer in this manner, it would be the last of my customers they spoke with. There is no need for employees to send you on a wild goose chase when phone trees and AI do such a stellar job. I was locked out of PayPal for 9 months because AT&T couldn't validate me. Now a health insurance site can't validate me, every attempt to login says I'm locked out until I reset my password, and every reset does the same. I can't get out of the phone tree jungle, there are no human or tech support options, that system won't validate me, and there is no other number to give me. And nobody stopped talking long enough to understand my problem. So I guess I'm just supposed to wait to die now, lol. I didn't (TOS roulette aside) agree to a security throuple connecting my data through these two companies. My relationship with each of those companies was separate and private. They decided otherwise without involving me. Social Networking became Social Media. (boo!) Please stop calling whatever this is Customer Support because it's absolutely not that. It is corporate support for firing humans at the expense of customer experience, which is short-sighted at best. We are locked out of accounts for months while algorithms, not humans, decide if we may access our assets, or services we paid for. Usually over mistakes THEY made (and money they don't want to give back). "Script reading" coupled with complete lack of authority may as well be AI already. I won't do business with a company who doesn't think enough of my needs as a customer to allow me to express them. If that means my next phone is 2 tin cans and a string, or I never stream another show, so be it. There's never anything good on anyway.