A buyer rings your contact centre at 09:14. The agent answers, opens the order, and the caller asks if they can finish paying on their phone instead because they're on a hands-free in the car. The agent says yes, drops a payment link into a text, and the caller taps through while still on the line. The card never enters your network. The agent stays in the conversation. Total time: 78 seconds.
That's multichannel customer service in 2026. Not channels stitched together with screen-pops and prayer. Channels that carry the same identity, the same payment rail, and the same security boundary, with handoffs that don't drop context.
We see this every day at Paytia, and we've watched the definition shift hard in the last two years. So here's what's actually changed, what's marketing fluff, and what your contact centre needs to plan for now.
What multichannel customer service means (and what it doesn't)#
Multichannel customer service is the practice of letting customers contact you, and complete the thing they came for, across more than one channel — voice, web chat, SMS, WhatsApp, email, in-app messaging, social DMs, and occasionally video. The bar isn't availability on each channel. The bar is completion on each channel without forcing the customer to start over.
The word "multichannel" used to mean something narrower. In 2018, a contact centre with a phone queue, a web form, and a Facebook inbox called itself multichannel. By 2022, the bar moved to omnichannel — the same conversation history visible to whichever agent picks up next, regardless of channel. By 2026, the bar has moved again: the same payment and verification rails available on every channel, so a customer can pay, verify themselves, or sign something without you having to bounce them out to a different system.
If your agents can answer on six channels but can only take a card payment on one of them, you're not running multichannel customer service. You're running a switchboard.
Why the bar moved: three forces driving the 2026 shift#
1. Payment trust collapsed in the audio channel
Phone calls used to be the trusted channel. Customers read out card numbers, agents typed them into a CRM, and nobody worried. That's gone. The Office for National Statistics reported 1.2 million card-not-present fraud cases in the year to March 2025, with the average loss above £600. Customers know. They hesitate to read card numbers out loud now, even to brands they've used for years.
Multichannel customer service in 2026 has to assume the customer doesn't want to say their card number — on any channel. DTMF masking on voice, tokenised payment links on SMS and chat, and stored credentials for repeat purchases all have to work together. Pick one and you've still got the friction problem on the other channels.
2. The agent is no longer the bottleneck. Compliance is.
The 2024 CCMA Voice of the Contact Centre report found agent capacity is roughly stable year-on-year — UK contact centres aren't short of people the way they were in 2022. What they are short of is channels they can take payments on without dragging PCI scope into the chat tool, the WhatsApp gateway, or the social DM platform.
That's why the channel-mix conversation has moved from "can we offer chat?" to "can we offer chat without the chat vendor becoming in-scope for PCI DSS v4.0.1?" The answer depends entirely on how the payment leaves the conversation. If the card data ever touches the chat platform, the chat platform is in scope. If you redirect to a hosted page or send a tokenised link, it isn't. Channel separation is the design pattern that makes this work.
3. Customers expect channel persistence
Forrester's 2025 customer-experience index found 64% of UK consumers expect to be able to switch channels mid-interaction without re-authenticating or repeating themselves. They don't care that your phone system and your chat platform are different vendors. They care that they've already proven who they are once today.
Multichannel customer service in 2026 means a customer who's authenticated on the phone can finish on chat. A customer who started a payment on a web form can complete it via SMS. A customer who got cut off can pick up where they left off without an agent having to manually piece the session back together.
The channel matrix that actually works#
Here's the rough split we see working in contact centres that handle payments. We're not going to tell you to add every channel — most contact centres should run four or five well, not nine badly.
Voice
Still the anchor channel for any transaction over £200 or any account-sensitive conversation. Voice carries trust nothing else does. The trick is keeping payment-in-conversation flows secure without forcing the customer onto a separate IVR. That's what DTMF masking does — the keypad tones get intercepted before they reach the agent or the recording, the card data is tokenised at the PSP, and the call continues. No transfers, no hold music, no "please wait while I put you through to our automated payment system".
Web chat
Now the dominant channel for first contact in retail and utilities. The risk: most chat platforms aren't designed for payment data, and customers will sometimes paste a card number into the box. Don't let them. Either drop the payment into a hosted modal that the chat platform never sees, or hand off to a payment link that opens in the customer's browser.
SMS and WhatsApp
Best for asynchronous payment requests — "here's the link to settle your invoice, valid for 24 hours". Don't try to do interactive payment back-and-forth in WhatsApp. Use it as a delivery mechanism for a tokenised payment link and call it done.
Slowest channel for resolution, but still the channel customers trust most for confirmations and receipts. Don't take payments in email. Send the link, log the click, move on.
In-app messaging
Underrated. If you've got an app, the in-app channel can carry stored credentials and biometric reauth — which means a £45 payment can settle in two taps without an agent ever being involved. Treat in-app as the lowest-friction channel and route the simplest transactions there first.
Social DMs
We don't recommend taking payments on social DMs. Use them for triage and routing to a channel that can actually close the loop.
What "good" looks like in 2026: five hard tests#
If you want to know whether your contact centre is doing multichannel customer service properly, run these five tests this quarter.
Test 1: The mid-call channel switch
Pick up the phone, get the agent to start a transaction, then tell them you'd prefer to finish on text. Can they send you a link without dropping the call, ending the session, or making you re-verify? If yes, you've got channel-aware payments. If they have to transfer you to a different team or end the call, you don't.
Test 2: The PCI scope check
Ask your CISO or PCI consultant which of your channel vendors are in scope for PCI DSS v4.0.1. If the answer includes your chat platform, your CRM, or your call recording vendor, your channel separation isn't working. Voice and chat should be in scope only at the point of payment, and not for storage or transmission of cardholder data.
Test 3: The first-touch resolution rate by channel
Look at your FTR rate per channel. If voice is 60% and chat is 35%, you've got a handoff problem. Customers are using chat to start things they have to finish on the phone. The fix isn't usually "train chat agents better" — it's "give chat the tools to close payments".
Test 4: The repeat-data test
Count how many times in a sample of recent multi-channel interactions a customer had to give the same piece of information twice. Name, email, postcode, order number, last four digits. If the average is above one, your channel persistence is broken.
Test 5: The dropout test
What percentage of customers who start a payment on chat or SMS abandon before completion? In our customer base, contact centres with good channel design see under 12% dropout on payment links. Above 25% means your link UX or your channel handoff is failing.
The PCI question, asked properly#
You can't talk about multichannel customer service in 2026 without talking about PCI DSS v4.0.1. The standard went live in March 2024 and the future-dated requirements all kicked in by 31 March 2025. There's no "we'll deal with it next year" left — it's all current.
The biggest change for contact centres has been requirement 6.4.3 and 11.6.1 — script integrity for payment pages — and requirement 8.4.2 for multifactor authentication on any access to the cardholder data environment. If your chat platform's iframe renders a payment form, every script on that page is now in scope. That's a problem for any contact centre that pasted a payment form into a chat widget. The cleanest solution is to send the customer to a hosted payment page on a domain you control, with strict script governance, rather than embedding the form in the chat vendor's environment.
We've written more on this in our PCI v4 script integrity guide if you want to go deeper. The short version: multichannel design and PCI design are now the same design problem.
What we don't recommend#
Some patterns we see contact centres reaching for that don't work in practice:
Adding every channel because a competitor did. If you can't staff WhatsApp 12 hours a day, don't offer WhatsApp. A channel with a four-hour response time damages trust more than not offering it at all.
Routing payment data through your CRM. Tempting because the CRM is where the order lives. Disastrous because now your CRM is in PCI scope. Tokenise at the PSP, store the token in the CRM, never the card.
Using AI voice agents for outbound payment calls. The tech works. The compliance picture doesn't, yet. Outbound calls from synthetic voices are heading toward FCA and Ofcom scrutiny, and we'd wait at least another 12 months before betting a customer-payment flow on it.
Where to start if you're rebuilding#
If you're planning a contact-centre refresh and multichannel is on the brief, here's the order we'd put it in.
First, fix the payment problem. Get DTMF masking on voice and tokenised payment links on at least one digital channel. That's the floor.
Second, sort the identity problem. If a customer authenticates on one channel, they shouldn't have to authenticate again when they move to another. Use a session token that's portable across your channel stack.
Third, sort the data problem. Same conversation history, same agent view, regardless of how the customer arrived. If your agents are reading three different consoles to handle one customer, the channels aren't really integrated.
Fourth, sort the routing problem. Get the customer to the right place first time, with skills-based routing across channels — not just voice. Most contact centres still route chat with first-available logic, which is why their FTR rates are 20 points lower on digital channels than voice.
The contact centres that get this right in 2026 won't have the most channels. They'll have the fewest gaps between them. That's the bar we're setting with our own customers, and it's the bar buyers should set when they're evaluating any contact-centre platform this year.
