All articles
Webinars

RCS for Business and AI Agents: Inside Our Webinar With Google and Centerfield

Clerk AI sat down with Google and Centerfield to talk about RCS for Business, AI agents, and Google click-to-message entry points. Real numbers from a Centerfield pilot paused on day 13 because conversions doubled.

Clerk AI Team|May 20, 202610 min read

Clerk AI recently hosted a webinar on RCS for Business and AI agents with Alex Allemand, who runs North American go to market for RCS Business at Google, and Aniketh Parmar, the CTO of Centerfield. Centerfield is one of the bigger performance marketing companies in the country, handling millions of messaging conversations a month across telecom, insurance, and home services. So when Aniketh talks about what is working with click-to-message and AI agents on RCS, it is not theory. It is from running it in production.

The full recording is below. If you want the highlights without the hour, keep reading.

From Transactional SMS to RCS for Business

Our CEO Alex Haque opened the session with a framing that ended up being the through line for the whole conversation:

If you're like me or Aniketh or Alex, you're probably receiving one of these millions of text messages that consumers receive every single day, which is a coupon, maybe a FedEx delivery notification, if it's not a scam. That is what's leading us to say that transactional messaging is dead on arrival.

Everybody in the room recognized it. Phone buzzes, you glance, you ignore. Apple and Google are pushing those notifications into spam folders. Consumers are doing the same on their own. SMS as a marketing channel is quietly being abandoned by the people it was supposed to reach.

The question we wanted to answer was whether anything is filling the gap. Based on what Google and Centerfield are seeing, yes.

Google: 275 Million RCS Users in the US

Alex Allemand walked through where RCS sits from Google's vantage point. The headline:

We have 275 million RCS-capable users. That makes it really the number one reach and conversational channel for brands to interact with their customers.

A year ago this was still considered emerging. Now it is the biggest conversational channel in the country, sitting natively in iMessage and Google Messages, end to end encrypted, with verified branding and rich media built into the inbox people already check. No app to download. No new behavior to teach.

Alex got specific about why performance teams are leaning in first:

Performance marketing in general sits at the intersection of where RCS has the most value. You have this branded message, verified branding, verified sender ID, rich conversational. And so you have trust. We see 3x performance conversion rate, and these kinds of things.

A 3x lift without changing the audience or the offer is the kind of number that usually only shows up in pitch decks.

Voice + RCS Pilot Paused After 13 Days

Aniketh had the story that got the most chat reactions during the webinar.

Centerfield had been running SMS for somewhere between five and seven years before they ever touched RCS for Business. It worked, but it never felt right:

It always felt like missed opportunity where you send a message, someone reads the message, but there was never a conversation happening. More often, we would send a message and then redirect into a different channel. So it was kind of like a moot point to have SMS as a channel for a conversation.

When Centerfield partnered with Clerk AI to pilot RCS with AI agents, the original plan was 5,000 contacts, 15 touchpoints, 45 days.

3x RCS for Business Conversion Rate

They paused it on day 13.

In first five to 10 days we almost started seeing the double conversion rate. So the more number of transfers are getting to the humans. Since we picked up a small program, we only had limited capacity on the call center side. So in 10, 15 days we almost doubled or tripled to a point where we are trying to abandon calls.

The pilot was not failing. It was overwhelming the call centers downstream. Centerfield regrouped, then scaled from 5,000 contacts to more than 50,000 over the next few weeks. Aniketh's read:

With the help of Clerk and Google we were able to do that, and we saw huge transformational results in our ecosystem since then.

The channel works. The bottleneck is usually whatever sits behind it. Read more about it in our internet provider case study.

AI Agents With Memory Across Voice and RCS

One thing we kept coming back to is how broken the voice plus SMS handoff has always been. Aniketh put it in a way every CX team will recognize:

Oftentimes we send a text message and then we put a phone number and then we move the consumer to the voice. When the phone call comes in, there is no context about the consumer or what the customer is interested in. With RCS we took a different approach where if we have ever spoken to a customer, then we have all that context within our ecosystem. The same memory shared across both ecosystem. That actually gives a much better experience to the customer than just sending a message across a singular channel.

This is the piece we have been building at Clerk AI. One agent that holds memory across RCS and voice, so the customer is not asked the same question twice and the brand does not lose the thread when the channel changes. Alex framed it from Google's side:

The old model was really siloed into voice and messaging separately. Now we see that combining those can really help create those amazing experiences. A messaging agent can really pick up the conversation from voice and vice versa. You can really build these rich and interesting and really great experiences for brands combining these technologies.

The remarketing side of this is where the math gets interesting. If a customer does not pick up a sales call, the RCS agent follows up in the same thread the brand already owns. If they have a question about pricing or service area, they can ask the agent at any hour and get an answer in seconds.

Google Click-to-Message: Search & Ads Entry Points

Google click-to-message sitelink entry point for RCS for Business shown under a brand listing in search results

The second big piece of the webinar was where these conversations start. Alex walked through what Google is rolling out:

We are introducing entry points from Google search and Maps. To me, this is the most critical piece that brands should really look into this year, is how to drive conversation, how to handhold conversation from these entry points onto RCS.

Click to message sitelinks already appear under organic search results for both branded and generic queries. Optimum, IHG, United Airlines, and Aetna are live with it. Maps integration is coming next quarter. Paid search ads now support a message extension that routes straight into an RCS agent, and Centerfield is running that pilot with Google and Clerk AI right now:

You combine both the power of search to the power of ads targeting with specific keywords, a specific audience that you can target, multiplying the performance of your ads. This is really a key pilot that we're running right now. If you're interested as an advertiser, feel free to reach out to Alex at Clerk AI. They are really helping us in running that pilot.

Search intent and conversational engagement in one continuous journey, with a verified brand identity from the first impression all the way through to conversion. That is new.

RCS for Business Compliance: FCC, CTIA, and 10DLC

The biggest blocker for most brands is not the technology. It is the internal review cycle. Aniketh's take matched what we hear from almost every conversation we have:

The technology is not the issue at this point. It's navigating the internal compliance and all that good stuff within the company. If you create a small POC with clear goals and a small team that focuses on those goals, and getting a buy-in from the leadership is really important. Otherwise you can be in the cycles of approval from one department to the other, and by that time it'll be six months and then you're going to lose that opportunity of early adopter.

This is where a lot of our work at Clerk AI happens. We take FCC rules, CTIA 10DLC, and state by state regulations seriously, because the brands we work with cannot afford not to. Our agents respect opt in and opt out by default. We provide consent capture widgets, double opt in flows, and the kind of documentation security and legal teams ask for before they sign off. The point is to compress a six month review cycle into a few weeks of real data the compliance team can sit with.

How Long It Takes to Launch an RCS for Business Agent

If you start this month, what does the rollout look like? Alex put a number on it:

Between signing with Clerk AI and designing the agent, maybe one or two weeks. And then getting the agent vetted is probably another one to two weeks. So it's pretty quick.

Two to four weeks from kickoff to a live, verified, branded agent talking to customers. The carrier vetting process has gotten a lot faster in the last year. Brand size is not the gate it used to be either.

What's Next for RCS for Business and AI Agents

We want customers to be obsessed with talking to brands. That is the version of this that gets us out of bed. Conversations that feel useful instead of conversations that feel like coupons.

Aniketh closed the webinar with the same advice he opened with:

Don't be shy, just go ahead and test it out. That's probably the only way to know the true power of this channel. We can sit down and try to convince everyone for months and years, but the only way to know is by doing a small pilot and then they can get enough real data, and then data will speak for itself.

The recording is above. If you want to scope a pilot for your own brand, that is what we do at Clerk AI. We will help you pick the SMS journey to convert, handle carrier registration and brand verification, design the agent flows, and stand up the analytics so your team can measure conversion against the baseline you already report on.

Scope an RCS pilot with us

We will help you pick the SMS journey to convert, handle carrier registration and brand verification, and stand up the analytics.