What Is Agentic AI, and Will It Become a Game-Changing Marketing Tool?

Agentic AI is no longer a far-off concept – it’s already here, with platforms like Salesforce’s AgentForce and OpenAI’s ChatGPT Operator already beginning to change how businesses engage with customers. Agentic AI is a different beast, and has huge implications for productivity, automation, and even safety, because the more autonomous an AI system becomes, the more important it is to understand its limits, reasoning, and goals.

The potential of these AI agents is massive, which is why LBB’s April Summers wanted to sit down with Brad Bettinson, chief of innovation and solutions at Toronto’s JAM CRM, to gain a grounded perspective on the subject of how marketers can benefit from this autonomous system. Crucially, he doesn’t see this as a replacement for marketers, but as an empowering tool, explaining how JAM CRM places the focus on helping businesses get their CRM foundations right so that AI can truly add value and vice-versa.

Yes, the AI future is coming fast. But as Brad sees it, we’ve got some back-end plumbing to finish before handing over the keys.

LBB> You describe agentic AI as a game-changer, but also point out that most businesses aren't ready for full autonomy. What’s the biggest gap you see between where companies are today and where they need to be to really benefit from this tech?

Brad> The biggest thing holding our clients back hasn’t changed in 10 years: disconnected systems and disconnected data. If one part of the company knows something, we, as marketers, want to do something with that knowledge. We want to create a helping moment for your customer.

However, we need data to be centralised in one place so that the AI can tap into that. That's really Salesforce’s big push with Data Cloud, which a lot of companies are getting on board with. But it takes time to clean up all the disconnected data – there's a lot of band-aid ripping to upgrade and migrate systems. The promise of the tech is here, and you'll see some amazing stuff, but for the average company, there's a bit of a roadmap to get there.

LBB> You mention that sales and service are ahead of marketing in terms of data maturity. Why do you think marketing lags behind, and what will it take for marketing teams to catch up in an agentic AI environment?

Brad> Catching up to sales and service is a process that requires three things. First, we must help clients articulate a clear vision of what their CRM 3.0 looks like and how AI can empower it. Once ‘designed on paper’, it's easy to get C-level support for the vision - because once they see the customer experience they should be delivering - and on the scale they need to deliver it - it’s impossible to unsee it.

Secondly, they need a clear understanding of the data they want. We have a ‘Marketers Data Dictionary’ process where we build a roadmap for Marketing to get the data it needs. If marketing has the data it needs in its platforms, then the AI will have access to it too.

Thirdly, they need a seat at the MarTech transformation table. Yes these initiatives will be led by IT - but they need to act (and be treated) like one of the transformation's most important customers. Typically, we are guiding VP’s and CMO’s to make sure they get a seat at the table, and advising a list of ‘must haves’ they should bring with them to ensure that once they get the much pined for upgrades, they actually empower marketing and lay the ground work for AI acceleration.

LBB> Many businesses are eager to adopt AI but haven’t laid the “back-end plumbing,” as you put it. What are the non-negotiables they need to have in place before agentic AI can actually make a meaningful impact?

Brad> Generally, compliance comes first. Even for our agency, we're pretty far out, but we've looked at a few AI tools that couldn't explicitly clarify who owns the copyright of stuff coming out of these generative tools. We want to make sure information won't be used for training data, and that anything we put in is secure and safe. When it comes to AI, the non-negotiables are pretty huge – it's compliance, it's security, it's trust.

LBB> You talk about AI as something that should empower, not replace, marketers. What does that empowerment look like in practical terms? Can you share an example of how AI can elevate a marketer’s day-to-day work rather than automate it away?

Brad> The number one thing for us, or with marketing in general, is acceleration. Somebody needs to write a brief – they need to write content. Nowadays, they are briefing an AI to write the first draft – what we call the ‘bad draft’. A human would write the bad draft before and maybe take two or three hours. But now, a human with an AI can write a bad draft in two to three minutes. However, you need the human to take the bad draft and make it 10 out of 10. The speed at which our copy teams are working allows us scalability, so instead of just creating one version of the message, we can now create 10 and make them segment specific.

This means we're saving a lot of time by doing things faster, and then we're reinvesting that time into doing things better and handling more complexity. Because, if you want to get to, again, that CRM 3.0 of highly relevant and individualised communications, it's more complex. You need more versions, you need more considerations, and you need more data segments.

LBB> What do you think it will take for businesses to feel confident letting AI take the wheel in more meaningful ways?

Brad> AI will be taking the wheel on very repeatable transactional experiences first so, again, sales and service will lead there as well. For CRM and 1:1 there is a lot more nuance required. Marketing departments may not necessarily hand over the keys to the AI right away. They will need a ‘human in the loop’. If you're not a lawyer, are you going to let ChatGPT review and sign your legal documents for you? You might expect your lawyer to use AI to do it faster, but ultimately, you still need that trusted expert in the hot seat.

We are the trusted marketing experts in the hot seat that are going to help our clients do things faster. There should always be a wariness and a human in the loop – someone who can validate it before the button is pressed and the work goes out to clients. That’s why there's agentic AI behind the scenes (which we're all using), but then there's also AI in a live environment that has the autonomy to talk to clients directly. I think we are a while away from handing over the keys – any smart company is cautious about that.

LBB> There’s a lot of hype right now around ‘autonomous marketing’. How do you distinguish between the flashy demos and what’s realistically implementable for your clients at JAM CRM today?

Brad> It's a good question. It changes with the technology because there's a huge spectrum. It could be generative to write content on the platform to do things like send time optimisation, or on the data layer to look at all your customer records and find patterns and build segments.

As we validate these things for one client, we can then provide a very educated opinion to our other clients. This is part of the reason why clients rely on us, because a lot of times, internal teams have full time jobs already. They are busy getting work done, they can’t drop their current workload to research, test and validate every new AI tool that comes along. And they are working on the same problem in the same vertical, whereas we are busy solving that same problem for 20 different clients in 20 different vertices. So the learnings and insights tend to stack up for us a lot faster, and we can come in to accelerate change and evolution in ways fully maxed out internal teams just can’t. We're able to exchange feedback with the clients in rapid fashion. We can report back on something that came out three weeks ago, demonstrate how we've experimented with it for client X, and tell them if it seems like a good fit.

One other thing to point out is, there is the decision to add a product or tool and then there is the proper implementation of that product or tool. We often see the ‘promise’ of what is purchased falls down in practical application. A beautiful car without gas, a licensed driver,a place to go, and roads to drive on is just an expensive driveway ornament.

LBB> Looking ahead, what excites you most about the potential of agentic AI in CRM – and what do you think will surprise people as adoption deepens?

Brad> We have always been a customer-first strategy company. For any client we're helping, we always put our cards on the table. If you can be relevant, help the customer and do what's right for them – even sometimes at your own short-term investment as a business – long-term gains are there.

As all end clients, and all customers, get better AI, it's going to be filtering everything that's coming in. Their agentic AI is going to go out to company websites and use their apps – it's going to act on their behalf. It’s going to skip over anything non-relevant, filter out the noise that has permeated marketing by people, maybe try to play the shell game with end customers and bait them with benefits that aren't real benefits, and lead them on in ways that aren't customer-centric.

We are big proponents of customer-first, creating helping moments, and the AI is just increasing the need for companies to do that. Because, if you don't, you'll be part of the noise and you'll be filtered out.