Is Salesforce a CRM company? Or is it an AI company? The pivot towards generative AI everywhere over the past six months has made that an entirely valid question - and it’s not one that’s had the most consistent answer. I’ve had different responses from different sources in recent times. Some proudly declare that the firm is absolutely an AI firm. This week at Dreamforce, the message has effectively been ‘CRM company that does AI’.
I caught up with Ariel Kelman, recently returned to the Salesforce mothership after stints at AWS and Oracle and now three months into his role as Chief Marketing Officer, to get the definitive answer to the question. He told me:
We like to think of ourselves as a CRM company, just to really focus that we're not building general purpose AI and we're not building AI for cooking recipes. We’re very focused on making our business applications smarter.
Kelman has a very clear idea of the role of marketing in the tech sector and it’s very much grounded in educating the customer:
As a technology vendor, you need to put yourself in the shoes of the customers. They're dealing with lots of different technology vendors. They have complicated IT architectures, where it's always changing and we're always trying to look to simplify it. It's really important to them to understand where do we play today and what problems can we help them with? So that's why we try and be precise and try to think through how we can best communicate to our customers where we're investing, so they can think about where they should be looking.
At the moment that means getting across Salesforce’s position on AI at a time when the generative AI hype cycle is arguably out of control. Keep it real, is Kelman’s stance:
The practical way to do AI marketing is to be very use case focused. The more general and vague you are, the more risk there is you're not going to do something that's actually useful. So, we tend to be really practical. We build a lot. We have a core AI research team that's doing foundational work with a lot of AI functionality, working with each of our product teams, sales teams, and marketing teams. We're out there talking to customers and exploring different areas we can automate with generative AI.
That’s very sensible, but given that generative AI is only six months old, isn’t there an inherent risk in ‘betting the farm’ on such nascent tech? Kelman demurs:
I wouldn't look at it as a massive risk for the company. It's an opportunity, where the more effective we are in building this functionality in a way that works well and meets our customers trust expectations, the faster we can make them happy with the AI vision that they have in their head.
I think AI is top of mind for CEOs. It doesn't take a lot of imagination to see this as incredibly disruptive to your business in both good and bad ways, depending on what you do. If you have complex products and you can solve customer cases five times faster, just think about how much more competitive that makes you. Conversely, if your competitors do that before you, watch out! So, it's not surprising these CEOs are all over their technical teams now and asking, 'What's our AI strategy?'. There's a sense of urgency in our customer base.
But it is important to set expectations, he adds:
One of the things we did this year at Dreamforce was to be very clear in the keynote and in our press releases on the availability of everything we talked about, with very specific dates. A lot of this functionality is available today or will be available in the coming months. We also invested a lot of energy on our website on providing all of the AI use cases that we've talked about. There's AI pages for every product area, where we're giving our customers very thorough explanations of not only what you can do with it, but how it works. That’s information attendees can use to go back to other people in their companies and explain, 'Well, here's what we're going to do'.
On a wider note, Kelman has a marketing mindset that he pitches as ‘back-to-basics’
I'm just very practical in my approach. I have in my mind the things that you need to be good at marketing, to be successful and really to make good and effective use of the company's money. The company is giving you money to promote our business and we need to get the most value out of it.
For me, it always starts with metrics and measuring the major areas of marketing around awareness, around generating leads for our salespeople, helping our salespeople educate customers and close deals, helping drive our self-service businesses in the areas where we have that, and tuning some metrics that maybe were out of date. We've updated some of them simply because new technical approaches emerged, like using Deep Learning for attribution. You know, how much how much did a certain program contribute to your sales pipeline? After Dreamforce, I'm going to go look and say, 'Well, how did this do? Where are we getting a lot of value? What do we need to be better at?'
Three months in, he’s clearly been thinking about how Salesforce marketing is working:
I've just been going through different areas and making some adjustments around some of the basics. I've kind of been talking about back-to-basics marketing, just making sure we're consistently doing a great job of explaining our products on our website, of showing up on people's Google searches when they're looking for products, and measuring everything that we do effectively, making sure that we're aligning sales and marketing.
How this plays out in practice and what changes may result remains to be seen. I was curious about how Kelman views Salesforce+, the company’s ‘Netflix for business’ media platform. He tells me:
Salesforce+ does two things. It is a platform for streaming our live events and then it also is a way for on demand content to be sent out. I haven't seen that strategy of doing both on the same platform elsewhere. I think it's been very effective as a broadcast platform and we're going to continue to invest in it and make it easier to use. We added a bunch of features this year. For on demand content, I think it's still good, but it's one of the things I'm going to look at after Dreamforce.
There's always a debate around content access, he adds:
Do you use content to get people to register and put it behind a wall or do you use it just to educate people and give up an immediate opportunity to gather information? I’ve always viewed this as a data-driven exercise. For example, do we put a video on Salesforce+ and you have to register to see it or do we put it on YouTube or on Salesforce+ without a registration wall, where more people might look at it, but you lose that immediate opportunity. That exercise I haven’t gone through yet.
What I’d like to look at is how much business are we generating for those leads compared to how many people wanted to look at the content, but bounced when they saw there was a form? A lot of people don’t look if there’s a form to fill. That’s why I say back-to-basics. Often you have different departments looking at these numbers. That’s why I like to bring people together and say, ‘Let’s get back to first principles. What are we trying to accomplish?’
Another debate revolves around the degree of standardization there needs to be around marketing. Can one size fit all? Does a marketing campaign that works in San Francisco translate unchanged to Paris or Tokyo? Kelman argues:
I think most American technology companies think default American first and then internationalization is an afterthought versus coming up with an international content approach from the beginning. I think we could be more international from the beginning. It’s a common problem. A lot of times you have the people working on content at headquarters. I've always wondered about the arguments for having more decentralised teams that are closer to the customers. That is something I've always been in favor of, because the closer you are to the customer, the more you can iterate on what works for them, and the less you have to communicate internally.
If you make your UK operation more self-sufficient and make them more autonomous in their decisions, they're quicker to iterate with customers and more able to experiment. But your cost on that is you lose your consistency. You have to decide, is that consistency important? Part of that consideration is practical. There are some businesses where consistency is more important than others. You see American tech companies often swing back and forth depending on the leaders.
At Salesforce, he concludes, the approach is pretty balanced:
I personally like to have central groups, building systems and infrastructure for the marketing team so that you then enable distributed marketing teams to be able to execute their marketing campaigns in a creative way, but still follow our compliance regulations. The more that you can build your CRM tools to be great, that actually enables you to have more local creativity, but you don't sacrifice kind of a high quality bar standardization.