IBM’s head of strategic partnerships on dealmaking in the AI age

It’s no secret that partnerships can supercharge growth: Plucky upstarts can hitch their wagon to an established incumbent with brand recognition, and large enterprises can turn the ship faster when the winds start to change.

Lately, it feels like AI partnerships are the only ones that really matter: OpenAI, Google and Anthropic are all looking to keep pace with, well, each other.

Revenue Brew spoke with Kareem Yusuf, SVP, ecosystem, strategic partners, and initiatives at IBM, about his transition from product to partnerships, and how he approaches AI in a very human way.

I noticed you have a robust background on the product and tech side of IBM. Now you head up strategic partnerships; how did that transition come about?

At my core I will always say I’m a product person. Most of my career has been spent in and around product. Indeed, I even consider this role very product-centric.

The partnerships that we need to build, the distribution that we need to build is obviously of our technology, around our technology. So I think the view is essentially having a technology leader come work on this side of the house.

It seems like tech, and specifically AI partnerships, are being announced everyday. What is it like to be in your position while there is so much movement in the space?

It really comes back to what is one’s core strategy. Because you’re right, in a world where things feel so frothy, you can very easily just get caught up.

The journey begins and ends with client value. Our technology is of no value if it’s not deployed at a client.

We hear about leaders struggling with the amount of applications, or getting those applications to work together. What is the biggest challenge right now to making that happen?

The space is moving so fast [that] yesterday’s architectures are being replaced by today’s architectures.

Because we’re in this heady time of real discovery and exploration, probably the biggest challenge is making sure you create appropriately and take an approach that customers can actually adopt and absorb.

When putting together these partnerships how do you ensure they will benefit all of those different revenue streams in a balanced manner?

In those discussions, you begin to say: What else might we do together? What other kind of value can you land at the client? So we begin to work on all of the angles. When you think about the landing of the value, then actually the mechanics of the financial simply become: What’s the best way to translate that into value capture?

I want to talk about the recent partnership with Anthropic. How does something like that come together?

In a very human way. You see that whiteboard behind me, usually you begin to lay out some frameworks.

Think about swim lanes. I’m a big fan of archetype thinking, so we’ll say, okay archetype one: big needle movers. I want to do how many of those. Archetype three: a volume play like, my idea on that agent catalog. I need to curate a lot of agents but in the context of domains. So that’s a volume play. How you configure for that is very different from how you configure to go land an Anthropic.

I want to focus on Groq for a little bit. How excited are you about that partnership, and where does it fit into the IBM ecosystem?

One [thing] they focused on is the optimization of the speed and cost of inferences, because every time you’re calling these models, that’s what they need to do. We have customers working at very large scales, so being able to now augment our [watsonx] Orchestrate, —which is where we do all our agent orchestration—running, building—being able to augment that with essentially what I would call an accelerator is a great value prop.

The Anthropic and Groq partnerships are very splashy. But are there other, maybe quieter, partnerships that not as many people are talking about?

I would look at all the work we’ve been doing in our agent catalog as an example because that’s where you get a lot of what I would call the smaller ones that make a real difference on the ground, when you think about agentic AI activation within the enterprise. There are a lot of companies that are building agents into their technology, whether on our platform or others, that we know become important to our enterprise.