IBM Channel Executive Kareem Yusuf Eyes 50 Percent Ecosystem Revenue Goal With Upcoming Program Changes

‘I want to drive volume and velocity,’ IBM Senior Vice President for Ecosystem, Strategic Partners and Initiatives Kareem Yusuf said.

IBM’s quest to reach 50 percent of its revenue coming from partners means aligning its programs and incentives with business strategy, working with service partners on measuring influence, and better mapping of products to the customer segments they best serve.

These are some of the ways the Armonk, N.Y.-based provider of products ranging from mainframes to hybrid cloud to AI is looking into evolving its partner program for the AI era,

IBM Senior Vice President for Ecosystem, Strategic Partners and Initiatives Kareem Yusuf told CRN in an interview, promising more concrete details in the early weeks of 2026.

“I want to drive volume and velocity,” Yusuf said. “It’s not (just) about AI work. It’s about technology. We build products. … That, to me, has been one of the distinct things that I think has been important to elevate again as we think about partners.”

[RELATED: IBM To Acquire Real-Time Data Leader Confluent In Blockbuster $11B Deal]

IBM Partner Program

IBM reaching 50 percent of its revenue coming from partners would mark a jump from the about 40 percent share CEO Arvind Krishna reported back in 2023–itself an increase from 15 percent seen in years past.

Yusuf–a 26-plus-year IBM veteran who took on his current role in January–told CRN that this AI era of the channel is seeing business model changes.

For example, some services partners are now building assets and pursuing technology embed opportunities and managed services with their own assets, he said. At IBM, the vendor has been exploring the pricing models for AI and how its consumers like to consume the technology. (Yusuf spoke with CRN before IBM announced its planned $11 billion acquisition of Confluent).

So far, many enterprises still prefer the predictability of a licensing model, with Yusuf reminiscing over excitement around a token-dominated pricing scheme imagined in the early days of AI proliferation. “In some ways, as much as everything is different, everything is just the same,” he said.

Here’s more of what Yusuf told CRN about the vendor’s partner program plans in 2026 and the trends he sees in how IBM customers buy and consume AI products.

What’s your advice to solution providers as we enter 2026?

It’s going to be about scaling, volume and velocity of business.

Business expansion means the growth and reach outside of what I would call the largest concentration of our clients. And that means tweaking and tuning the ecosystem to go in that direction.

It’s also in terms of the products that then fit, and even the form in which they’re consumed. You begin to lead more towards SaaS (software-as-a-service) and subscription-type businesses than the traditional on-prem or perpetual licensing.

You begin to think about more of a focused product set that speaks in clear functions. Can be use case driven.

Platform notions so you can begin to do more cross selling, so we get more into land and expand.

It’s a lot of change, and it does require, in many ways, aligning and tuning our ecosystem, our partners.

Are you seeing solution providers combine business models more in this AI era?

The service partner–traditionally, their entire business model was really around, I’m oversimplifying, the work billable work around the installation, configuration, deployment, running of various technologies.

Part of the refactoring we have done this year (and) refocusing is to remember that from the service partner, the most important thing we need is influence, that they choose to take our technology into the solution space based upon what their priority is.

We’ve really changed our conversation with our service partners this year to really come back to–’What solution areas that you focused on? What technology do we have that maps into that? How can we design that into your offering?’

It’s back to that core basics, but it also means you begin to focus in and measure that and how much influence they’re driving.

They recommend you. The customer transacts. They get the deployment.

We’re beginning to see some service partners begin to do more of the, ‘We recommend (this product), but we’ll also do the resale.’

That, I’ve really seen a lot of change over the course of this year itself. There was really a clear class that said, ‘Only if the customer asks, we’re going to put it on our paper because it doesn’t really fit our accounting.’ Now, all of a sudden, accounting is changing. And more (partners) want to do that.

As service partners begin to fixate heavily on the assets they build, the embed opportunity comes in where these very technologies they were recommending, they now begin to embed in assets and managed services of their own that they then sell to the client. (That shift) is occurring more and more across the board.

What patterns are you seeing on how enterprise customers purchase and consume AI?

In some ways, as much as everything is different, everything is just the same. The most predominant path for AI to land within business is embedded within applications the business already buys.

You’re using (Microsoft email application) Outlook, now there’s (an AI chatbot) copilot in it. And if you think about that licensing model, what did they do, I think they tacked on some additional dollars per user.

You begin to look at most of the places where we will therefore embed AI in technology that we deliver, whether that’s Instana for application observability or Maximo for enterprise asset management. The licensing models for those products are established.

What we have discovered is those AI features really need to be ‘licensed’ in the context of the overall application. We all went through this. Are we going to sell it standalone? Is it a big add-on?

In that particular thread, AI enabling an application is becoming–tablestakes is the wrong word, but it’s like, of course you’re going to do that.

You’ve got the middle ground where you are building a completely new product that you couldn’t have built before the advent of, let’s just call it, generative AI.

Classics like that would be things we’ve done with products like IBM Concert around app resiliency.

Or even the code assistants–(IBM’s integration development environment (IDE) and modernization assistant) Project Bob most recently. Watsonx (Code Assistant).

While it’s a completely new product and literally would not exist without the AI within it, you have got to license it to how the customer feels they’re deriving value.

Across the board, subscription licensing as the model is obviously becoming the norm. Which, technically, is not quite the same as consumption. But are very closely related. I am not seeing a massive shift in terms of business to pure consumption because I don’t think it really works with their buying pattern.

When you’re a business and you’re trying to buy for your department to run your function, to run your procurement process, they don’t totally like that. There needs to be a little bit more structure, I will say, than just runaway consumption.

From a partner perspective, in my opinion, then you begin to make sure you do your incentive against our various licensing types. That, in effect, is what we are doing and what we are focusing on. Whatever license type we use, then you know what you get paid on against that.

How about agentic AI?

Agentic AI is going to heavily impact business process automation and all of that because the agents begin to embody the business processes.

We’ve got a product, Watsonx Orchestrate (with) a license structure. The additional element there that plays into the partners, though, is with our focus with the agent catalog.

And a core value proposition that we’ve pre-tested, pre-done all the interoperability with all these agents, so if you want to lay down a business process that touches these various endpoints, we probably have got an agent for that from the vendor in our catalog good to go, accelerate your time to value.

We have made as part of our value proposition to that build partner effectively landing the sale. The customer goes, ‘Yeah, I’m going to just use that.’ The sale could just flow from that notion. You bought it from my app store.

(Because build partners and ISVs) are really treating these things as add-on features, it’s a lot more about the convenience of, oh, the agents are in our catalog, it’s running, enter your license key or whatever your credential is because it’s captured in that application buy that you’ve already done.

While it’s very different in many ways, it’s very much the same. (I remember when) we all got ourselves wrapped up around inferencing, tokens and all of that.

The commodity price of tokens is so low–why bother? It’s not a pricing metric. But if you go back when we started, we all had pricing per token.

(An outcome-based approach to pricing), you see that more with the service partners.

SIs have been dabbling in outcome-based pricing for a while. It is a lot more difficult to do in terms of core technology purchase.

Sometimes it gets wrapped up in a bigger idea. But regardless, you’ve got to know what you’re going to carve out for the technology from the get go.

How do you see the consumption pricing model playing out looking ahead?

It’s a lot more about making it easy for someone to get started.

Not many of us do user-based pricing anymore, but I’ll just use user as an analogy–you want to use a bit of tech. You’re just a standalone user. Just get started consuming.

But by the time you get to a certain threshold–whether it’s number of users or amount of whatever you’re drawing in–it probably makes sense to move to a more standardized (pricing).

The other phrase people like to use is ‘pay as you go.’ And if you think about it in mobile phone terms, pay as you go is always more expensive than a standard plan once you hit a certain core volume.

What’s the future hold for IBM’s partner program?

The most important thing for IBM partners or potential IBM partners to understand is I am very, very focused on aligning programs and incentives with business strategy.

Hybrid Cloud, AI as plays. Subscription license, SaaS as delivery models. … There should be clear alignment.

What you have to ask yourself is, ‘Where are we focused? Where are we placing our investments? And does what we have here align?’

With my focus on expansion and reach, then the incentive … will obviously be more attractive in those kinds of segments (because) I want to drive volume and velocity.

Consumption as in deployment and use of what has been sold is critically important. So we begin to talk about an adoption accelerator and stuff like that. We announced that at (IBM conference) Think. And I’m leaning into those kinds of thinkings.

Co-marketing has to be tied to a purpose because–what are we all trying to do? Drive leads, progress deals, get yield, win revenue.

It’s so easy when these programs occur over time for them to just step back and become almost like a ‘benefits’ track. It’s not ‘benefits’ standalone. It’s a co-investment to drive outcomes.

As we step into the new year, … those are what I would call at the most generic level things to look out for.

What should partners know about how that strategy operates with IBM products?

As we look at different segments, we understand and appreciate, by the way, that we need specific products that really serve well in different segments.

You’ll see alignment and enablement. This is across the board where we say, ‘OK, these are the right, focused products for this kind of area and opportunity and here’s the work we’re doing to make sure they work better together and enable cross-sell and all that good, fun stuff.’

You’ll see that come through. But if I was to just use one simple, pithy term, my major focus is alignment with strategy. It’s not about willingness to spend. It’s about what we spent on. And for me, that is everything I can do to ensure that our partners are successful in the context of our strategy.

What do I expect from them (partners) every day–that’s evolving. These are discussions that we’ve been deep in.

What does it mean to effectively manage and support a partner? And what are the metrics that tell you you’re doing right or wrong? What does an ideal partner profile look like based upon the different things that you want to do?

What is the definition of a healthy channel? And, by the way, what levers do you pull to affect channel health? So I’m leaning a lot more actively into what I would call active management of the ecosystem.

I am actually leaning a lot more into that notion of what it means to actively manage a channel for growth.

If our big vision is that 50 percent of our revenue comes through the channel, how do you actually get there?

The channel and what it can touch–rough math, what, three times better than the rest, at least. Probably 5x if you’re kind of just trying to do all rounded math. So where does that volume come from?

How do you measure it? Because actually, it means their growth has to be going faster than what we can provide. … These are the mechanics in my head right now.

That brings you to a very important discussion on ways (internal IBM employees and partners work together). What it means to do handoffs, hand-ins, work together. (And) I’m actually maniacally focused on that right now because that’s where it breaks down if you don’t get that right.

How are you thinking about IBM’s relationship with distributors?

It serves a very important purpose for us. I think the emphasis is on the word ‘value-add.’ I think the good ones are also thinking deeply about that.

You want to do more volume, then that means you have to think very carefully about automation and integration with the VADs. And so that’s a big topic thread for us–the work we’re doing and actually work that we will continue.

How do you ensure that you’re actually enabling this volume to move through with velocity? That means we’ve got to have the right integrations and automations to make it really easy, and, quite frankly, drop their cost to serve.

We have been blessed with some pretty good VADs who engage and want to talk in that value-added way. We’re having conversations with them, similarly, around all the thinkings and directions we could go and getting feedback.

They remain an essential element of how we deliver value to market.

Are there non-AI opportunities you continue to see for the solution providers who’ve held back so far?

The customer who needs flash storage to run backup–yes, that business remains as important.

All work is not AI work. Then I’ll chuckle and I’ll say, ‘But what about if that flash storage box has some content-aware storage capabilities in it so it knows what’s on it?’ And all of a sudden, was the sale of that flash box an AI sale or not an AI sale?

At the end of the day, I actually come at this the other way. It’s not about AI work. It’s about technology. We build products.

That, to me, has been one of the distinct things that I think has been important to elevate again as we think about partners.

It is the distribution of a product. It amazes me how much this conversation can be had with no mention of what’s actually being distributed.

You have to tune and manage your channel based upon product. Different products have different characteristics and different things that need to be done. So that is not lost on me.

Why do we use a word that sounds like a singular noun to describe the most diverse, multivariate entity? We talk about ecosystem like it’s a singular thing. Meanwhile, it’s not.

You have to really think about the personas, the partner types in our context, and the products, these ingredients that come together.