IBM CEO Arvind Krishna Aims To Grow Revenue Coming From Partners To 50 Percent, Touts ‘Huge Amount’ Of Money To Be Made In AI

“We want to work with the channel. I’ve stood here the previous few years and said, ‘Look, we are about 30 percent of our total revenue from clients coming from the channel.’ … I’d like to make that 50 percent of the revenue, not 30 percent of the revenue. And that is a hard commitment,” IBM CEO Arvind Krishna said on stage at the XChange Best of Breed Conference.

(From L-R) CRN Executive Editor Jennifer Follett, IBM CEO Arvind Krishna, CRN Editor News Steven Burke

(From L-R) CRN Executive Editor Jennifer Follett, IBM CEO Arvind Krishna, CRN Editor News Steven Burke

IBM Chairman and CEO Arvind Krishna repeated his pledge to create more opportunities for partners, laying out multiple opportunities for them to make money in artificial intelligence alongside the tech giant.

“We want to work with the channel,” Krishna told an audience of solution providers at CRN parent The Channel Company’s 2023 XChange Best of Breed Conference being held through Tuesday in Atlanta.

“I’ve stood here the previous few years and said, ‘Look, we are about 30 percent of our total revenue from clients coming from the channel.’ … I’d like to make that 50 percent of the revenue, not 30 percent of the revenue. And we will go down that path. And we’re going to continue going down this path. And that is a hard commitment.”

[RELATED: IBM Channel Chief Woolley: Our AI Is Not For ‘Writing Poems About Your Dog’]

IBM CEO Arvind Krishna

The Armonk, N.Y.-based tech giant has invested in its channel partner program, adding more than 4,000 new partners in the last 12 months, Krishna said on stage Monday.

IBM has about 55,000 worldwide channel partners, 12,000 of them in North America, according to CRN’s 2023 Channel Chiefs.

Phil Godwin, president and COO of Clear Technologies, an Addison, Texas-based IBM partner and No. 389 on CRN’s 2023 Solution Provider 500, told CRN that he welcomes IBM’s efforts to grow revenue from partners to 50 percent.

The nw IBM Partner Plus program has been a positive for his partnership, Godwin said.

And while IBM has shown impressive AI technology, the partner role in bringing it to market is still in early days.

“I think the AI thing is interesting,” Godwin said. “I think we have to figure out where we fit in that.”

Krishna also used his time on stage to express condolences for the people affected by the Israel-Gaza conflict, to dismiss concerns around a Google AI monopoly and to illustrate partner opportunities with IBM subsidiary Red Hat.

Here’s what else Krishna had to say.

The Israel-Gaza Conflict

All of our sympathies and our help has to go out to the people who are impacted. And we do have a number of employees in the region, and some of them are deeply impacted.

So we should try and help as much as we can. … If I look at the impact on the economy … we can see it in the gyrations of oil prices this morning. And I think the last three and a half years have shown us that supply chain shocks are really bad for prices and the overall economy.

And I expect there to be some supply chain shocks from what’s going on there right now. That much we can see.

The rest, I’m hoping it gets resolved quickly. But this is not my area of expertise.

80 Percent Of AI Opportunity With Deployment

We can debate whether [AI is] the most important technology in the last 30 years. Second, third. It is definitely in the top three.

When I say ‘top three,’ I would say internet, and I’ll give that circa 1995 with the advent of Netscape. … We can take smartphones, that’s 2007. And within two or three years, almost everybody had smartphones. … What’s the impact on the world [from AI]? … [Consulting giant] McKinsey’s number is $4 trillion and change … of productivity by 2030.

OK, normally if you go into the tech space … about 20 percent of that we would say is the opportunity in tech. So you say $4.4 trillion, 20 percent, $880 billion—that’s a lot of opportunity. … I do think it’s going to impact almost everything we do.

And that’s a massive opportunity for us to all lean into and participate in.

A lot of the attention right now is on the inventors. … The vast majority of that hundreds of billions is going to be in the deployment, not in the invention.

And that is where it’s important for everybody to understand—bringing it to life, to bring value to somebody’s business, is what most of you do. Eighty percent of that opportunity is going to lie kind of here, not with those who invent AI.

IBM’s Place In The AI Market

If we go to IBM’s solution, we are not going to do a B2C model.

We’re not going to go put out an app that anybody can use with hundreds of millions or billions of people using it. That’s not our skill set.

Our skill set is going to provide much more in B2B use cases.

So when I think about that, when I think about people’s internal HR systems, I think about recruiting. I think about accounts payable. I think about trying to improve how you deploy your IT systems. … That is where we’re going to go.

So with the approach we’ve taken, we are going to endorse people using other models. That’s part of the whole goal with open innovation. That’s why you see us partnering with both Meta, with Hugging Face and anybody else who would like to get their models brought into this world.

We’re also going to do our own models. We brought the first set out, it’s a family of models that we announced about two weeks ago. Our first one is the Granite set of models.

These you can think of the range as anywhere from 10 [billion] to 100 billion parameters. … I’ll give you an example. So one of the Granite models we have does 130 programming languages … The first instance we were using is for COBOL.

So it can understand the COBOL language. It can take a whole environment—a few of our clients have anywhere from 3 [million] to 30 million lines of COBOL. And it can take chunks of COBOL and make it readable. Because a lot of these as you know produce unreadable stuff, stuff that is not maintainable.

The whole goal is to make it maintainable, make it easy and go forward. So that is one example.

Cyber is another. Ansible is yet another. Climate change is not in the Granite model, but there is another model that we did with NASA. So you can imagine quite a few like this.

And our goal is to bring these together with an environment that you can use to run these so you don’t need to run these only on a public cloud. You can run these, and people care about privacy also.

And our goal then is to work with partners to really go and help our clients in all the different domains and on the fly. Maybe an audit. Maybe in compliance. And then you have got to worry about sovereignty when you’re talking about some clients who want to keep the data on their own premises or certainly in their own nations.

AI And The Workforce

Do I believe that repetitive white-collar work will get replaced by AI, which is in the larger family of automation? Absolutely.

Now, the key is repetitive white-collar work, how many jobs are 100 percent repetitive only? Very few. … I do believe that 20 [percent] to 30 percent of back-office roles will get automated away with AI. But that’s an aggregate. This will be over five years. You get that from churn, by the way, in the workforce. So this is not a reduction of workforce statement.

I actually went further. I said, if we become more productive, that means we are more competitive. If we are more competitive, that means we win more business. It will actually lead to a total addition of the workforce. So when I said these 8,000 [roles can be replaced by AI and automation]… about 20 percent of our people could be thought to be in back-office roles, that’s sort of the upper case. … Then we go to the first six months of this year.

We did automate away probably about 800 roles. Although we hired 8,000 people. … The last 150 years of the economy has shown that companies that are more productive tend to hire more people because they’re actually winning more business. They’re actually more profitable.

They can afford to reinvest. ... There is a lot that gets talked about that this is going to cause labor displacement. So the onus goes on us. We have to help people with reskilling.

But if I think about deployment, then the vast majority of people need to have domain knowledge.

So for IBM, I do expect that G&A does decrease. … If you can drop your G&A down from 8 percent to 6 percent, that’s a great thing.

You’re not going to necessarily add 2 percent to the bottom line. You can put it back into sales, into marketing, into R&D, into those things that are more value-creating for your clients.

AI And Services Business

If you have businesses [focused] on … deploying information technology, that’s some of you.

Maybe doing process work for your clients, that’s some of you. … Maintaining applications for your clients. … If I begin to go through all those, I think then … if I look at scripts like Ansible, 60 percent of the work in doing those scripts can be done by AI already.

That is not 60 percent total productivity. We should be careful. That’s only half the time. … The other half is still spent on deployment, gathering, or working with others and understanding what to do. But that does mean it’s 30 percent productivity.

If as a result—I will turn around and tell all of you, as we intend to do—you have got to give some of that back to your client. … I think those of you who don’t embrace AI for the fact of making higher quality and making work more productive will be disadvantaged compared to those who do.

Now, this is not a clean four or five times, but it is going to be 30 percent more productivity in the next one to five years, depending on which topic you’re in.

If you don’t actually code, maintain older applications, it is probably going to be near the three-year time frame. Not instant.

But in terms of writing new things, in terms of maintaining, in terms of deploying IT, it absolutely is going to come in the next six to 18 months. So that’s the task that I will tell all of you, get ready and get going.

Now for those of you who do more BPO [business process outsourcing] work, that’s going to be here and now. You can see that sort of coming. … How many people who use call centers already have chatbots? Already have AI answers? Already have text-to-speech? … This is kind of the next wave that is going to come in on many of those functions.

Will Google Dominate AI?

How many of you have your own ChatGPT paid account? … And how many times is it 100 percent accurate? Zero.

To me, this is pretty intuitive. As you think about it, it is learning from lots and lots of data. … If you have lots of data, there is inconsistent data. … Think about if you’re talking to a friend, how often do they have a perfect recollection of the same event that you were at? … So that’s the nature of these technologies.

So I don’t think it’s a nightmare. Those two [Microsoft and Google] are going to go fight it out in the search market. I think that’s good for them. More competition is good. They can go do that.

Will it really impact all of us to that extent? No. I think it is a more useful tool. … When I was in college it was all about, ‘Oh, my God. You’ve got programmable calculators. People are not going to learn math in college because you’re going to use programmable calculators.’

Really? I don’t think it makes us better or worse. These are the tools. And this is yet another tool. These tools are imperfect. They’re not perfect. But that’s the nature of tools. But it frees us up from mundane work.

So if I can make search better … if I can get summaries, if I can get all of that, it makes us more productive as human beings. … The number of people of working age is decreasing. My guess is some of you see that. Wages are going up. It’s tougher to hire people.

We need technologies that help make people more productive. We need that, otherwise our quality of life is going to decrease.

So for me, using all this, using these technologies, is what is needed to make our companies more productive. But in the overall competitiveness of the economy, it is essential. Otherwise, we lose going forward.

AI Dangers

One danger that people talk about is is this going to aid terrorism and bioterrorism more than before?

With respect, I don’t think those people need the help of AI. There are a lot of things they can do that are damaging all by themselves. Just the fact that there is social media. There is [the ability to search] digital archives. This gives people a lot of access to a lot of techniques out there.

Could AI maybe help make it faster? OK, fine. But these things take years of planning in any case.

Second danger they’re concerned about, which probably is a concern, is misinformation, specifically in terms of attacking elections. And it’ll be termed ‘election fraud,’ but it’s more about trying to sway public opinion. … I probably agree, it could make it a bit worse.

It’s not really the issue. This issue has gone back to the days of the printing press, of newspapers, then radio and then television. Each time it becomes faster and faster. … In AI, you can personalize the misinformation to go convey it to somebody. That’s a possible danger.

That will be addressed through legislation. Not technology. … Think about it right now in terms of the danger of the use cases. … When you use it to answer questions on life and death issues, maybe be a little bit more careful. In that case, put up really tough guardrails around where the answers are coming from.

But if you are using it internally for your employees to answer questions, I don’t think there’s much danger. Go ahead and use it.

If I use it to create systems with full autonomy in the physical world where you can have it drive heavy machinery. Oh, I would be careful right now by doing it in the wide open. Putting it on a mock road … I think that’s a fine use to do.

AI Regulations

I think that in the near term, we should talk about regulated use cases, not the technology, part one.

Part two, whatever we do, we must be allowed to do open innovation. Otherwise you get revenue captured by just a few companies.

And three, I think that people who develop models should be held accountable. So is that going to be legal liability? Is it going to be accountability to some kind of publishing of what you do? … I actually believe that legal liability is fine. If I produce a model, then I should be held legally liable for what’s in that model. … I know that’s not a widely held view. But I think that’s a way to put teeth into some regulations.

IBM AI Strategy

I believe that as we go through every enterprise process, you can begin to automate away pieces of it.

So I take our HR process. I divide it into two. … The first is not going to get very automated. What is that? Leadership development. Helping decide the composition of a team. Helping assess what should be people’s development path forward, etc.

That is still going to be very, very human, at least for now. And I don’t think that they’re going to be without people for a long time.

Then there is the other side. ‘Please write me an employment verification letter so I can get a mortgage. Please give a raise to somebody. Promote somebody from one department to another. ‘

Yes, you could say this could be self-service. But that is still a person, a manager or somebody else, doing all that.

Why can’t we automate all that? We took in our case, Slack—which is a Salesforce product—and we just injected a model into it. … Think about the old days. If an employee was trying to get a mortgage, they probably went to the manager and said, ‘I need a verification letter with my salary and number of years.’ … They went to the HR partner, who immediately went to look up the system because they didn’t know the person.

They looked up two different systems. … Then they went to the back office. … Think of the number of steps. … It probably took two hours of human time. … OK, if you’re on Slack … you’ve actually got an ID so Slack knows who you are.

That means you can look up if everything you’re asking for is correct or not, five seconds. You can print the letter and just ask, ‘Who do you want to email to or mail to physically?’ And you could do it 24x7 as opposed to waiting for four days.

So this is an example of leveraging what can be done.

Now go forward. How about employee retention? It’s actually easy, unlike what people think. Thirty [percent] or 40 percent of the time, there is enough data around that we can guess if somebody wants to leave and they’re a great performer.

How about salary raises? We need to give somebody a raise. … All of this, you can automate. That’s the few 100 roles that we could make more productive and go away. … We get, I think, around 7.5 million [support] calls.

People call it and say, ‘The system is down. Can I get some help?’ Or, ‘I need to find something.’ … All the initial steps around gathering information, around entitlements … if I remember right, 75 percent can be addressed right away using AI.

That means that you’re actually freeing people up to do the more complicated ones. That means that the complex ones are getting solved faster because people don’t have to spend time on the simple ones.

So as you think about example after example after example on the enterprise side, you can begin to make things much, much better for your clients. … [Red Hat subsidiary] Ansible is a technology that is used to deploy IT. …

Put together a web server with a firewall with a load balancer with an application server. … The fact that there is a runbook to go do it is great because that’s better than doing spreadsheets and yellow stickies.

Now the runbook itself, if you can go ahead and say, ‘Write me the script to attach a Linux server with a web server and a … load balancer.’ Say it in English. … You’ve just taken two hours of work and made it in about 15 minutes.

You still want to look at it to make sure that it’s about right. But to make two hours into 15 minutes … you can imagine this could save a lot of time inside our CIO team. … As we go forward on this, I can see this come up with C++, in Java … and all the programming aspects.

Solution Provider AI Opportunity

Opportunity is in the hundreds of billions [of dollars]. … In 2007, the smartphone came out, how many of you were making money writing apps for mobile phones? Zero. … By 2010, that became a world of hundreds of billions.

Every enterprise went to—you remember this term—‘mobile first.’ Meaning internally we now spend an equal amount of time writing apps that are mobile-enabled as opposed to desktop- or just web-enabled. … Now I’m telling you that people will use these tools to do AI-driven programming and productivity. By productivity, I mean the enterprise side of the apps.

Programming meaning that what you’re doing for application development, that’s the task. What should you do? … Take a look at Go take a look at what we’re doing around code. Go take a look at what we’re doing around customer service. And get people trained on it. … What does it take to get trained on it? … We went to our people and we said, ‘We’d like you all to go take a week and go learn how to use our tools.’ … People in sales jobs: How can I get a better response to an RFP?

People inside accounts payable: How can I begin to get much better about collecting all the data I need before handing the task off to somebody else?

So your domain knowledge—that is much, much more important. How can I automate the tasks? How can I make it more productive? It’s much more important than understanding AI at a deep level.

And then being able to apply all this knowledge? That is the magic. [The number of people who have to worry about how Large Language Models work], in the end, that’s going to be a few 100 people. … One model, every one of you can do on one GPU.

And that is where I’ll tell you … take Come to our client engineering team and say, ‘Put the onus on us. Help us do this class.’

Once you get it going, we will be off to the races. And I think it is a couple of weeks of work. It’s not a month of work to be able to use these technologies usefully. And I really do fundamentally believe that. And I think that’s a big opportunity for all of you.

Customer service, code and digital workers. Just think of those three.

Customer service in the end is a good 20 percent of the total opportunity. Code is going to be the first one before we get to digital workers.

And then automating—how do I take some fraction of the people doing the repetitive work and make them into digital workers? … Talk about the world of automation and RPA [robotic process automation]. I think this is actually the real technology which makes all those things come to life. … I’ll use a baseball analogy.

Cloud is probably in its fifth inning. So the game has been around. We still have a while to go. … AI is in the first inning. That means the whole game is ahead of us and there’s a huge amount of money to be made. … Quantum, probably getting into the minor leagues. But watch it because in another two or three years, it is going to be one of those that’s what AI is today.

It will look sudden. But it’s not sudden. AI was not sudden. … It was … years in the making.

## IBM Partner Strategy

If you want to partner with our consulting organization, great. But I’m actually going to put that a little bit on the side.

Our consulting at the end of the day is 5 percent, maybe. I doubt that it is even 5 percent of the total opportunity. … By the way, we partner with EY. We partner with Deloitte. We partner with NTT. We partner with Wipro—many of whom are announcing centers [of excellence] around AI with us. So I’d like to go broader.

But for all of you, in the last 12 months we’ve added 4,000 partners. We have upped our incentives. We want to offer you … client engineering and technical training that we want to work with all of you on. … If you want to work with our consulting people, great. But I actually would put that a little bit on the side almost to say … there’ll be a 5 [percent], 10 percent overlap, worst case.

In the rest of the world, we’d rather have all of you do it. The way that we are approaching it is very simple. There’s about 300 to 400 clients where we tend to approach them directly.

Even there, by the way, on the hardware side is a really good channel revenue, but that’s about the max that we approach directly.

There are about another 1,000 [clients] where we will say, ‘Yes, there is some directness from us.’ But it is also partners.

Everywhere else, which means about 80 percent of the market, is going to be channel only, meaning we will not approach them directly. Our commitment is we are not going to change that.

We want to work with the channel. I’ve stood here the previous few years and said, ‘Look, we are about 30 percent of our total revenue from clients coming from the channel.’ … I’d like to make that 50 percent of the revenue, not 30 percent of the revenue.

And we will go down that path. And we’re going to continue going down this path. And that is a hard commitment. We don’t want to make that the other way.

The 2024 Economy

So I personally think that ’24 will be a better year than ’23. … So I’m one of those who kept saying—and people look at me like I had my head twisted on wrong—‘I don’t really see a recession coming. I don’t really see a hard landing.’

I said that in ’20. I said that in ’21. I said that in ’22. The reason is that I see … that technology is the only counteracting force to demographics, interest rates, supply chain shocks, cyber, geopolitics. To all these things, technology is an answer, not a problem that is exacerbating any of those. … They talk about 2.4 percent real GDP growth for 2024.

I think technology will be three to four points ahead of that. I think you now put that together with 4 percent inflation, because that’s kind of a current number … 4 [percent] or 5 percent in real numbers is what I think growth should be. And so 4 [percent] to 5 [percent] is kind of what I think is kind of real going into 2024.

Now, given all the added geopolitics because there is the U.S.-China relationship to technology. Then you can add in what’s happening in Ukraine and Russia. … Hopefully, in the Middle East, [the conflict] will resolve itself.

Could all that cause maybe half a point of hiccup? Maybe. But if it doesn’t get any worse, it’s in that range. So I’m pretty optimistic going forward right now.

And I think there’s a big opportunity for all of us to get all these technologies into the hands of our clients. … I mean, look at the pandemic. …. Who would have thought that your corner restaurant [would] become an omni-channel retailer with takeout and delivery and back to in-store dining? Every retailer became like that, not just restaurants.

Those are all technology opportunities because then you begin to think about how do I worry about serving my client.

Red Hat

I would urge every one of you to be part of both [IBM and Red Hat channel partner] programs.

The reason that we keep the Red Hat program separate—Red Hat works a lot with every hardware and software vendor in the world. They work with Dell. They work with HP. They work with AWS. They work with [Microsoft] Azure.

They work with Lenovo. … And that’s the nature of Red Hat. … And so we got feedback pretty strongly at the beginning. And still do. We want to make sure that Red Hat does not look like it’s [giving an advantage to] IBM over any other partner. … It’s as simple as that.

But … we urge every one of you to be a partner in both programs so you can leverage both going forward.

Is there a lot of Red Hat inside IBM? Of course there is. … But I would urge you to be in both programs because there’s a lot of money to be made.

Culture In A Large Workplace

People know what is the right thing to do. But they tend not to deliver because of fear it’ll cause displeasure in somebody who’s in a position of more responsibility or more influence. Perhaps the boss. Perhaps not the boss but [someone in] the boss’ sphere.

And so people don’t do the right thing. They know what’s the right thing for a client. Or for to grow the business. Or for an employee. Or for the outside world.

So I’ve always taken the view—step back. And you know what’s the right thing to do. Don’t make it about yourself.

So not the right thing for you, but for the greater organization. If you do that thing, I fundamentally believe that on average you’re going to be rewarded. On average, you’re going to get the organization to grow.

And as the organization grows, then you are going to be able to thrive with it.

And so what that means, you have got to have transparency. You’ve got to be willing to take sometimes brutal feedback. … People often talk about small-company culture versus large-company [culture].

So any organization that is large, I don’t think of it as a negative. Be it government, be it nations, be it companies. Large becomes sort of quote unquote bureaucratic because you do need some process. … But bureaucratic does not mean you can’t be nimble. And you’ve got to be very, very aware of what brings real value. … If you can focus on doing the right thing, then you’re not going to become bureaucratic and slow. But you’re going to be large and be able to leverage that scale to succeed even better than you could before.