Welcome to your IBM i update for April 2026, your monthly digest into what’s happening within the IBM i community.
In this episode of the IBM i update, we discuss AI on IBM i and ask if we’re heading for a productivity breakthrough or is this just a bubble that’s about to burst?
You can watch the video above or read the blog post below…
Opportunity vs risk of AI on the IBM i
AI has become impossible to ignore in the IBM i world.
Gartner, the huge multi-national technology analyst firm is rather cautious on AI.
If you search ‘bubble’ on their website you get a slew of results that paint a tone of tempered optimism, where the opportunity is acknowledged, but the risk of inflated expectations and short-term disillusionment is never far behind.
As reported by The Register, Gartner’s view is sobering. Estimates suggest that up to 70% of AI projects could fail, while as many as 75% of vendors in this space may disappear.
Scary right? But let’s put some context behind these figures.
The Gartner report was mainly geared towards mainframes, a place where scale and interdependency of these systems make full migration physically and financially unrealistic for most enterprises.
So, the failure rate is largely about ambitious mainframe transformation programmes, not practical IBM i augmentation projects. I would suggest that the 70% failure reflects expectation – not technology failure.
Indeed, Gartner themselves advise caution due to what they describe as “The gap between the ‘marketing promise’ of generative AI…” As such, set your expectations accordingly.
As for the 75% of vendors disappearing, well, every emerging tech scenario goes through this.
The AI bubble vs the .com bubble
If you’re as old as I, then you’ll remember the .COM bubble where all anyone needed to throw money at a project was an idea – valuations of companies exploded with no proven business model.
Thousands of startups launched and investors piled in on potential, not delivery. Most of these startups disappeared within a few years.
Today, we’re seeing the same with products being rebranded as ‘AI-Powered’ but I think we’re learned enough (and with hindsight) understand that expectations shouldn’t run before reality.
So yes, many startups and business born on the AI revolution will appear and disappear just as quickly, and poor projects will fail – thousands of them.
But also, today the .COM the market is now larger than ever anticipated with Google and Amazon leading the charge.
AI is different from the .com era.
The fundamentals are stronger, the technology real and adoption being driven by tangible business value.
The lesson from the dot-com era isn’t that the technology failed – it’s that the hype did.
And we’re seeing the same pattern emerge with AI.
For IBM i organisations, the opportunity isn’t in chasing the bubble, but in applying AI where it genuinely adds value – on top of systems that already work.
How you can achieve good AI outcomes
What should we apply to ensure that we’re not just another nearly-run?
Well, good AI outcomes come from strong foundations:
- Clear governance
- Clean architecture
- Well documented
- Trusted data
- Well-defined processes
- Good people
So, this isn’t a signal to avoid AI, it’s a warning to approach it with clarity, structure and good dollop of realism.
AI-assisted tools can significantly reduce the time, cost, and risk of modernisation, but they need to be applied with a clear understanding of where they add value.
But perhaps the most critical component of this your data.
Indeed, IBM are saying the same thing.
Within their, ’Your AI is only as good as your data’ white paper, IBM explain that “In the AI era, no matter what industry you’re in, you’re in the data business.”
So, AI is only as good as the data that goes into it.
And this alone makes your organisation’s data its most important asset.
You can download a copy of the white paper here.
According to IBM, high-quality data isn’t just about volume, it’s data that is
- Accurate and correctly reflects the business, its processes and outcomes too
- Complete, with all data being present with no missing pieces
- Consistant and uniform across the systems and matching formats [dates, currency, decimal place…)
- Timely, up to date and available for when it’s needed (even archived data)
- Fit for purpose, so relevant for the task in hand be that for process, reference or even audit
IBM are clear – poor data quality is one of the most common reasons that AI initiatives fail.
Put it simply, AI models learn from whatever you give them, garbage in, garbage out.
If the data is flawed, biased or incomplete, the output will be flawed – at scale.
So, where success previously meant good code, in the AI era, success means good data.
AI-based code modernisation on the IBM i
When we have that good data in place, what can we do with that foundation layer?
Well, AI-based code modernisation is moving from concept to commercial reality on IBM i.
As highlighted recently in IT Jungle, new tools are emerging to help organisations understand legacy applications faster, uncover hidden business logic, assist with refactoring, and accelerate productivity.
We’re seeing the likes of Profound Logic, ARCAD and Remain Software providing teams with tools that analyse, convert and test complex estates.
Fresche too provide such tools but in addition, their XA-AI extracts the business logic and leverages this scematical information to support the generation of meaningful – connected code.
And of course, there’s Project Bob that’s pushing AI-assisted development and code understanding further into the mainstream.
But be aware, this isn’t push-button modernisation.
AI doesn’t replace skilled developers; it amplifies them. Human judgement is still critical for preserving business logic, validating outputs, and governing change.
And that’s the big opportunity: not AI replacing developers, but experienced IBM i teams using AI to modernise faster, reduce risk, and do more with scarce skills.
Amid the AI hype, the platform itself keeps evolving… albeit it slower in other areas as IBM i Technology Refresh appears to be arriving later than usual, reportedly to align with a broader Power announcement.
Rather than cause for concern, it may point to a bigger coordinated platform story involving IBM i 7.6 TR2, and potentially new Power hardware.
So, what may be coming downstream?
Well, this is all speculation but we may see better support around AI-assisted development tooling while improvements for project Bob have already been mooted.
Sticking with the AI theme, enhancements to SQL services for metadata access (of which will enhance AI-assisted analysis) plus features aimed at making IBM i data easier to expose to AI workloads may become present too.
Security is always a priority and I’m sure we’ll see more around zero-trust too. It would be a joy so see improvements in performance maybe enhancing DB2 (although this is more akin to tuning a Ferrari).
What are your thoughts, you do you think will be announced – let me know in the comments below.
PTF Guide – Updates for Java Security within the Java SDK / Runtime
And finally, this month’s PTF Guide is really a security story, centred on significant Java vulnerabilities affecting IBM i.
IBM has issued fixes across 7.4 through 7.6, with updates to Java, defective PTF and QMGTOOLS groups.
The takeaway is simple: patch discipline remains a key part of IBM i risk management.
Another reminder that on IBM i, stability doesn’t happen by accident… It’s maintained.
And that’s it for this months update.
I’m Andrew Nicholson, and we’re Proximity, your application support, maintenance and development partners that are in your corner.
If you missed it, catch up on our March IBM i update.
Follow this link to access all the IBM i Update’s in one place.
Alternatively, if you’d like to receive the latest IBM i Update hot off the press to your inbox, subscribe to the newsletter version on LinkedIn.
