Human behavior is full of idiosyncrasies, whether it is a CEO, a Manager or an executive, people are prone to blunders that can affect livelihoods.
A bad call may cause layoffs, and an error may mean the loss of a colleague’s job.
The hubris of the human mind is that these things can be avoided with systems. If spreadsheets need to be tracked across globes, over a team of <500k employees, then an ERP would solve that problem.
Unfortunately, ERPs lose millions of dollars annually. It’s a meme at this point. And here comes AI, the great saviour, that may help organizations break free from this ERP curse by unifying all siloes together.
Here we find hubris in effect again- to fixate on one solution over the rest. Remember, the ERPs are not the problem here, per se. The errors are caused by behavior within the organization- through rigidity and a lack of focus on what matters in the long run.
But will AI systems save organizations from human error, or will they exacerbate it?
Yes modernization has taken place within the ERP ecosystem- it is all AI-powered. But without a guiding focus, this project will fail, too. Here’s what you can do instead.
ERP Modernization and AI: Technical Challenges Facing Enterprises
Legacy ERP Systems and Technical Debt
ERP implementation is haunted by vast disconnection- what started out as an endeavor to end data silos has instead caused a huge legacy software problem- enterprises running old software that is too clunky to update tacked over layers of new software.
It is quite the conundrum
ERPs are disconnected from their original purpose and have now become the job that has to be managed and updated so that organizations don’t run into errors- tack on mergers and acquisitions or redundancies that haven’t been updated in 10 years and enterprises find themselves dealing with cascading issues.
This affects modernization in two ways: –
– AI does not have context to manage data effectively
– It might act on erroneous data, causing cascading failures.
Will AI Make ERP Systems Redundant?
Many enterprises, let’s wager, would gladly replace the ERP estates associated with processing delays, inventory discrepancies, customer-service failures and a growing burden of manual reconciliation.
And AI seems like a guiding light- why can’t they now migrate to a platform that is AI-native? Many enterprises are running experiments that would do away with clunky legacy software that needs additional job roles.
ERP vendors have sensed this change and have long started to actively tackle redundancy. But the market view (uncited) is that ERPs are changing too late- the wave has passed and a new tech must take place.
However, even with this view- vendors need not worry, for enterprises are slow to change and they need ERP solutions. They will run pilot programs that read and retrieve data for now- and eventually move to a full-scale implementation.
This does not mean ERPs won’t become redundant- especially if an organization can connect all disparate tools into one layer.
ERP Data Silos and AI Data Quality
Let’s run a thought experiment- answer to yourself this question: what is data?
Then: –
– Do your subordinates also have the same definition? Which is usable data and which isn’t?
– Does everyone in your leadership agree with what is usable and what is not?
The answer to the questions will shed light on a simple problem: no one knows which data is worth something and which is not.
This is the data silo that affects organization- not physical siloes but rather informational gaps in what matters and what doesn’t. If your product team believes data A is invalid but your marketing team doesn’t- how will you decide to keep it for future use?
Data siloes are not just physical containers keeping the data disconnected but people who disagree on what matters and what doesn’t in the longer run.
Why ERP Modernization Projects Stall
Deloitte reports that investment in ERPs is up by 10%. However, there is a caveat- they want lightweight solutions.
But Deloitte overlooks something in the report: legacy software is not so easily overcome. Pilot projects can and will take flight, but what about long-term implementation?
Herein lies the opportunity and a major failure mode for ERP vendors and enterprises alike: dependency on each other. Enterprises need lightweight ERPs but are stuck with legacy software and data siloes and ERP vendors, on the other hand, must make their ERPs more agile, lightweight, and agentic without losing their original essence and power- on an unreasonable time frame.
This is stalemate.
How AI Is Changing ERP Modernization
We may have painted a grim picture. But let us move away from abstractions and take a grounded approach to this problem.
Let’s look at an organization selling a piece of the puzzle- it is a pain point for many enterprise-level organizations. So, for vendors: what happens when such upstarts encroach on your moats? Or, when they are added as another layer on top of another layer.
The mess here seems palpable- how many data points is the ERP going to convert and keep track of?
But thankfully, now organizations have a layer that can be called upon and speaks to other tools via API calls.
AI.
This intelligent layer can connect and call upon different tools, presenting a single source of truth- the question here is simple: can you trust that truth?
Think of all the data you have- think of the data that is considered useful but isn’t. And data that is considered useless but isn’t. How will a leader or operator decide what to act on?
This is where modernization of ERPs really take shape- with humans- operators working everyday to make sure your organization runs smoothly.
Let’s loop back to the start and reintroduce the idea: ERP when stripped of jargon is the process of unifying management processes- hiring, sales, etc. But in the search for unified systems- leaders have become dependent on them, trusting numbers on a sheet more than their process.
Here lies a grave error. One that can be fixed with a perspective shift.
Why ERP Modernization Requires Organizational Change
There are two major bottlnecks to AI implementation: –
– Time and ROI.
– Lack of management structure.
Back in 2020, the current CEO of IBM saw that the organization was dependent on stock buy-backs, bloated vendor relationships, and misaligned incentives. When he took seat, IBM, from an outsiders perspective was a failing organization- massive layoffs, and low innovation had become synonymous with the company.
But CEO Arvind Krishna restructured the entire IBM organization behind high-quality partnerships, distribution as moat, and using its own competiton to sell its product- how? By ensuring that the seller ecosystem gave the seller $3-$4 for every $1 sold for IBM.
It is nothing short of revolutionary.
The point here is simple: organizations, when introducing change into their systems must change with the tool and environment.
Misaligned incentives, legacy thinking and a lack of clear direction will haunt the change and no amount of AI implementation will solve that.
How AI Can Automate Institutional Bias in ERP Systems
The issue here is clear: organizations want custom ERPs that work for their use-cases and not break during updates or new roll-outs.
But what happens when the data stored in your systems is based on bias or practices that harm your organization in the long-run? The AI will provide these data points as facts, and produce these half-truth confidently.
This causes breakage of trust in the long-run- because your teams may be making decisions based on faulty data- now backed by an intelligent entity- when moving fast and when moving nuanced data- this becomes a major problem.
Who Is Accountable for AI-Driven ERP Decisions?
Now, ERPs help enterprises make decisions, when a leader makes a bad approval based on “right” organizational practices, confirmed through AI- who is to blame?
The leader might pin it on the bad data- but who will look at the organizational philosophy that got them there? Perhaps, this might be mentioned in a report after people have lost their jobs or after a major pivot.
The Future of ERP in the Age of AI
Vendors and users alike have become tired of bloated software- disconnection and silos are here for most enterprises.
Yes, AI will help but if and only if organizational behavior changes. Modernization isn’t just about tools, leaders of the market.
Modernization means changing with the tool and adopting practices that minimize damage and revenue loss- not increase it. This framework is meant for you to think beyond mere systems and look at what your team is doing and pulling information from.
The question here is: is the direction the one you want to go in or you’re forced to because of legacy practices.
