That new system won’t fix your current process problems

Illustration of a rabbit coming out of a hat, referencing a magic trick.

AI is obviously ever present in any conversation about technology and transformation, but the more practical challenge I keep coming across is much less futuristic.

Organisations know their current ways of working and the systems supporting them are no longer serving them properly. The mistake I see too often is jumping too quickly from recognising that problem to shopping for the replacement, before properly designing and testing how they want to work instead.

The issue isn’t that the current system isn’t part of the problem. Often it is. The issue is starting with the system, or treating it as the total solution to the problem.”

When I’m speaking with software vendors, implementation partners and people who work around large systems projects, I often run a scenario past them.

It usually goes something like this:

  1. The project scope starts with Systems or IT.

  2. The business agrees on the broad problem.

  3. Requirements are pulled together and an RFP goes out.

  4. A platform is selected and implementation begins.

  5. Somewhere through discovery, everyone realises the future processes the system is supposed to support haven’t actually been clearly defined.

  6. Those processes now need to be designed while the implementation is already underway.

  7. Change management starts catching up behind it.

  8. The organisation ends up expecting the new system to deliver the new way of working at cutover.

The interesting bit is what happens when I get to around point four. There is usually a very familiar expression across the table, that unintentional recognition of “I know exactly where this is going.” By point five, they are often finishing the sequence for me.

This isn’t a criticism of software vendors or Systems teams. The good ones know exactly what the problem is. Their challenge is that they are often discovering it after the organisation has already selected the solution.

The problem is sequencing

There are usually legitimate problems with the existing system. It may be old, patched together over years, poorly integrated or painful to report from. People may be running shadow processes through spreadsheets and other tools because the core platform no longer meets the needs of the business.

None of that means replacing it is the wrong investment. The problem comes when the organisation jumps from knowing the current way of working isn’t good enough to deciding it needs a new system.

There is an important piece of work missing in between.

Before deciding what technology you need, you need to properly understand how the business works now, what is genuinely broken, how you want it to work in the future, and where the limitations of the current system actually sit.

Otherwise you are effectively buying the solution before fully understanding the problem.

This becomes particularly important with large platforms. Salesforce, Microsoft Dynamics, HubSpot and others are incredibly capable systems, but they are not interchangeable and they do not all solve the same problems in the same way.

If the requirements are based on a quick review of the current process plus a list of things people wish the existing system could do, it is quite possible to choose a platform that technically meets the RFP and still discover later that it does not naturally support the way the business actually wants to operate. That is when the customisations, workarounds and additional costs begin.

A new system should enable the way you want to work

Technology should enable the desired way of working, not dictate it. That sounds obvious, but it requires more work up front than many organisations allow for.

This isn’t a new or controversial problem. Most major consulting firms have been arguing for years that technology transformation needs to start with the business and its future operating model.

Yet the failure pattern continues to appear. Metcash’s Project Horizon grew from an $80 million transformation into a program approaching $300 million, with its scope and timeline reset along the way. Westpac’s In the Moment program cost more than $70 million before being written off when the custom-built solution failed to deliver the capability the business needed.

The theory is well understood. The difficult part is doing the work properly before the technology decision gets ahead of the business thinking.

Fix the playbook before you automate it

The reason for moving to a new system should be to make the desired way of working easier, faster, more accurate and more scalable. It should not be the point at which you start rewriting the playbook.

That means properly documenting how the business operates now, ideally through process maps and real user journeys. Where are the hand-offs? Where do things get stuck? Which approvals genuinely add value? What information is required at each point? Where are people working around the current system because the process demands something it cannot support?

From there you can design the future process. What should change? Why does it need to change? How does that improve the experience for customers, the team or the business? What does the future process need to achieve that the current one cannot?

Only then can you properly separate a process problem from a technology limitation.

This is also where I think Systems and functional leaders are often put in an unfair position. They do their best to work through these questions, but redesigning an end-to-end operating process may not be their core capability. They are often doing it alongside their actual jobs, from within one part of the organisation, with all of the internal history, competing priorities and politics that come with changing how people work.

The answer isn’t handing that problem to the software vendor either. There is a big difference between asking a vendor to configure a platform around a future way of working you understand, and asking them to discover what that future way of working should be after you have already selected their product.

One is a technology implementation. The other quietly turns a technology implementation into a business transformation project halfway through.

Don’t let the demo write the requirements

Even organisations that start with a sensible set of requirements can get pulled off course once they start looking at platforms.

SaaS vendors understandably show their product at its best. You see workflows, automation, dashboards and features working beautifully for another organisation, often in a completely different sector or with a different use case, and somebody says, “We’d like that.”

The requirements start growing around what the technology can do rather than what the business has decided it needs to do.

The problem usually becomes obvious later, when the feature is configured against the actual operation. Your use case is different. The workflow doesn’t quite fit. The feature isn’t as flexible as everybody assumed, or making it work requires additional configuration and customisation. At that point, the process starts being redesigned around a feature that was never part of the future way of working in the first place.

There is nothing wrong with seeing something in a platform that gives you a better idea. Good technology should challenge your thinking and show you possibilities you hadn’t considered. But there is a big difference between a feature that genuinely improves the future operating model and one that makes its way into the requirements because it looked impressive in a demonstration.

If you have already designed how you want to work, that distinction becomes much easier to make.

Prototype the future before you buy it

One of the most useful things an organisation can do before replacing a major system is prototype as much of the future way of working as possible before committing to the platform.

That doesn’t mean pretending the old system is good enough or delaying an investment that clearly needs to happen. It means testing the process before paying someone to automate it.

Take a business development process as an example. The current system might provide poor visibility across opportunities, activity, deal stages, approvals and conversion. That sounds like a CRM problem, and eventually it probably is.

But before selecting the replacement, you can define the actual deal stages, agree what needs to happen before an opportunity moves forward, test approval gates, track how long opportunities sit at each stage and work out what information leadership genuinely needs.

Until recently, doing this often meant shared spreadsheets, forms and other fairly crude temporary tools. They still have a role, but this is one area where the new generation of AI tools has genuinely changed what is possible.

Tools such as Claude Code and ChatGPT Codex now make it possible to take a well-defined business case and set of functional requirements and turn them into a workable prototype far more quickly than would previously have been practical. Anthropic is already using Claude Code internally for rapid interactive prototyping, including turning static designs into functional prototypes, while OpenAI reports non-technical teams using Codex to build internal apps and dashboards.

The point is not to build your own CRM or create another piece of shadow technology that becomes permanent. It is to create enough of the future process to actually use it.

Put it in front of the people who do the work and let them run real examples through it. You start to find out which fields really matter, where the hand-offs fail, which approval rules make sense and which ideas sounded good in a workshop but become painful in practice.

The temporary prototype is not the solution. It is a way of learning what the solution actually needs to enable.

By running the process before automating it, assumptions get tested and problems become visible. Some things that sounded essential turn out not to matter much in practice, while other requirements only become obvious once people start working in the new way.

By the time you go to market, the buying conversation changes considerably. Rather than asking a vendor what their platform can do for you, you can show them how you need to operate and ask how well their platform supports it.

You can also benchmark the platforms against functionality you have already tested and ask vendors to price against realistic scenarios for how the organisation is likely to grow over the next three to five years. You are not just asking whether the platform works for the business tomorrow, but whether it continues to make operational and commercial sense for the business you expect to become.

That is a much stronger place to select technology from.

Change management should start before the platform does

Prototyping the future way of working also changes the nature of change management.

Too often, change management becomes something that appears towards the back end of a systems project. The platform is configured, implementation is approaching, and attention turns to communications, training and getting people ready for cutover.

By then, a lot of the important change has already been decided.

If the people doing the work are involved earlier in understanding the current process, identifying the friction and testing a better way of operating, change starts well before the technology arrives. They have had an opportunity to shape the process, see why things need to change and understand what the future model is trying to achieve.

That doesn’t remove the need for proper change management during implementation. It makes that work considerably stronger.

Instead of introducing a new platform and a new way of working at the same time, the organisation is introducing technology that supports a way of working people have already started to understand and, ideally, test.

The conversation shifts from, “Here is the new system and this is how you now need to work,” to, “This is the way we have been working towards, and this platform allows us to do it properly.”

That is a much easier change for an organisation to absorb. It also provides a useful test before committing significant money. If the proposed process still doesn’t work when you prototype it with simple tools, buying a more sophisticated platform is unlikely to solve the underlying problem.

The real process should look more like this

The traditional sequence often looks something like:

Problem → Requirements → RFP → Platform → Discovery → New processes → Implementation

I think it should look more like:

Understand the current operation → Design the future operation → Prototype and test it → Identify the real technology gaps → Define requirements → Select the right platform → Implement and scale

There will obviously be overlap. You are not trying to perfect every process before speaking to technology providers, and good vendors can bring enormous insight into what is possible and where standard platform functionality should influence the design.

But the organisation should enter that conversation knowing far more than what it doesn’t like about its current system. It should know how it wants to work, what it has already tested, what really matters and where it is prepared to adapt.

I’ve seen this pattern for decades

I first recognised a version of this problem decades ago when I was working on digital platforms for large institutions, including Royal Bank of Scotland’s Global Trade Services platform.

The technology was obviously very different, but the underlying issue was the same. Infrastructure decisions could be made before the customer journeys and ways of working the platform needed to support had been properly understood. Once the technology was locked in, teams were left trying to deliver the right experience within limitations that had already been designed into the solution.

The technology has changed enormously since then, but we can still make exactly the same sequencing mistake. We select the tool and then work out how the business needs to use it, when the order should be the other way around.

Before you issue the RFP

If your current systems are genuinely holding the business back, replacing them may absolutely be the right decision. But there is an important difference between knowing what the current system cannot do and knowing what the next system actually needs to do.

Before issuing the RFP, I would want to be confident that the organisation properly understands how it operates today, where the real process problems sit, what is genuinely constrained by technology and what the future way of working should look like.

I would also want to know how much of that future model can be tested before the new system arrives.

  • Have the people doing the work actually tried it?

  • Have the assumptions held up?

  • Have the important information, hand-offs, approval points and user journeys been worked through properly?

  • And have you worked out what should be standardised and what should deliberately remain different because it creates value for the business or its customers?

If not, you may be asking vendors to help define the solution after you have already started shopping for it.

Doing this work first can feel like slowing the transformation down, particularly when there is pressure to get on with replacing a system everybody already knows is inadequate. In reality, it gives the organisation a much better chance of choosing the right platform, reducing unnecessary customisation, making implementation easier and getting to value faster.

The new system should make the better way of working easier, faster and more accurate. It shouldn’t be where you first discover what the better way of working is.

This is a part of the strategy and transformation work I genuinely enjoy working through. Getting clear on how the business needs to operate, prototyping what can change before the technology arrives, and turning that into real requirements before the platform decision gets locked in.

If you’re staring down a systems transformation and want a hand working through that piece first, reach out. I’m always happy to help.

Cheers, Craig.

Written by Craig Bridge, Founder and Strategy Consultant at CB Strategic Consulting , Melbourne Australia.

Previous
Previous

Building a higher-performing team could be your biggest strategic unlock for the year ahead

Next
Next

SMEs have been using AI throughout 2025, but how and what for may surprise you