Reevaluating Technology Portfolios in the Age of AI

SaaS products are under renewed scrutiny. Rising license costs, underused applications, and overlapping capabilities are forcing leadership teams to take a harder look at their technology portfolios. At the same time, AI tool builders and “vibe coding” platforms like Copilot Studio and Anthropic’s Claude promise new ways to consolidate services, automate work, and reduce operational overhead.

An effective tech stack improves process efficiency, communications, visibility, decision-making, resource management, AI application outcomes, and more. But in many organizations, the push to optimize enterprise technology is long overdue. Poor governance has led to departmental tooling, fragmented data, and bloated software spending.

Recent research findings tell the tale:

• More than 25 percent of global data and analytics employees estimate their organization loses more than $5 million each year due to poor data quality, according to Forrester research, and 7 percent say that figure is $25 million or more.
• A recent study found that the average enterprise manages 897 applications, but less than 30 percent are integrated.
• And about two-thirds of businesses spend more than $2 million annually to maintain or update their legacy systems, according to a 2024 report.

The application landscape often grows organically rather than strategically, as teams are wooed by the latest shiny object. And the perceived difficulty, risk, and cost of modernizing legacy platforms lead teams to layer new tools on top rather than address root causes.

What’s more striking is how many organizations are excitedly embracing “AI” as a solution concept, while still relying on deeply outdated systems to run core operations. AI, however, is not a silver bullet. It is not going to magically replace or fix broken processes overnight.

If you don’t fully understand the business processes and dependencies behind the applications you plan to add, consolidate, or retire, you risk creating more disruption than value.

So what should leaders actually do when approaching a technology rationalization effort? Based on Integrated Project Management (IPM) consultants’ experience guiding efficiency, digital, and transformation initiatives over the decades, here are six recommendations.

1. Start with Business Outcomes, Not Tools

Technology optimization should never begin with a list of licenses, vendors, or software. It should start with clarity on business outcomes. This is the most important part of the process, because all the other work—process improvement, change management, data analytics—derives from it. IPM has found that companies who don’t define strategy and outcomes first rarely succeed with large digital projects.

For example, a growing company with a strategy to expand through acquisition should consider systems that are flexible enough to enable integration. What enterprise platforms must be consistent throughout divisions, and which software can remain local? What will it take to train the new people on the enterprise system and processes? Alternatively, what will it cost if they don’t adopt them? How much organizational change will it require to sunset existing applications supporting business units?

Likewise, an organization reliant on FDA regulatory compliance has different challenges—and thus technology needs—than one manufacturing millions of commoditized components.

When evaluating your tech stack, consider business needs and long-term goals, as well as capabilities, processes, and risk. Which capabilities are truly differentiating for the business? Is technology actively constraining growth, compliance, or customer experience? Are there processes that are fragile, slow, or overly manual?

Cost is certainly a consideration as well. But when rationalization is framed purely as a cost-cutting exercise, teams tend to protect their own tools and resist change. Frame optimization instead as an opportunity to improve resilience, speed, and clarity. It will be far easier to build alignment.

An industrial products company that had grown quickly through acquisitions intended to reduce costs and increase revenue. This would require getting everyone working together quickly to take advantage of buying power and best practices. But the different companies, and often their multiple functional divisions, used different enterprise resource planning (ERP) systems and other point solutions. IPM helped the company define the global standards and processes, then configure and build a solution—along with data governance and decision-making frameworks—to support them. The new ERP and CRM will ultimately reduce costs, but, more importantly, they enable consistency, scalability, visibility, and competitive advantage, meeting the company’s objectives.

2. Map Processes Before You Map Applications

Once you identify the business outcome, don’t jump right into the technology solution without considering processes and people. One of the most common mistakes in rationalization programs is attempting to remove, change, or add applications without understanding how work actually gets done.

IPM worked with a medical device company that planned to implement a new collections management system. Too many of the company’s customers had unresolved billing problems, and it wasn’t clear why. Some were even reaching their credit limit, so they couldn’t order additional products. Executives realized that to accomplish the goal—to make the company easier to do business with—the entire order-to-cash process needed work before the software would be effective.

An IPM consultant assessed the process from end to end and identified pain points, one of which was that sales reps were not requiring a PO when taking orders. They designed a new process and clarified roles and responsibilities. For instance, the correct documentation must be in place before billing. And if there is an issue, the type of problem dictates who should contact the customer and when. The new process steps, required records, and escalation points then informed the software requirements. Now, customer problems are resolved quickly, enhancing service and increasing orders. In addition, because they are tracking what type of disputes they get, they can use the data for ongoing improvement.

Another organization, a pharmaceuticals packaging company, hired IPM to help implement a new digital tool to track time out of refrigeration. The cold-chain tracking software would replace manual processes that took a lot of time and on-the-spot calculation. The cumbersome method caused a lot of deviation and scrap. And it was inconsistent from plant to plant. If the software and hardware rollout occurred without documenting and standardizing processes, the results would have remained inconsistent and difficult to track. The tool could have caused more problems than it aimed to solve. So IPM mapped current processes, designed streamlined future processes, and gained alignment with leads from each site across the globe before implementation.

In any change to the technology portfolio, map end-to-end business processes, not just system architecture, before making any decisions. Identify manual workarounds, downstream dependencies, and reporting needs. What steps can you streamline? Are there processes you can automate with AI applications? Where does data originate and how do you use it?

This level of detail is essential. AI tools can help automate steps, but only if the process design is sound. Also, a low-use tool could turn out to be critical glue holding together brittle processes. Eliminating it without redesigning the underlying workflow will create future failure points. And you may find an underused tool is an untapped enterprise solution.

An industrial products company IPM consulted for has a variety of manufacturing sites. Some produce a broad mix of custom products, others consistent ones. Most managed their new product development on paper and manual spreadsheets. One site used an off-the-shelf product lifecycle management (PLM) system. A PLM holds product data including design and specifications, work instructions, engineering changes, and quality records, and the company wanted to benefit from the faster introductions and consistent quality it would enable. It tried to implement the PLM at another plant but failed due to widely different processes.

The manufacturer hired IPM to assess the end-to-end processes and technology, to help determine if the PLM was an option. IPM clarified the company’s business goals and mapped the ways that people at different plants were working. They found that the existing PLM would work across the other sites, with minor customization and some key system integrations. Importantly, cross-functional representatives across the sites were involved in designing consistent processes for new product introductions and engineering changes and developing the system requirements. Getting input upfront is also part of effective change management.

IPM is now leading the team to prepare data for migration and users for new ways of working in advance of end-to-end testing. Ultimately, more consistent processes will accelerate product introductions and improve consistency and quality across sites.

3. Establish Strong but Flexible Governance

Many organizations are rationalizing today because they lacked clear governance yesterday. And AI experimentation—spinning up agents, automations, and local tools—increases the need for guardrails around verification, data security, and other risks.

A recent study found that the average enterprise manages 897 applications, but less than 30 percent are integrated.

An IPM consultant recently drafted an AI governance proposal for the regulatory affairs team at a client engagement. He had been supporting the global team on a handful of initiatives, all of which had an AI component, and saw inconsistencies in use and outcomes. For example, two dashboards would provide a different result based on the same data—which was correct and why?

The proposal is meant to help coordinate across teams, apply consistent and responsible practices, and mitigate risk. Recommendations included prompt engineering standards, output validation, access controls, tiered training, and ROI scoring.

Effective governance doesn’t mean heavyweight approval boards or piles of paperwork. It should enable innovation, not block it. It means clear ownership of tech portfolios and investments, defined criteria for introducing new tools, data security and access, transparent cost and use reporting, and regular review cycles tied to business priorities.

The industrial products company noted above, which implemented new ERP and CRM software to help integrate business units and sites, used a formal transformation management office (TMO). As the tools gain traction, the TMO will scale down to become the ongoing governance team. They will oversee data and process governance, as well as the company’s technology decision-making framework, providing visibility and accountability, and enabling faster conflict resolution and risk mitigation.

4. Invest in Change Management as Seriously as Technology

Technology implementation and optimization fail far more often due to people and behavior than technical limitations. Adding, retiring, or consolidating applications changes how people work, how decisions are made, and who’s accountable for what. Without strong change management, even the most logical plan will be undermined through shadow IT, workarounds, or outright resistance.

In one of the examples above, in which a pharma packaging company rolled out tablets and software to record time out of refrigeration, gaining alignment with each of the facilities was key. IPM partnered with leadership to ensure site leads and users understood why the tool was important and the risks associated with not changing processes. On-the-floor observations informed the new processes as well as training and feedback methods. Ongoing communication ensures users, site leads, and executives know the status of the rollout and any changes to processes or the application, and it always ties back to the desired outcome—to reduce scrap and risk.

In any technology program, it’s important to clearly articulate why change is happening, not just what is changing. Involve end users in process redesign and tool selection, and provide training that focuses on outcomes. Actively manage the transition period and continue to reinforce new ways of working.

It’s worth noting that AI-enabled tools can accelerate change, but they can also amplify confusion if introduced without context. People need time, support, and clarity to adapt, especially if they are concerned about job security.

 5. Be Honest About Legacy Constraints

Many organizations delay rationalization because legacy systems feel untouchable, especially in highly regulated industries where software may be tied to an approval agency. Companies might use AI to extend their life by improving the interface or automating interactions. However, they may only be increasing the long-term risk and cost of those systems.

Two-thirds of businesses spend more than $2 million annually to maintain or update their legacy systems, according to a 2024 report.

When IPM worked with a veterinary diagnostic company to help them launch a new point-of-care business segment, consultants had to overcome a hurdle. They designed new processes and reconfigured the ERP system, so veterinarians could order and receive the new equipment, materials, and service plans. Billing was the problem; the MS DOS-based accounting system limited the amount and type of data on invoices.

The company had been accommodating the old system for years, so it was deeply rooted in technology across functions. Even though fewer people have MS DOS expertise and other solutions are far more advanced, the system remains. The team had to work around the constraint, configuring the ERP and training users to overcome character maximums and fill in data gaps. AI could help set and follow the rules, but it would only cover up the fundamental problems.

Whether adding to or subtracting from your technology portfolio, it pays to be explicit about which legacy systems are candidates for modernization. AI may be able to prop up outdated technology in the short term, but it’s not a reliable foundation. Again, serve the outcomes you’re trying to achieve, the optimal business processes, and your organization’s capabilities to make decisions.

6. Treat AI as an Enabler, Not the Strategy

Likewise, AI should sit within your optimization strategy, not replace it. Used poorly, AI becomes another layer of complexity, cost, and risk, and it operates in silos, unable to scale and solve real business challenges. Used thoughtfully, AI can reduce manual effort within stable processes, replace niche point solutions, accelerate development and experimentation, and improve access to information across fragmented systems.

For example, IPM led a pharma company’s initiative to better compile and share health authority questions and answers, so they could save effort and be more consistent. The program collected data from the regulatory information management system (RIMS) as well as from correspondence, word documents, and PDFs across the company. The team built a relational database so they could access the data, then used AI to parse the data and pull out agency questions and answers, which could be found using metadata or a keyword. The team can respond to FDA inquiries more quickly and consistently.

Now, as AI has evolved, the team has been able to provide access to the data via a chatbot. It’s even faster and easier to use because searchers don’t have to rely on keywords.

The organizations seeing real benefits are those that apply AI after simplifying their technology landscape, not those trying to use it to compensate for unmanaged complexity.

Foundation for the Future

Optimized technology aligns with how your business operates and its aspirations, whether that means growth, efficiency, customer retention, competitive advantage, or any other goal. It can also prevent the millions of dollars average companies lose every year due to poor data quality, out-of-date legacy systems, and failed IT projects.

And while AI changes the economics and possibilities, it doesn’t remove the need for discipline, foundational hygiene, or deep understanding of processes and people. If anything, it makes those fundamentals, along with strong change management, more important than ever.