5 Reasons You Shouldn’t Fragment Your AI Rollout

If you’re searching for best practices on how to roll out artificial intelligence (AI), a look back at your company’s source-to-pay (S2P) implementation is an excellent place to start. One lesson many companies learn is that a fragmented approach doesn’t deliver the value you expect from the project. You might have automated payments, for example, and achieved some quick wins — but maybe reconciliation was left as a manual process, which still kept your AP teams bogged down in manual work. It’s much the same with AI. As companies dig into real use cases across S2P, we’re seeing the proof that AI is powerful when implemented strategically and can solve complex problems, improve efficiencies, and generate insights. But a siloed implementation squanders this potential.

Today’s post focuses on why a fragmented AI rollout is something to avoid. It’s the first part of our three-post series on how cross-functional collaboration compounds the value AI brings to any company. Stay tuned for the rest of the series!

To learn more about how Coupa defines AI, check out our AI Glossary Guide.

How a fragmented AI rollout happens

There are numerous ways technology implementations can become disjointed. They often happen organically and with the best of intentions.

• Rogue teams: One or more people on a team may be interested in AI technology and will experiment on how it can serve their particular role. This experimentation can be helpful for an individual or a department, but other aspects of the business can be left out of the loop and processes may become disconnected, with some tasks assisted with AI while other tasks remain manual.
• Excluding IT: Without IT’s guidance on implementations or initiatives, competing technologies can create issues including breaking up workflows, increased costs for various tools, and added security risk for unvetted tools and applications.
• Unmapped workflows: Undocumented workflows and processes lead to missed steps that can either be skipped altogether or remain manual, reducing the value AI can provide.
• Change management strategy: Implementing an AI strategy requires organizations to change how they think and work. Failure to adopt a change management strategy slows down the benefits of an implementation because employees resist change and take longer to adopt new technology.

How a fragmented AI rollout limits value

1. Margin erosion

Fragmented AI solutions often lead to siloed data and incomplete analysis. This prevents you from gaining a holistic view of spending patterns and supply chain dynamics, making it challenging to identify inefficiencies and opportunities for cost optimization, budgeting, and reallocation of resources. Without a comprehensive view, organizations miss out on valuable insights that could drive margin improvements.

2. Less agility

Disconnected AI tools may not communicate effectively with each other, creating duplicate work and wasting time and resources. Instead of responding faster to market changes, teams struggle to stay on top of day-to-day operations. This can drive up costs, increase employee burnout, and make it more difficult for companies to attract and retain talent that expects to work with leading technology. AI has the capability to automate even complex workflows and free up teams for higher-order work, but a fragmented approach limits this potential.

3. Fewer confident decisions

Fragmented AI solutions may provide conflicting or incomplete information. Without the right data, leaders are at risk of making decisions that don’t support larger business objectives. Spend analytics based on a unified AI approach help teams make more informed financial and strategic decisions that drive better outcomes and faster time-to-value.

4. Limited access to predictive, prescriptive insights

Unified AI platforms like Coupa’s can leverage community-generated insights across a vast $8 trillion dataset. Fragmented solutions miss out on this collective intelligence, potentially leading to less accurate predictions and recommendations across the S2P process.

5. A costly and cumbersome tech stack

Managing multiple fragmented AI solutions can increase IT complexity. How do you handle integrations and upgrades in specific areas, for example, and where do you find the IT resources to do this — without introducing cybersecurity risks? This is where IT leaders should compare how short-term purchases costs and the total cost of ownership impact overall margins. A unified approach helps reduce this overhead.

A framework to avoid a fragmented AI rollout

Every organization will be at a different stage of AI maturity and has a unique risk tolerance. While some may jump ahead in the journey, the framework below is a solid starting point for those still on the fence about adopting and realizing value from AI-powered procurement solutions.

To unlock the real potential of AI as it evolves, the most effective approach a company can take is one that provides a unified solution — one that creates a positive ripple effect across workflows and efficiencies. Based on more than 3,000 platform implementations with customers around the world, we know how important it is for everyone involved to take the time to understand the workflows and department needs and then research and implement solutions that will streamline processes to see real results.

Continue to the next article in this series, The AI Multiplier Effect: How Collaboration Between Teams Can Amplify Value, to learn more about cross-functional AI and how a collaborative rollout compounds the value company-wide.

* Gartner, Gartner for Data & Analytics Leaders, Jul 2024

** IDC, IDC Executive CIO QuickPoll Series: Operationalizing AI, Sep 2024