With AI being the topic of conversation for procurement professionals right now, it’s easy to get lost in the maze of conflicting information. Vroozi is a procure-to-pay platform powered by robust AI capabilities to deliver meaningful use cases. CEO and Co-Founder, Shaz Khan, takes approaching AI the right way very seriously.
For Vroozi, the use of AI is a two-sided coin. It’s an organisation that talks about AI both in production and consumption. AI is a tool that has been a game-changer, because it has enabled Vroozi’s software and technology engineers to be able to rapidly prototype and develop code. And that code is beneficial for creating feature sets and capabilities that the company wants to introduce to the market.
“Similarly, we take steps to look at how a customer interacts with our software for the first time,” Khan explains. “The implementation process is also ripe for consuming and producing great results with AI. Imagine you go through some type of interview wizard where you prompt the system based on your region and industry. The system will self-configure according to your business unit. This is real intelligence that understands your business at a different level, as well as the competitive landscape, and brings in best practices to deliver incredible results.”
Getting the approach right
Having said that, Khan freely admits that we’re in the early innings of AI adoption. For him, leaders should adopt a multi-pronged approach to implement AI. The first move is to assemble a team. “One key area with AI is that a lot of companies are relying on outside experts that don’t know the business and the goals that they’re trying to achieve,” he explains.
“You should invest in your own people before you invite outside parties in. Bring that education and assemble a use case, before assessing the problems you’re trying to solve and determining whether AI is a good tool set or capability to solve the problem. If these things match up, execute the game plan, bring in the right technologies and the right expertise, and only then bring AI capabilities into your workforce.”
The challenges
With this being the “early innings”, there are also barriers and challenges. The main issue, from Khan’s perspective, is security. “There’s a trust aspect that has to be looked at,” he explains. “There’s also an ethics aspect. Are you delivering the right results? And how much autonomy are you giving AI and its agents to go out and deliver those results for you without any human interaction? I think the companies that get it right will strike a balance between the trifecta of automation, really great AI technologies, and a balance of human interaction to create an overall output.”
There’s also the question of data. If the data isn’t clean, output will be compromised and lead to poor results. We haven’t seen the worst of what can happen, Khan believes, and AI has the potential to create scenarios that are hard to recover from, if used poorly. “We need to prepare ourselves now to prevent those types of potential calamities from happening,” says Khan. Which is the entire point of DPW: for procurement and technology leaders to educate and learn about best AI practice.
This allows people to cut through the, as Khan puts it, “hysteria” around AI that can cause problems for businesses. They’re rushing to solve problems, and while leveraging AI can be a component of a complete holistic toolkit, it can’t be the only answer. “A lot of companies today still struggle with getting their businesses off spreadsheets,” he states. “AI should be an equaliser and enabler to get it right.”
Structuring unstructured data
For Khan, in order to ready themselves for AI, procurement professionals and practitioners need to be absolutely committed to data management and governance. “What companies often forget is that much of today’s data is unstructured. It’s not neatly stored in databases – it might be a chat, an image of a spec sheet, or a contract never digitised. This unstructured data often can’t be used by AI models today, so companies risk only addressing a small part of the challenge. Data governance has to be an ongoing exercise.”
Having said that, Khan is keen to differentiate between clean data and perfect data. In fact, many procurement professionals we spoke to at DPW New York 2025 said the same. The message is: don’t wait around for everything to be perfect, or you’ll never start.
“Good enough data is just fine,” Khan says. “But if you’re going to continue to feed your AI engines and algorithms bad data, your outputs will be compromised. Companies need to have data governance strategies and upfront policies in place so that they can manage this, independent of the people that offer them.”
AI creating a complete picture
While treading carefully is important, Khan is equally keen to extoll the many virtues of AI for procurement professionals. There are many incredible use cases already, and AI tool sets and algorithms can effectively interrogate a company’s data and give them the answers they require. AI enables these users to have a complete picture of their buying cycle, and allows them to get additional information for where they can pivot.
“This is where the true power of agentic AI will come into play,” says Khan. “When you can fully trust the system inputs, AI will be able to orchestrate those processes autonomously, and present that information to an end user for final decision.”
Khan is very excited about what Vroozi is doing within its own AI layer. The business looks at AI and intelligence as a pervasive thread across its entire tech stack. Every aspect of its platform has some kind of AI enablement, although it’s not an AI-first company.
“We follow three distinct areas where we are thriving on the AI front,” says Khan. “First is intelligent document processing. Can we take structured and unstructured data such as contracts, quotes, work orders, and invoices, and populate them automatically onto a screen without any human touch? Processing invoices might require an army of people typing in data, and they might not capture it all. But an AI toolset can take millions of records and process them simultaneously. That’s the power of AI.”
The power of hyper-personalisation
The second area is what Vroozi calls hyper-personalisation, where it intensely personalises the platform to meet a company’s preferences and needs. It’s about how AI can find trends and not only predict the user’s needs, but also help take the next steps. This includes finding suppliers and ordering things that are needed, so that workflows aren’t disrupted.
“Then we also have what we call the push economy,” says Khan. “AI’s power is in pushing and giving people head starts. So when you talk about AI algorithms and look at analytics, it’s about how AI can present to companies in the procurement space when they need to lock in favourable pricing on products and services, and predict when you are seeing potential fraud scenarios based on trends and patterns. You need a lot of data for those AI models to train on, which is why I say we’re in the early innings. It takes time, but it’s incredibly powerful when you get to that point.”
The benefits ahead
At such an exciting time for procurement, 2025 and 2026 look bright for leaders in this space. Not only procurement, but also supply chain and FinTech, are set to benefit from what AI can do with data.
“There’s going to be a focus on how to capture and harness data, and feed it into AI in a way that produces results,” says Khan. “What we’ll see in the next two years is that AI has now learned from the data that’s been fed into it. You’re going to see higher-quality results and better outcomes. Again, I would caution companies to define the problem first. Then determine if AI is an absolute enabler and game changer. We believe AI can be an influencer and supercharger in terms of productivity. However, there needs to be specific use cases that make sense for corporations.
“In 2025 and beyond, you’re going to see great technologies embedded into organisations that really work.”