AI vs. Spend Analytics: An Honest Comparison

If the difference between Spend Analytics and Artificial Intelligence (AI) in procurement is less than crystal clear in your mind, you are not alone. It’s almost as though whoever created these categories followed a recipe for confusion:

  1. Start with the name of a common activity
  2. Tweak that name just slightly so it means something different
  3. Fold in the label of hot new technology
  4. Season generously with marketing jargon
  5. Bake in the oven of media buzz and high expectations

No wonder it’s foggy. At the simplest level, both AI and spend analytics are technologies that can support procurement and sourcing activities. But there are meaningful differences. Understanding both types of solutions and how they interact can help a team choose the right tools and be more productive. That’s the goal of this article, and we’ll start by looking back before either term was coined.

Spend Analysis, then Spend Analytics

Spend analysis describes a common activity for procurement. It’s a fitting label for analyzing the spending of an organization, with a goal of identifying opportunities to save. Spend analysis typically includes separate steps to gather, cleanse, organize, and analyze information about a company’s expenses.

In decades past procurement analysts waded through physical invoices, filing cabinets, general ledger entries, and vendor reports to conduct spend analysis. The spreadsheet was a huge step forward, and then amid the wave of business intelligence (BI) software emerged the procurement-focused set of tools called spend analytics. Leading spending analytics tools include GEP, SAP-Ariba, Coupa, Jaggaer, Zycus and Spend360, among others.

Spend Analytics: Foundational and Historical

Spend analytics tools have provided and continue to provide a powerful foundational capability for Procurement. From the start, these tools were designed to bring automation and ease to the basic steps of spend analysis:

  • Importing data
  • Facilitating classifications / categorization / taxonomy
  • Providing summary level full-enterprise visibility into spend with standard reporting
  • Allowing drill down by category, product, vendor, or business unit

These solutions rarely stand alone; they are usually connected to a consulting engagement or a broader enterprise system. Those adjacencies reveal two important aspects of spend analytics.

  1. First, most spend analytics solutions emerged as modules or add-ons to accounts payable, procure-to-pay, or other enterprise-wide systems.
  2. Secondly, as powerful as good software can be, its value always depends on the available data. Collecting, formatting, and importing data has historically been the most labor-intensive activity in spend analysis, and is an essential (and often costly) component of spend analytics.

Once the data gathering phase is complete, spend analytics provide reports and dashboards that summarize high-level and category spending. Further, they allow users to delve into the components and causes behind the numbers: vendors, invoices, time periods, and sometimes line items.

Spend Analytics Are Backward-Looking

It’s important to note that spend analytics solutions are designed to report the history of what’s happened and provide a drill-down to understand why it happened. In essence it provides an intermittent, backward historical view of spend that then can be used to manually plan and take future actions.

How AI is Different Than Spend Analytics

There are at least 2 main differences between spend analytics and AI in procurement:

  1. Backward vs. Forward Orientation
  2. Internal vs. External Focus

Backward vs. Forward Orientation


Spend analytics are inherently descriptive and focused on evaluating what happened in the past and what caused it.

AI, on the other hand, synthesizes historical and near real time data, trends, and statistical analysis to predict future events and prescribe specific courses of action.

Spend analytics tools typically are passive and reactive….you have to run the report, manually conduct drill down, conduct analysis and plan future actions.

AI is persistent, vigilant and operational. AI can constantly be scanning data to identify trends, conditions and anomalies to proactively alert specific users to opportunities and risks as well as recommend specific actions.

This move from reactive, static information to proactive, persistent operational insights is key to unlocking latent value from legacy procurement. In some ways I consider the full transition from descriptive periodic spend analytics to vigilant AI like moving from driving your car only using the rear view mirror to driving using your front windshield, constantly observing, and making changes in advance of changing conditions.

Internal vs. External Focus

spend-analytics-internal-vs-externalSpend analytics tools also tend to be internally-focused, using data internal to the systems of record within an enterprise.

The intent and promise of AI in procurement is the ability to leverage not only internal data, but also to actively capture and use data from external sources to predict outcomes and and prescribe actions.

Supplier risk provides an instructive example. External news and data regarding labor unrest, financial incidents, natural disasters, etc. can be used in real time to not only identify the risk, but more importantly prescribe the movement of orders to alternative suppliers.

AI in Procurement vs. Spend Analytics: Different Purposes, Different Expectations

Spend analysis is often represented with a magnifying glass or bar chart icons to convey its power to see details and produce reports.

spend-analytics-vs-ai-in-procurementAI is more proactive, predictive, and prescriptive.

An AI symbol might be a factory and a mail icon, representing that the machinery of AI is constantly analyzing data to turn out messages that lead to action. While managing cost using spend analytics has been described as “driving a car by looking in the rear view mirror,” AI solutions promise forward-looking insights that combine past usage data with emerging trends and external marketplace data.

In addition to having different functional emphases, there are differences in technology between the two solutions. Comprehensive spend analytics require a full view of the enterprise, so data integration has always been critically important. When spend analytics systems were first developed, this integration involved feeds from CD parsers, magnetic tape, different EDI standards, and optical character recognition (OCR) from scanned paper invoices. Most companies integrated spend analytics with their enterprise financial systems, which were typically based on the installed on-premise software model.

Because AI solutions in procurement came late to the party, they enjoy important technological advantages. Several generations of electronic billing have worked out the kinks of data transmission and translation, so more detailed vendor data can be imported with ease. Also, because the AI solutions are native and cloud-based, they are nimble and free from client-server management overhead.

Perhaps more importantly, AI in procurement is not designed to be a primary resource for purchasing, accounting, or anything else. The “system of record” burden meant that it was hard to get substantial value from a spend analytics solution until nearly all the data was accurately loaded and correctly classified, and for some companies that became a multi-year effort.

By contrast, AI can be applied as a light solution. There’s no rip-and-replace cutover because it operates best as a SaaS solution with feeds from existing systems, and it can provide value with access to even a small portion of company data. Many times, AI solutions can be implemented without IT involvement. This translates to lower cost and delivering value in weeks instead of months.

Spend Analytics Working With AI in Procurement

At this point it may sound as if AI is the favored little brother to spend analytics: not only is he smarter and more charming (with a cooler name!), he doesn’t even have to take out the trash.

But it’s not a competition; in this case both can play important roles in their family.

In fact, the greatest benefits will come when they work together. AI procurement solutions tend to be facile with many types of inputs (e.g., GL, expense reports, credit card statements, vendor feeds, and contract management tools). When AI brains are connected to a mature spend analytics software system, companies can realize spectacular benefits by applying cutting-edge analysis on their full body of detailed enterprise data. These benefits come from moving from reactive, periodic, manual, and descriptive to proactive, persistent, automated, predictive and prescriptive.

Is AI in Procurement Right For Your Company?

But would you want both solutions? For many companies that have already made significant investments in spend analytics, the practical question is whether they need to add an AI solution to the mix. Is a layer of AI worth the investment?

That question is sort of like someone who owns a shovel asking if they’ll need a snow blower. To give a helpful answer requires more information.

The first question might be, where do you live and how much snow do you get? Increasingly this is the same for all of us. Given our world of increasing vendor spend and exponentially more data from inside and outside our organization… we all live in Buffalo. (There’s a LOT of snow in Buffalo.)

That’s relevant, but it doesn’t necessarily settle the issue. We might ask: how long is your driveway, and how important is it to you to have it clean? How much does that snow blower cost?

And, importantly, what if my competitor has a snow blower and all I have is a shovel?

Beyond the snow metaphor, companies must consider the competitive marketplace. As AI solutions are adopted more broadly, those teams who delay or pass may fall behind. According to a 2016 study of over 340 large enterprises by Keystone Strategy, those companies that use data effectively enjoy gross margins and operating margins that are respectively 18% and 4% higher than their peers who lag behind.

Given the potential benefits, it seems that companies over $1B in revenue are prime candidates to leverage an AI for procurement solution. And depending on the industry and procurement maturity of an organization, smaller companies – say, several hundred million in revenue – could also potentially benefit.

More than anything, the decision comes down to priorities.

If your procurement organization is still developing its foundational processes, policies, tools, and team, then it may not make sense for you to jump right into the AI pool just yet. Get you basic spend analytics and other foundational elements in place first.

If your procurement organization is on the more mature side of the spectrum – you’ve implemented a procure-to-pay system, have well defined processes, tools (including spend analytics), then I’d suggest giving AI a look. Enterprises of all kinds now compete largely on data, and procurement owns the data associated with stewardship of spend representing 30-60% of every revenue dollar that comes into the company. Gaining ever-increasing value from that data via cost reduction, innovation, growth, new products, risk reduction, and improved responsiveness and service is Procurement’s core responsibility.

Let me know your thoughts. What experiences do you have with spend analytics and these new AI capabilities?

jack-quarles-expensive-sentencesJack Quarles is a 20-year sourcing veteran and author of the bestselling books Expensive Sentences, Same Side Selling, and How Smart Companies Save Money.

If you have questions about this article or want to learn more about AI in procurement solutions, contact Suplari.

To schedule a demonstration of Suplari Insights schedule a demo.