Coupa Inspire – new leadership and AI tools, same community buzz

A lot has changed for spend management vendor Coupa in the past year, in particular the departure of its founding CEO Rob Bernshteyn last May, to be replaced in November by Leagh Turner, former co-CEO of enterprise HCM vendor Ceridian. The management shake-up followed the company’s buyout by private equity giant Thoma Bravo in February 2023. All this while the broader tech industry was coming to terms with the rapid rise of generative AI. Customers gathering in Vienna this week for the EMEA edition of its annual Inspire conference may be wondering how much has changed since they last got together.

Judging by what transpired at the US edition of Inspire, which took place in April, they’re likely to be reassured by what they hear. There are some tweaks to the messaging, with Coupa now positioning itself as “the margin multiplier company,” and some significant new AI offerings, but the new leadership has doubled down on what is probably Coupa’s greatest asset — its customer community and the huge volume of spend data accumulated with its customers’ consent. Coupa has been applying AI to this data store for many years, but with the addition of generative AI, the results are even more powerful. In her opening keynote at the Las Vegas event, Turner leant into this theme:

The data has been safely and ethically sourced for years, from thousands of customers and millions of suppliers. It’s protected and it’s safe and it’s secure. And it’s only used with private Large Language Models.

The results that you get from this community-generated AI will help you drive your business forward. You contributed your data to this community, so that everyone can benefit from the most accurate and useful AI that is embedded throughout the entire Coupa platform. This community made that possible.

The collective experience of what you do every single day is training your AI. You’ve made it real. It’s not hype. And you are exactly as you always have been, you are a force for good. And that is really big. And it’s really unique. There is no other community like this one — there isn’t — with this data store. In the world of enterprise software, it does not exist. And this is a direct result of this community.

New generative AI powered capabilities

Bringing generative AI into the mix alongside the existing work that Coupa has done with machine learning and AI over the years is opening up new capabilities that weren’t within reach before. One example is the newly launched contract intelligence feature, which uses generative AI to ingest a contract in any document format and automatically identify all the key data points within the text, including risk areas, obligations, conditions and so on. I recently spoke to Donna Wilczek, EVP, Strategy & Innovation at Coupa about the Inspire announcements, prior to her announcement last Friday that she has joined the exodus of longstanding executives leaving the company. She told me:

We tried to solve for that use case about 18 months ago using traditional AI methods, machine learning, deep learning. In fact, we were unsuccessful, Our rate of success of reading this unstructured data and properly structuring it was not very good…

[Then] GPT-3 launched. We said, let’s try that same use case with GPT-3, and were successful… We were able to natively build that using our own GPT, trained on the Coupa big data sets.

It’s for me spectacular, because if we can rethink through all of these use cases that have surfaced through the years that we were unable to deliver on, we now have a great new tool to be able to solve and deliver them natively.

This tool can be used to analyze both both buy-side and sell-side contracts, as well as many other types of agreement with customers, partners and so on. For many organizations, this provides an opportunity to surface and keep track of often long-forgotten commitments and undertakings. She added:

It’s still surprising to me to see how many businesses don’t have a handle on all the contracts that they’ve signed. They’re sitting in file cabinets, or they’re sitting in digital folders that are unstructured. You ask the CFO, you ask the CEO, ‘Are you aware of all your contractual obligations?’ Nobody can tell you. So this is a way to create that control and visibility.

With the launch of Navi, a generative AI-based agent, the new AI technology provides a more automated, conversational means of interacting with the Coupa system. Rather than the user needing to know how to search for information or build an analytics report, Navi is able to understand the user’s intent and take the required actions on their behalf. For example, it can create recommendations on supplier risk or generate a CFO dashboard, without the user having to work through a set of manual configurations. Wilczek told me:

[It] helps a person accomplish their tasks faster and in an automated fashion. Not just retrieving answers to questions, but actually doing the work and doing the task. For example, saying in natural language, ‘Build me a dashboard for my CFO that gives the CFO a great lens of the last 12 months.’ AI interprets that request and determines the configuration options, and then automatically builds those dashboards right in Coupa.

Optimizing results

Here again, the pairing of generative AI with Coupa’s historical data set of what procurement and finance professionals have done over the years helps to optimize the result. She added:

As it develops that configuration for you, it’s optimizing what would be the prescribed recommendation for what the CFO should be aware of. What it’s using for that interpretation and that prescription is that big data set of customer-contributed data that is guiding to, what are other CFOs in this community looking at? How do they think about their business?

Generative AI is also adding new layers of functionality in various aspects of supply chain collaboration, such as around forecasts, category planning and purchase orders. For example, where suppliers are communicating order status via email, generative AI helps automate the ingestion of those emails to speed up reaction time. Wilczek explained:

We were able to leverage Gen AI to read the inbox, to read the emails, to interpret what the email is saying, and to take that data, that non-structured data, and turn it into a structured response in Coupa. So the supply chain planner no longer needs to do that transcription. They come into Coupa and they see the draft messages that have been converted from those emails into draft confirmation statuses, or delivery schedules, or other actions in Coupa.

The system also learns from adjustments that are made to the draft messages so that it can improve the accuracy of its responses over time. She went on:

The person can make adjustments if needed, or accept the draft as is. Each action that that supply chain planner takes is used to then auto-train and reinforce the Gen AI. So if the draft was accepted as is, it’s a confirmation, you read it properly, keep doing it that way. If a change is made, the Gen AI continuously learns from that change, and says, ‘Oh, I made a mistake on that confirmation, edit and revise what I’m doing for the next time.’

This data of these reinforcement actions are coming in across the entire customer community. So you’re getting a big data set that is constantly reinforcing the work that was done, or saying changes need to happen.

My take

A change of ownership followed by sweeping changes in leadership always creates a period of uncertainty for a vendor, when customers wonder whether their experience will change. Despite ongoing departures such as Wilczek’s, that period of uncertainty seems to have largely blown over in Coupa’s case. The culture remains intact that helped it build its customer data store into the immensely valuable asset it represents today, and the new management can hardly ignore how much that value is now amplified by the addition of generative AI.

Following Inspire US, I spoke to Anders Lillevik, an experienced procurement professional and now founder and CEO of Focal Point, a Coupa partner which helps organizations manage strategic sourcing decisions. He agreed that there was some nervousness after the Thoma Bravo acquisition, but that this had largely dissipated by the time Inspire came around, with incoming CEO Leagh Turner providing a sense of continuity, based on what he heard at the event. He told me:

They have always come to the market touting a very strong financial ROI. That hasn’t changed… The devil’s in the details, but if they were able to save money, mitigate risk, speed up the process and manage ESG, it’s a pretty strong and compelling argument.

It’s quite a change since last year’s Inspire, when Bernshteyn was still holding the reins as CEO. But the strong culture that Coupa built up under his tenure appears to have survived the transition to a new leadership that still values its community ethos. The shared data store that this community has built up makes for a potent combination with the latest AI technologies, and Coupa seems well positioned to continue to make the most of this unique asset.