AI and ML to Enable More Intelligent P2P Solutions: An Interview with Jesal Mehta

AI and ML to Enable More Intelligent P2P Solutions: An Interview with Jesal Mehta
The Siliconreview
14 August, 2019

With an evolving landscape of technology innovations, Artificial Intelligence (AI) is clearly at the head of the pack. Machine Learning and AI’s potential power and influence is far-reaching across industries and lines of business. Both Procurement and supply chains professionals are planning to leverage AI/ML to address long-term challenges related to digital procurement workflow.

The Procure-to-Pay solutions have been there for ages and there are some high-level advancements around it. However, the level of maturity and changes required to get the most out of what AI & ML has in store for P2P use cases appears challenging. Very few organizations are thinking of the future and actually coming up with innovative real-world AI-enabled solutions to make work-life easier. One such company, Aavenir is at the forefront when it comes to providing revolutionary levels of AI-enablement in Procure-to-Pay domain with disruptive technologies like Machine Learning (ML) and Natural Language Processing (NLP).

After having decades of experience and building break-through products in the procure-to-pay domain, Jesal Mehta has recently founded Aavenir with a vision to deliver the future of work. Armed with intelligence, Aavenir’s cloud-based Procure to Pay solutions enable enterprises to automate manual and time-consuming tasks with advanced machine learning and artificial intelligence technologies. Aavenir’s P2P solutions are primarily built to empower employees across enterprises to work efficiently in a collaborative manner.

We spoke with Jesal Mehta, the Founder, and CEO of Aavenir about the future of P2P solutions landscape and how ML and AI can enhance the way enterprise manage Procurement to Payment operations today. 

Can you shed some light on the current market scenario? How is Aavenir is going to disrupt with mindful innovation?

In the typical Procure-to-Pay (P2P) workflow, once a purchase requisition is placed, you look for a supplier, do the negotiations and contract agreements with the supplier, and issue a purchase order. The supplier sends an invoice and you release the payment. In this process, many of those things are still being done like it was done in the earlier century. With Aavenir P2P solutions, we are introducing innovative and new technologies in P2P domain that can help enterprises to analyze P2P operations, identify bottlenecks and define a new competitive strategy to streamline procurements. For example, let us say that you receive an invoice; as a typical large organization, you would have upwards of several thousand vendors and each vendor brings in its own different invoice format. When the thousands of invoices are coming to you, your accounts team need to manually read and decipher those invoices. However, since those invoice formats are so different from each other it is extremely difficult to put it in one unified software for account payable automation. Therefore, many organizations work with a BPO (Business Process Outsourcing) service to extract the data or they are just happy with entering the data manually.

Aavenir Account Payable Automation Solution, an integral part of Aavenir Procure-to-Pay solutions, comes with advanced Machine Learning (ML) engine which reads any invoice in any format and extracts the data for you. You may have several thousand different types of invoices and we can handle them all. Incidentally, we are working with one of the leading global organizations who happens to have sixteen and half thousand unique vendors with an equal number of invoice formats with four hundred thousand invoices per year and we are working on solving the problem for them.

With many enterprises claim to offer AI-enabled solutions, how do you measure AI maturity, what is AI adoption blueprint for P2P solutions?

Unstructured Data extraction using NLP and finding anomalies is a very challenging problem. At Aavenir, we have leveraged industry-standard Machine Learning models and we continue to customize them with high volume industry-standard training datasets to build an intelligent ML engine trained for P2P operations. These ML engines are being deployed on our P2P solutions, so it can decipher different variants of invoices or contracts with less amounts of data. When I say less amounts of data, we are talking about roughly in the single digits of thousands. For example, in the case of our customer with four hundred thousand invoices per year, I think we will be able to train that data with less than ten thousand invoices.

On the other hand, we use different Machine Learning algorithms for Contract Lifecycle Management (CLM). That is our second product, which is already available to use on ServiceNow App Store. Aavenir CLM also offers some interesting things to improve overall enterprise productivity and enhance user experience. For example, when a third party sends a contract on their own paper, we need to extract and import the important elements of those contracts; the terms and clauses of those contracts. In this particular case, we are dealing with unstructured data, such as text, word document or a PDF document, so our ML engines have to be smart enough compared to the others, i.e. invoice extraction ML engine. In the case of CLM, we have trained our Machine Learning model with twenty thousand different clauses and terms, for us to reach a satisfactory extraction from contracts.

What sets Aavenir apart from the other Contract Lifecycle Management systems in the industry? 

This is a very interesting question for all our customers and future prospective users. As I mentioned earlier, we deliver the future of work, we are actually on the leading edge of that. For example, in the case of CLM, the clause recommendation engine is fascinating where AI algorithm assists lawyers in negotiating contracts or using smarter terms and clauses to reduce business risks. Weallknow that contract negotiation takes weeks to months resulting in extended procurement and sales deal cycles and overall business slowdown. During the negotiation phase, when the other vendor/supplier sends a redlined contract, a procurement executive needs to be able to read those modified clauses and terms, compare them with the company’s internally approved language and then provide a recommendation to reduce business risks. It is a fascinating idea because we are now talking about understanding a natural language and applying machine learning together to really compare the intent. We are not comparing the actual words here, but we are comparing the intent of two clauses; one in our own library and other that is coming from the other vendor/supplier.

How do you find the current market scenario after the recent launch of your product?

I think the momentum and the market interest really surprised me! Considering the fact that Procure-to-Pay and Contract Lifecycle Management are in the market for a while, I was expecting it to be slow. However, in just a few days since we launched our product, we are seeing tremendous traction. Both of our products have captured the imagination of the people, especially when we talk about machine learning in that space. As I said earlier, momentum and interest in the market really surprised me quite a bit!

What’s your take on innovation? How do you foster innovation across all levels in our organization? 

We started this company with three core values; customer success, innovation, and joy- joy for our employees and joy for our customers. Innovation is one of the core values of our organization. We are not just trying to make noise about some of the things that we are doing. We are actually trying to do some interesting research on usability before we put across our ground-breaking work.

One of the commitments that our team lives by is—“Let’s find something new for this industry.” Every day we ask ourselves, “Did we do something new?” It’s okay to fail but, did we, as a team, learn and experience something new. We firmly believe that if you are just going to stay the course, it is not really going to take you anywhere. At the core of this belief is - allowing each employee to lead the new initiatives without the fear of failure. 

Describe the feedback you received after you launched your latest P2P solutions. 

Yes. Just to give you context, I have been in the Contract Lifecycle Management space for over ten years. I have a fairly good idea about the market and the gaps in the market. We are addressing those gaps right out of the gate. So far, we are getting really keen interest in Aavenir CLM launched on ServiceNow app store, even though it is just a few days! Also, we noticed that entire P2P space is an unmet need for ServiceNow customers and we are addressing for them. Almost 75% of Fortune 500 companies and many other global enterprises are using ServiceNow as a cloud vendor, getting value-added modules like CLM on top of the same platform is a clear win for them. No need to work with multiple cloud platforms, no need to acquire new administrative capabilities! ServiceNow is growing at a rapid clip of 33% YoY, and that certainly helps to get more traction to our product portfolio.

Finally, the whole advent of machine learning and artificial intelligence is really capturing the imagination of people, especially the visionaries in this legal and procurement space. That is because they can really see the potential where productivity can be bumped up significantly using these innovations.

Tell me three strong reasons to choose Aavenir.

Jesal: Advanced ML-based P2P solutions, ServiceNow as a platform, and very intuitive user experience are three primary reasons to choose Aavenir Solutions.

We have talked about ML and ServiceNow earlier, but it is important to elaborate on 3rd point, i.e. user experience.

Currently, when corporate procurement and legal professionals are buying something on Amazon for themselves at home, that experience is very clean, very contextual, very advanced with a lot of AI-based recommendation and so on. However, the moment they turn towards their work terminal they go back ten years. We are changing that. Aavenir UX gives intuitive user interface along with AI-based guidance and recommendation. Essentially, we are consumerizing the work experience! And, that is one more reason for anyone to consider Aavenir.