Apttus - applying artificial intelligence to quote-to-cash

Profile picture for user ddpreez By Derek du Preez May 5, 2017
Summary:
Machine learning and AI assistants will help organisations improve the performance of their quote-to-cash lifecycle, says Apttus’ Elliott Yama.

Apttus Max digital assistant 370px
Fast-growing cloud vendor Apttus is helping enterprises improve their quote-to-cash processes by not only breaking down the silos between the functions that have traditionally been applied to these actions - which sit between CRM and ERP, and include quote, contract and revenue management - but by also changing the behaviour of the organisation through machine learning, artificial intelligence and virtual assistants.

I got the chance to speak with Elliott Yama, VP of best practices and knowledge management at Apttus, this week at the company’s Accelerate event in San Francisco, where he explained how behavioural change is at the centre of of the company’s approach - not just a focus on improving business processes. For more insight into Apttus’ strategy, read my colleague Phil Wainewright’s piece here.

Yama explained that companies should be looking beyond automation to gain efficiencies and boost performance in their quote-to-cash processes. He said:

If we think about how Apttus has organised, we have process automation for those key steps, but we have behavioural nudges that we provide to the various constituents, whether they’re my direct sellers, my partner sellers, or even the end customer. We use incentives, promotions, rebates to basically inspire the type of behaviour that drives the outcomes that I need. It’s absolutely nudge theory.

If that’s not integrated into the solution design, or that information isn’t provided to the seller at the point of decision making, then I’m probably losing an opportunity to drive the type of behaviour that I’m looking for.

Layered intelligence

Apttus believes that intelligence should be embedded throughout the quote-to-cash lifecycle. However, this is combined of a number of intelligence layers, says Yama.

At the base level, enterprises can make use of descriptive intelligence - charts and dashboards, to try and understand what has happened historically. From there, predictive intelligence can be applied - where if you have enough data, a company might be able to anticipate what might happen.

Going beyond those two base levels comes artificial/cognitive intelligence, where the solution can advise on the best course of action. Yama explained:

It can actually give you a set of actions - I think you should sell this offering to this customer and I think you should sell it at this price, to be consistent with the actions that you’ve taken in the marketplace in the past.

So I can give that set of choices and even automatically bring that up as the starting position for that seller, if I have enough of that data and I can bring those analytics to bear. At this point we are beginning to bump into what we would call cognitive intelligence, the ability for the solution to understand, reason in a human like way and learn over time. So the system becomes more valuable in the insights that it provides as it goes forward.

Yama argues that this can be done instantly if a company has enough comprehensive historical data and results can begin to be seen immediately. However, that doesn’t mean that barriers don’t remain. Namely, the human factor.

Yama highlights two areas where cognitive intelligence used to nudge behaviour may face resistance from people within an organisation - when it is provided without context and out of fear. He said:

Number one, does the user trust the recommendation? Do they have any context for this recommendation? We have seen this going back to the very first customers that we started working with, which is: I can give you a recommendation that is mathematically absolutely the best recommendation that can be generated, but it may not inspire you to actually promote that recommendation or accept it.

We go to extra effort to provide context to that user about the recommendation. A simple example is, ‘this product is recommended for this customer’. Why? This customer looks likes these customers, and here’s some characteristics about all these customers, they all fall in this sector, and they’re of this size, and they generally have practice areas that look like this - you name it. That’s the first piece, just proving context.

And on the rise of the bots, he added:

The other piece is kind of the rejection factor, which is more born of fear. ‘The robots are taking over the world’. Our design point is one where we keep the human user at the centre of decision making. The beautiful thing about that is that a number of research studies have shown that a human supported by a smart model outperforms the human alone and it outperforms the smart model alone.

So we really see our mission as providing intelligence to the user to help them become more efficient and more effective about the decision they’re trying to make at this particular part of the revenue making process.

Introducing Max

Building on these intelligence capabilities, which are largely being powered by the Microsoft Azure platform, Apttus is now making a move to introduce new conversational interfaces via the use of an intelligent virtual assistant called Max.

The use of virtual assistants and chat bots are growing in popularity, not just in the consumer world, but also in the enterprise. There is a growing belief that conversational interfaces will become so sophisticated that it will be the primary way that people interact with software in the future.

During the Accelerate event this week, the live demos of Max seem very impressive, given that it is still early days. From what I saw, Max was able to understand and carry out fairly sophisticated tasks simply by text or voice command.

Yama said that enterprise apps haven’t traditionally been easy to use, but the future of virtual assistants could change this. He said:

It’s still early, but in terms of where we see it going, think about it, I interact with people that I’m closest with at work and at home. I text them. I use this mini supercomputer that I carry with me everywhere. So, I’m mobile and I’m social. That’s not the way that enterprise applications have been developed. They assume that you’re at a desk and you log in. They don’t want to log in, it’s kind of a drag to make a travel arrangement, or submit an expense report, or fill out a pipeline report, or advance a commercial opportunity.

What Max does it allows me to simply speak or text to Max, to get information or execute tasks in a way that I do with a peer, or a colleague. In that regard, I think that’s the way that we will interact with technology in the future. In fact, that’s probably the only way we will do it. Today it feels new, because we are learning how to do that.

However, again, human resistance to change will be a barrier - but Yama said that future generations will have less of a problem with a conversational interface, as this will be what they’re used to.

I don’t know if there are gripes, but humans in general are slow to change. And there are certainly people in Apttus that don’t use Waze, or Uber, or Alexa. But there’s a generation of people who are product owners for Max that can talk about the chat bots and applications that are familiar to them on a daily basis and I don’t know 90% of them. In that regard we have a whole generation of people that are coming along behind us that will never drive a car and they don’t want to drive a car.

And where does Yama see Max going? How much could it be applied to in the quote-to-cash lifecycle? Well, it appears that Yama believes there are no limits and that Max could even be of benefit to apps outside of the quote-to-cash remit. He said:

We literally see Max as a conversational user interface for enterprise software. Period. We naturally are kind of biased towards the quote to cash, but we see some of the inefficiencies in the CRM space, reps don’t like to log in to CRM to update pipeline because it doesn’t help them. In fact it gets them attention they don’t want. So if we can make that experience more efficient, more effective for them, it will businesses to do better planning, better forecasting. That’s just a very small subset of the enterprise universe, but the opportunity is boundless.

My take

A fascinating conversation with Yama and a good day with Apttus. It’s clear that they’re doing what they can do simplify what has traditionally be a very long, complicated, disparate and manual process - from configure price quote, to contract lifecycle management, to order management, billing and revenue. By unbundling these processes and rebuilding them into an integrated, automated and streamlined new bundle holds obvious value for the enterprise.

And as I’ve noted above, the move to nudge behaviour with intelligence and new virtual interfaces is very interesting. From what I saw, the use case is strong.

There are two caveats to this, however, which are closely linked. Smaller companies that don’t have the data scale and are doing a small amount of deals each year are unlikely to see much value in the more sophisticated machine learning capabilities. That’s not to say that the other intelligence isn’t useful, but the more advanced capabilities will likely be procured by the large enterprises.

Secondly, Apttus is currently only applying machine learning and AI on a company by company basis. There’s an opportunity here to aggregate and anonymise the data it has on its platform and provide insights on an industry-wide level to each company, which in turn could then be of use to the smaller players too (the likes of Coupa and InsideSales are already doing this).

Yama said that this requires some buy-in from customers, as it involves them handing over their data for the greater good. He said that some customers have shown interest in this and Apttus is working with them to explore the potential. That could make for a very interesting platform.