Oracle Live in review - Steve Miranda on how customers get from legacy to agile, and the potential of voice UI for the enterprise

Profile picture for user jreed By Jon Reed July 15, 2021 Audio mode
When Oracle Live comes around, Steve Miranda is always up for a good back and forth around the announcements. This time up? I ask Miranda about the challenges of getting to cloud agility - and he responds to my questions on voice UI for the enterprise.

Steve Miranda - Oracle Live
(Steve Miranda, presenting at Oracle Live)

Let's face it: virtual events have a reputation for landing on the bland side. During his  Oracle Live presentations, Steve Miranda avoids that trap, leaning on the spicy.

This quarter was no exception, with Miranda airing out on Oracle cloud customer wins via the legacy installs of a well-known competitor (whose name readers can surely guess, but you can also read about via Miranda's The future of business – conversations with customers who are leaving legacy apps behind). My colleague Madeline Bennett also profiled TTX, one of these customer wins.

I've already previewed Oracle Live via a surprising piece on ERP/applications support, ERP support as a competitive advantage? Yeah, you heard that right. Oracle's Maz Songerwala makes his case. Now, the entire event is available on demand. But what should audiences take away from Oracle Live? During my video chat with Miranda, he pointed to three key themes. The first is the customer wins. Next up:

Cloud applications have a different speed of innovation

Miranda explained: 

Because it's a SaaS platform, all 13,000+ of our cloud ERP, HCM and CX customers all get the same version. They all get these feature functions that they can enable and turn on. Particularly versus our ERP competitors, we believe we have a substantial differentiation, and a payback for the investment we've made those years of building out the SaaS platform.

Favorite new feature? Voice-enabled financial skills for cloud ERP

Asking an executive to pick their favorite new feature is probably terrifying for the marketing team, where all new features are equally beloved. But Miranda singled out Oracle's new ML/AI features:

This time around, we're highlighting machine learning as it applies to cloud ERP. One example is the financial skills reporting, adding to the digital assistant. You and I have talked in the past on the digital assistant, Oracle voice or conversational UI. We've had HCM skills in the past, but now it's around ERP, and specifically around ERP reporting.

Can the enterprise catch up with consumer-grade voice UIs? Miranda:

In our everyday lives, we can use our voice and find out what's going on with European Cup soccer. You can find the weather tomorrow; you can find out what stock is at its 52 week high. You know, 'Hey Siri, hey Alexa, hey Google, what's the Oracle stock price? And I get the answer.

Unfortunately, for most enterprise systems, using your voice find out, 'Hey. what's the application revenue so far this quarter? And how's that compared to last quarter?' Up until now, that's been difficult, if not impossible, and worse, sometimes difficult just to find that information, forget about using your voice. So having that digital system skill on top of ERP now, for this kind of basic reporting and questions, we think it's extremely valuable.

Cloud customer success - what are the challenges?

I shared some skepticism with Miranda about voice UIs for cloud applications, but let's table that for a moment. On Miranda's customer examples, I noticed all three were tied into speed and agility, enabled by modern business applications. It's hard to quarrel with agility - nor would I quarrel with the caliber of modern cloud apps. But that doesn't mean cloud application projects are easy. I asked Miranda: what are the biggest challenges to achieving these types of outcomes? He responded:

The first step is getting your team behind making the transition to cloud software. This is no offense to any of our customers - nor our competitors frankly - but most of the big ERP implementations were really centered around Y2K. So we have the theme that customers have done a lot of M&A, some of which has been rationalized, some hasn't.

When they needed to pursue new business models, customers found their legacy software didn't cut it:

Customers have a lot of new business models, and COVID has just accelerated that - dealing with your customer digitally. Oracle ourselves - we're one example of many that's gone from a product company to a service company. Just about everything today, you used to buy, now you can rent or have it on a subscription-type basis. These are major shifts in terms of the business processes, and the need to modernize from when the systems, on a grand scale, were put in.

But it doesn't stop there:

The second step is just to get the organization used to that, in our view, this is the last major upgrade you're ever going to do. What that means from the technical side is: 'We'll let Oracle worry about it.' What that means from the functional side is: don't rely on a landmark anymore, like, 'Well, let's just wait for this [major] upgrade.' Functional areas are going get very similar to what you get on your phone. Your phone probably has an alert right now that says, 'Update these apps,' and you press the update, and you've got a new feature.

My take

There was a time in the ERP, HCM and CRM markets where the limitations of legacy software prevented vendors from delivering what businesses needed. But today, as long as you select the right software, I think those risks have greatly diminished. Oracle has done as good a job as any vendor of documenting that, via customer use cases across industry.

And yet, to achieve the type of agility Miranda describes, where you are heading towards "continuously improving processes," you still need plenty of good old-fashioned change management. Miranda himself brought it up, and I agree.

As for my skepticism about voice UIs, that goes back to my own failed experiments pushing consumer-grade voice UIs to their (narrow) limits. I'd like to think that voice-enabled ERP could answer questions like "Tell me my most (or least) profitable retail customer," but as of today, I don't see it. However, Miranda had answers for me there. He said that even with the current limits on the sophistication of voice queries, useful information can still be surfaced from Oracle apps (in this case, financial information). Miranda believes Oracle's job is easier than Siri's or Alexa's, which has to field an enormous range of queries. The key for Oracle NLP is to get the right data trained and loaded:

When we added the digital assistant for vacations, we had an enormous number of synonyms. But we have the metadata from our customers... So we know all the different types of words our customers use for time off. So we were able to pre-seed a huge vocabulary of synonyms. 

But there is work with voice UIs ahead, specifically around compound questions:

You are correct that, especially when you create a compound request, or ask something like, 'What is my fastest-growing business in SaaS applications over the last two quarters?' We're not there yet. However, when I say we're announcing it, it's going to get better and better every quarter. We'll learn from what our customers are asking. Frankly, what we can't handle today, that gives a guide on what to build next. Two is, we'll get better in terms of the machine learning technology to do that.

Still, Miranda thinks the utility is already there:

I think the pragmatic speed will be very, very good, because we're not solving a generalized problem. We're solving a particular problem that a lot of people have.

I'm open to seeing how that works. The quality of enterprise search remains one of the biggest pain points for users in our industry. If voice UIs can make even a small dent there, that's a good thing. As long as users can find that sweet spot of "questions I can get useful answers about today," without too much lost time trying queries beyond today's scope.

There's always more new features released at Oracle Live than I could cover in one article, but you can check out the highlights here. Other additions Miranda is excited about? Machine learning enhancements to performance management for the EPM Cloud. And: improved workflows and dashboards around risk management, a fraud detection use case Larry Ellison has also emphasized. That's plenty to absorb for now.