SugarCRM CEO - ignore AI “parlour tricks” and focus on people
- In the first of a two part interview with SugarCRM CEO Larry Augustin, AI hype and sensationalism comes under scrutiny.
“Do you like that model?” asks Larry Augustin, SugarCRM CEO with a hint of mischief in his voice. Something extraordinary has happened. He's just talked about Apple – the company that (my contact at Apple says) signed an organisation-wide CRM deal with Sugar earlier this year, but about which Sugar has maintained a vow of silence.
But Augustin is merely asking after my laptop, which is sitting on the table of a West End bar that’s filled with all the shouting and hubbub of London’s pre-Christmas market. Indeed, the venue is so noisy that we repair to a lounge at the back of the hotel for a mid-Western fireside chat: much more fitting for the CEO of a company that is much quieter and more self-effacing than some of its brasher industry peers.
Augustin explains why Sugar sees things differently:
We’d observed that CRM products were among the most disliked among users. The goal was to build something that actually helped the salesperson, the end-user, to do their job and that they liked to use. As opposed to something that was about Big Brother, management, and tracking what they were doing. That was the original concept and it’s stuck through the company. We still focus on that.
We’re always looking for ways to enhance that and make it smart – we invest heavily in mobile, and then there are things like Hint [Sugar’s AI-enhanced relationship intelligence tool, which debuted six months ago]. We want to bring more intelligence to CRM: relationship management and intelligence, rather than just CRM. We think that core ability helps them manage business relationships, stay in contact with them, understand them better, and connect with them. And thinking of it in those terms helps us talk about how what we do is different.
In an enterprise market that has long been defined by big IPOs, big stories, and bigger noise, Sugar has maintained a low profile and kept itself private, in every sense, building brand by word of mouth rather than by the ‘bread and circuses’ approach of Oracle, Salesforce, and the West Coast jamboree, where CEOs make a show of rubbing shoulders with celebrities and politicians. Augustin smiles:
In the US, we would call that my mid-Western background. I was born and grew up in Ohio and went to school in Indiana, and we have this attitude of we just get things done, and it’s not about bragging. That’s part of the spirit of the company. We’re all about doing things well, and we want people to like doing business with us, because ultimately it’s about relationships.
That isn’t about bragging all the time; it’s about being grateful that someone is your customer, and acting that way. It’s not about command and control. And a lot of that has been lost in the market. If you listen to the way people describe their relationships with enterprise software companies, it’s generally not positive.
But given that marketing itself is all about generating noise, does Augustin ever look at the Silicon Valley carnival and think maybe there is some value in it? He says:
It’s not that there’s no value in it, I think it’s about how you do it. Doing it in a way that doesn’t compromise what you value as a business and doesn’t make people embarrassed to be your customer. We want to make some noise, but the balance is to do it in a way that’s honest.
Our customers have understood that command and control isn’t sufficient to drive a business. It’s not about activity tracking, it’s about helping them to do their job – there’s a whole market of failed adoptions out there for us, systems that companies have tried to force people to use, and it doesn’t work. The net promoter scores of many of these companies are zero, and that’s a pretty low bar. Sometimes these companies are more hype than delivery. That’s not the way we do business.
Sugar has refocused itself on relationship management and intelligence, which puts AI at the heart of its portfolio – a move that positions Sugar alongside IBM, Microsoft, Salesforce, and others who are riding the hype wave of 2017’s must-have (or must talk about) technology. Recently, Genpact CEO ‘Tiger’ Tyagarajan slammed the buy-side AI market, saying that many implementations are doomed to fail, because they’re being adopted tactically, rather than strategically. Of the backlash, Augustin says:
I’ve been a sceptic around how AI is often positioned as this magic technology that’s going to solve all things. And in our industry, people often hold up the idea that customers will talk to a robot, talk to the AI, rather than to a person. But I don’t think that’s going to happen – well, it may happen for mundane tasks, and that’s a good thing as it will take people away from things that don’t add value, but I think we have a long way to go before AI is doing people replacement.
Understanding the key concepts that people care about and being able to recommend the most likely fit in terms of content... things like that are not way-out AI applications. And if you’re a salesperson and you’ve got to talk to a client, you Google them, look them up on Facebook, LinkedIn, and their Twitter feed. It’s not science fiction to be able to do that for them: to go out, do those searches, bring them together, and aggregate them. The next step will be to summarise those for the person consuming them. Some people might not even call it AI.
Augustin agrees that the sensationalist, fear-mongering way that AI has often been talked about – mass unemployment, malign AIs, killer robots, and so on – been unhelpful:
If you want to sell newspapers, scare people or tell them they’re going to lose their jobs. But the new jobs that are going to be created will overwhelm all that. So it’s more of a change management problem: training, education, change management.
I don’t see AI replacing human beings, I see it replacing tasks that people don’t want to do, or which are repetitive, or not high value. Of course, they may have been high-value tasks 20 or 30 years ago, but they’re not today, and that to me is about augmentation. AI should help people to be more human, and stop doing the things that robots should do. Take ATMs. There are more bank tellers today than when ATMs came along, but they don’t give out cash, their role has expanded to solve different problems.
That said, there is a chasm between what most AI/robotics vendors and researchers sincerely believe they’re designing – assistive technologies to complement and augment human skills – and what many customers and think tanks believe they’re getting – a means to slash costs and headcount. That raises the question of how much should vendors focus on educating the customer that AI is a complementary tool, not an alternative to people? Augustin argues:
The vendor/customer mismatch, it happens a lot in AI. And in the space we play, people bought a lot of the legacy CRM products, feeling that what they did was relationship management, but what they got was command and control. With AI, there are vendors who provide technologies that are designed to automate tasks and replace or eliminate the need for people. They’re always going to sell that way, and as AI comes along it has a great story to tell about that – if you believe the hype. But we’re very clear that that’s not the business that SugarCRM is in. We deliver value by making people more productive, so you get more value out of the workforce.”
Of course, Sugar sells its product per seat, per user, and so decreasing the headcount is also not in its own interests. But what is Sugar’s strategic vision for AI, when the avoidance of hype, noise, and people replacement seems critical to success? Augustin explains:
Our focus in the near term is to build out the suite of available data sources, so we have that covered. Recently we added news, and we’re adding more information on companies. We’ve demonstrated personality insights – drawn from Twitter, Facebook, LinkedIn, and so on – but we don’t have a commitment to productise it as yet. And we’ve demonstrated content recommendations based on that. Another data source we want to roll into Hint is email: that’s an important data set that we want to productise.
Personality profiling – or insights, to use Augustin’s word – is a contentious, high-risk, area. So let’s hope that AIs can learn to differentiate between genuine emails and spam – something that many cloud-based email systems are still unable to do, despite nearly a quarter century of progress. The risk must be there that AIs move into the spam business themselves, simply by being deluged with misleading information. Augustin is cautious:
We don’t have a plan to productise it, as yet. The question is whether it will be useful. There is a lot of talk about AI, but we want to make sure that it isn’t just more noise. A lot of AI I would characterise as parlour tricks, and we don’t want to deliver a parlour trick. We’re looking at potential features, but we want to be sure that we’re delivering value, rather than something that looks good in a demo but no one actually uses.
He cites chat bots as case in point, adding:
I’ve seen tools that purport to replace your sales development rep, and automate the process of responding to emails, those seem to be a lot of parlour tricks. Canned interaction: it looks impressive, but in the real world, not so much. What I would say, though, is that sometimes the customer doesn’t understand what they should be buying, and so will lean towards the wrong thing. People need a lot more help and expertise. So we think a lot about how we can deliver more of that expertise. It’s not just about the technology.
With so much noise and hysteria around AI – the B2B market equivalent of a noisy bar at Christmas – there is value in vendors stepping forward to offer pragmatic, practical guidance and education. So let’s hope that, in a world of big shows and mega-budget distractions, buyers will still listen to the quieter, more sensible voices, and not just the shouting men in their stripy suits.