Interview with Conversica: autonomous AI-driven assistants generate big business

NL versie

For #WhatsNext, Artificial Intelligence is one of the most exciting developments we like to keep an eye on; with the potential to drastically change our lives and hopefully for the better. After all, the best technology can also be used at its worst, just ask Einstein.

Bill Gates calls AI the Holy Grail of Technology; Elon Musk, on the other hand, sees a threat in it and has set up NeuraLink to that end, in which he focuses on ‘improving mankind by linking our brains to computers’. We will soon have an article about this because Elon has something up his sleeve, that we’ll communicate about in the near future.

In any case, AI has taken off in recent years and is now successfully applied in various sectors, amongst others in B2B marketing and sales. We see a good example of use cases by the American company Conversica.

conversica-logo-horizontal.jpg

I had an interview with Amelia Farber, Director Global Partnerships at Conversica. We met at the Vlerick Business School in Brussels, at the LeadFabric ZEN event. An event focused on the latest trends in marketing and the increasing importance of Artificial Intelligence in B2B sales and marketing. Naturally #WhatsNext and Conversica had to attend such an event, so it was the ideal place to have our talk.

Conversia WhatsNext cover
Steven Dupont (#WhatsNext) and Amelia Farber (Conversica), at the Vlerick Business School

Steven Dupont: Amelia, can you briefly outline what Conversica is; what you do?

Amelia Farber: Conversica is a leader in the field of Conversational AI for companies and the only supplier of autonomous, AI-driven assistants for marketing and sales organizations. Our Conversica AI Assistants help companies find and retain customers faster and more efficiently by automatically contacting them through leads.

Basically, the assistant, or the AI entity, manifests as a junior member of your sales team – like an inside sales representative – and it fills that gap between marketing teams and sales teams.

We constantly see challenges when marketing teams hand the leads that they have generated to their sales team, and the sales team doesn’t have the bandwidth to follow up with all of those leads, so a huge percentage of those leads fall through the cracks – wasted opportunities.

We actually use Conversica internally, and our AI assistant is named Rachel Brooks. Rachel fills in that hand off between marketing and sales because she has both human and super-human capabilities: she is able to read, understand, and write back emails, have those two-way conversations with leads, but she also has infinite capabilities: infinite scalability (she can email thousands of leads at the same time but with totally unique, human-like conversations), infinite time (she works 24/7), infinite memory (she never forgets to follow up even if the conversation spans months and months of back and forth emails), persistence, politeness, and she treats every lead equally (she avoids the human fallibility of pre-judging leads).

This means that you get coverage of ALL of the leads that you generate, none will fall through the cracks anymore, and an equal chance to get a conversation started for each lead – which ultimately means higher conversion rate.

whatsnext-conversica-interview.png

Steven: AI technology therefore approaches these potential leads in a human way and then brings the right person within your organization into contact with them. In terms of marketing, it is of course a great advantage to have leads followed up on time and receive feedback quickly.

Amelia: Definitely, and sales gets much better leads so they can focus on having substantial, face to face conversations instead of spending hours just emailing and trying to get that lead engaged.

From AI Claire to Roger

Steven: Research agency ‘Markets and Markets’ estimates that the value of the AI sector between 2014-2020 will have increased from 420 million USD to 5.50 billion USD. How do you see it?

Amelia: There is a huge market for it. ‘Demandbase’ researched that 80% of top managers believe AI solutions will improve employee productivity and create new jobs, and simply for the specific business problem we’re trying to solve there is a huge need as well.

We’re focused on making business conversations more efficient, and taking the leg work of written conversation off of human teams and letting an AI do that grunge work – which makes a lot of sense because, according to a 2015 report by Radicati, business people send over 41 trillion emails a year. That in itself is incredible, and we all know from personal experience that most of those emails are mundane, back and forth emails trying to get ahold of people or set up meetings or simply get a real conversation started.

So yes, there is a huge need for conversational AI and we’re seeing the beginning stages of solving that need as the Conversica AI assistants are already used by more than 1,200 companies worldwide.

laptop on table
Foto door rawpixel.com op Pexels.com

Steven: On what basis do these Conversica (AI) sales assistants work?

Amelia: The Conversica AI Assistants are built on a foundation of natural language processing and understanding (NLP/NLU), which has to be very refined so that the assistant understands everything a human writes back to her, but it doesn’t stop there. A lot of conversational AI tools have decent NLP but then the AI doesn’t know what to do with the information it understands.

Our AI has a second component called an Inference Engine, which actually allows the AI to make decisions about what to do with the information it understands, without a human telling it what to do – so it can act autonomously (maybe decide when to send an email, decide to send our sales rep an alert internally, etc). And the third component is Natural Language Generation, which also has to be highly refined because all of the emails that Rachel sends to leads have to look completely, and naturally, human.

Rachel

Steven: For example?

Amelia: So Rachel always starts a conversation with a simple, short message introducing herself and then asking a question – she always wants to ask a question in order to elicit a response from that lead. She’ll keep emailing, just checking in with that lead, until she gets a response from them.

When Rachel receives a response of a like “Hi, No way I was ignoring you, was just busy – would be interesting to see a demo”, Rachel’s high-level Natural Language Processing allows her to not wrongly focus on “No way”, “I was ignoring you” and “busy”, but to understand the message correctly (that this person is interested in moving forward) and pass it on to the account manager.

“The combination of NLP, an Inference Engine, NLG and ML that Conversica uses is impressive. Among other things, the ability to interpret language, act autonomously, and respond with human-like messages is what makes AI useful or useless”.

Steven: Employment is a sensitive matter when we talk about AI and other innovations. I read several use cases from Conversica and see that one does not exclude the other.

Amelia: Absolutely, and this is a common concern, especially for sales teams, when considering adopting an AI tool like Conversica. However, what we end up seeing is that companies who “hire” an AI Assistant end up having happier sales teams because those sales people can now focus on more interesting, creative, and substantial human work, and they give the grunge work to the AI (all those email follow ups that they don’t have time to send!).

Conversica by WhatsNext

Ultimately, the Conversica AI Assistants are meant to augment a human team, and simply do the leg work that clutters up a human’s work day so that those humans can focus on doing more and better work. And because this means that a sales rep can now do more meetings, events, phone calls, and close more deals, your human job security is better protected.

Steven: Although AI will change sales and marketing, there will always be a need for ‘humanity’ and ‘personality’. New technology can be integrated to such an extent that it is experienced as an regular colleague.

Amelia: Exactly – in fact, the fourth, and key, component that builds our core AI is human intelligence – every AI tool out there relies on human intelligence as a model and a teacher. For example, our core AI went live in 2010, so it’s very mature for an AI. Over those eight years, every human being that Rachel interacts with, every email exchange she has with someone, is a learning moment. The AI has now processed over 317 million messages, and because she learns from every one of those messages, her artificial intelligence increases from the plethora of touch points she has with human intelligence. The incredible thing about an AI like Conversica is that it continues to learn.

IMG_5790
The virtual assistant Billie (Bol.com)

Steven: I can understand that, for example when I chat with Billie from Bol.com, he helps me a lot and I notice that I often use the same politeness as with a human being, while the chatbot doesn’t care less if or when I’m blunt.

Chatbots like Billie say they are a bot, but on some websites it’s not so clear and customers don’t know anything about it. There are already cases in the United States where customers declared their love to the ‘person’ they thought they were having conversations with.

Amelia: Definitely! Creating a human-like entity is key for creating smooth, and efficient conversations at Conversica. However, Conversica is not a chatbot, and in fact is quite different than a chatbot.

Chatbots can be very effective at quickly engaging someone who comes to your website, so, as a company, if you receive a high volume of web traffic, a chatbot might be a great lead generation or lead collection tool to get contact info of all of the people visiting the site.

But a chatbot is limited in scope and ability: it is anchored to a specific location (a website or certain landing page), so if someone doesn’t visit the website, they will never engage in a conversation with that tool. Additionally, a chatbot has a single-instance memory, meaning if you leave the website and then reopen that page in a few minutes, the chatbot will not remember who you are – that conversation you had with them is within a single instance in time.

Both chatbots and Conversica AI Assistants are based on conversational AI, but Conversica is actually a conversational AI platform – the AI Assistants are actual entities who are capable of having email conversations across months, if it takes that long to finish a conversation with a lead (maybe they’re out of office for vacation, then they say “hey, check back in with me in a few months, I’m finishing up a few projects right now”). But yes, Conversica AI Assistants definitely aim to have a human persona.

AI settings

Steven: Are there names that offer better results than others?

Amelia: Yes – and based on our clients and internal tests, we see that female names receive higher rates of responses from leads. But this is entirely up to the client – you’re “hiring” a virtual employee so you can decide who would be the best fit for your company.

We actually had a client that wanted their AI Assistant to be a 50 year old man, so they looked up names from 50 years ago and picked one out. The name can be very important per industry and per company.

WhatsNext Conversica interview 2

Steven: #WhatsNext is an independent blog that focuses on technology and innovation. We cover a very broad range of sectors, but place a special focus on innovations in banking and insurance. Can you tell us how Conversica can make a difference in this area? Practical examples?

Amelia: Certainly. Take for example Vern Fonk Insurance, an independent insurance broker from Washington state. The challenge there was to generate more internet leads, without impacting the number of employees in the company.

Vern Fonk chose to “hire” a Conversica AI Assistant named ‘Ava’, who worked on 11,780 leads in four months and assisted in the sale of 874 policies. She was even able to pick up old – seemingly dead – leads, or leads that had fallen through the cracks, and have them still pay off because of AI-driven consistency, persistence, politeness, and equal treatment of leads.

Steven: So Ava started following up with all of those old leads, or did she do more than that?

Amelia: Much more – thanks to those infinite-capabilities! She was still following up with all of their old leads, but they also assigned her to other lead groups. Initially, leads that arrive during opening hours were handled by Vern Fonk’s Call Center by telephone, and the internet leads were given Conversica as a first point of contact. But they realized that lots of leads would come in after hours, so even after the business closed in the evenings,

Ava was able to have email conversations with them because she worked 24/7. Or think of Spring Venture Group, another leading independent insurance broker from Kansas. Their Conversica AI Assistant increased the efficiency of the sales team and reduced the cost per lead.

Steven: I have indeed read these use cases and found that even seemingly lost leads can still achieve results, and that the cost per sale or cost per lead can be drastically reduced. In the case of Spring Venture Group, that was – after eight months – 30% less expensive per lead than before. Any other example?

Amelia: Absolutely. Another case is Assure Funding, a provider of loans and capital advances to businesses. Their goal was to detect prospects for financing in a cost-efficient way, regenerate previous leads and follow up with leads purchased from third parties. In addition, they wanted to up-sell other services and revitalize existing customer contacts, so a multi-faceted challenge that they approached using conversational AI.

Steven: And the result?

Amelia: Well, the Conversica AI Assistant that they “hired” as a virtual member of their team assisted in the 62% increase in financing requests, of which 42% deals driven by Conversica, ultimately giving them a 30% increase in credit renewals and strong cost efficiency.

Steven: Good results. Conversica already has more than 1,200 companies using your services. How do you see further growth?

Amelia: Two-fold. We continue to see growth across verticals – simply because conversational AI is applicable within diverse corporate and organizational environments – if you have a healthy lead volume or lead generation program, you’ll start to face the challenges of thorough and effective lead follow up – so cross-vertical adoption of Conversica will be a huge portion of future growth.

Additionally, Conversica is growing internationally as well – particularly within the EMEA market. We actually just opened our first EMEA office, based in London, with John Pincott leading the way as Conversica General Manager of Europe. With that new office and team, along with our strong network of partners who have the knowledge, experience, and wisdom within this particular market, we predict significant growth in EMEA.

In the UK our sales go through John, but in continental Europe we call on specialized parties who have knowledge of the business, are present on the field and therefore also take language and legal aspects into account. For the BeNeLux we work primarily with our longtime partner, Leadfabric.

Steven: Thank you for this talk and good luck.

Amelia: Thank you!

NL versie

Steven Dupont 

Want to stay informed about #WhatsNext? Follow us on:

Geef een reactie

Vul je gegevens in of klik op een icoon om in te loggen.

WordPress.com logo

Je reageert onder je WordPress.com account. Log uit /  Bijwerken )

Google+ photo

Je reageert onder je Google+ account. Log uit /  Bijwerken )

Twitter-afbeelding

Je reageert onder je Twitter account. Log uit /  Bijwerken )

Facebook foto

Je reageert onder je Facebook account. Log uit /  Bijwerken )

Verbinden met %s