How to Clone Your Best Sales Reps with AI from Drift
Drift is a conversational marketing and sales platform with $107 million in funding. The company streamlines how marketing and sales happen using conversations across chat, email, and video.
Artificial intelligence is a key ingredient in the company's solutions. Drift Automation uses AI to mine past conversations with buyers, so your on-site chatbots can answer buyer questions. Machine learning also helps these bots automatically qualify leads 24/7. That means marketers and sales pros free themselves up to pursue higher-value tasks.
We spoke with Jeff Orkin, Ph.D., head of the Drift AI Lab, to learn more about how the company uses AI.
In a single sentence or statement, describe your company.
Drift helps businesses use conversations to remove friction from their buying process with chat, email, video, and automation products.
How does your company use artificial intelligence in its products?
In June, Drift launched Drift Automation — which ingests chat transcripts from a Drift customer organization and mines a conversational model used to power chatbots. From there, we use machine learning and natural language processing to learn this conversational model.
The model is comprised of three primary parts — an ontology of domain specific words and phrases, a speech act classifier that can group utterances by semantic similarity, and learned flow sequences of utterances.
Through all of this, instead of just being able to ask questions, the bot can answer them too. It is also trained on what a good lead looks like for a sales team. The team then reviews the bot with the customer, perfecting its answers — and is then tested with a small audience on a website. It scales indefinitely, so the more it works, the more it learns and the better it gets. By utilizing Drift Automation, a customer’s buyer is able to get their questions answered 24/7/365 by a bot that the customer didn’t have to build or optimize, and it is always getting smarter.
What are the primary marketing use cases for your AI-powered solutions?
Drift’s Qualification Bot acts as an automated SDR (sales development representative) that is available 24/7 to qualify leads, while being flexible to answering freeform questions from the user. The bot is both proactive and reactive and will proactively pursue an organization’s lead qualification protocol, while reactively answering any question that the user interrupts with.
The bot is trained by Drift’s Annotation Team to respond using language that a customer’s best sales development representative would use. One customer, Keap, enabled Drift Automation and saw a 50% sales pipeline increase in just 50 days, without having to hire new people or build out new playbooks.
What makes your AI-powered solution smarter than traditional approaches and products?
With Qualification Bot, 100% of a customer’s leads are engaged with in real-time, it performs like the best SDR, it grows as traffic and lead volume grows, it’s deployed as a full-service solution, it’s continually learning and optimizing itself, and it uses AI technology so marketing doesn’t have to spend time creating and optimizing workflows.
Are there any minimum requirements for marketers to get value out of your AI-powered technology? (e.g. data, list size, etc.)
Drift Automation can start working on as few as 50 conversations.
Who are your ideal customers in terms of company size and industries?
We serve companies in various industries from 100 to 10,000 employees.
What do you see as the limitations of AI as it exists today?
Many of the existing solutions for automating conversations either use decision trees or intent based classification. These approaches lack the flexibility to context switch, and struggle with interactions that go beyond answering one question at a time. By learning a model at both the utterance level, and the conversational sequence level, it is possible to overcome these limitations, and maintain a coherent conversation that can handle context changes, and could continue for several minutes.
What do you see as the future potential of AI in marketing?
There are so many functions of marketing that are conversational, yet are also repetitive — e.g. lead qualification, webinar invitation and scheduling, event answering emails. AI has the potential to take the burden off marketers for these tasks, freeing them to do higher level tasks.
Any other thoughts on AI in marketing, or advice for marketers who are just starting with AI?
When implementing AI, look for solutions that augment your team, rather than black boxes that promise to automate away the need for humans. It’s important to remember that AI is not a replacement for good engagement, and today brands are winning and losing based on the quality of their brand as defined by the customer experience they deliver.
About Paul Roetzer
Paul Roetzer (@paulroetzer) is founder and CEO of PR 20/20, author of The Marketing Performance Blueprint and The Marketing Agency Blueprint, and creator of The Marketing Artificial Intelligence Institute and Marketing Score. Full bio.