Startup Spotlight #11: Connect and Conquer
Today's startup has proved that to succeed in the age of AI, sometimes all you need is... A little human touch.
1. Tons of startups are already building AI bots for online stores. Yet, live sales consultants still outperform them in driving sales. But what if chatbots could learn from humans — not just before launch but continuously?
2. This leads us to the idea that the ideal solution isn’t “all human” or “all AI.” It’s human-machine platforms where people and AI seamlessly complement each other. Today’s featured startup used this approach to enhance e-commerce sales — and raised $10 million in its first funding round.
3. But this powerful concept could be applied to creating custom platforms across many different industries!
Project Overview
In physical retail stores, sales assistants play a key role by answering customer questions or approaching them with offers to help, such as “what are you looking for?” or “how can I assist you?”. Although it can sometimes feel intrusive, these interactions often boost sales, which is why stores invest in hiring such assistants.
Remark proposes bringing this approach to e-commerce by connecting online store visitors with live experts who help customers choose the right product.
The Remark platform begins by asking visitors a series of questions to understand their needs and challenges in finding or choosing a product. Based on their responses, the platform connects them with a topic-specific expert available online. Through live chat, the expert answers questions, provides tailored advice, and suggests suitable products available in the store.
The platform also records each conversation, extracts key information, and automatically creates landing pages that assist other customers with similar queries. These pages are optimized for search engines since they reflect real-life questions people often search for online.
Remark’s experts are professionals in various fields, making them knowledgeable about the products they discuss. For example, clients include sports stores that employ alpine climbing instructors, yoga trainers, skaters, or even U.S. ski team members as their experts.
Involving experts significantly increases conversion rates and reduces product returns. Unlike the 25% return rate typical for online shopping, brick-and-mortar stores see only 3%, largely due to better-informed purchases.
Experts also boost the average order value (AOV) by recommending high-quality, often higher-priced items and suggesting complementary products that customers might not have considered on their own.
As a result, Remark’s platform has helped its clients achieve a conversion rate of up to 20% and a 9% increase in revenue. The average response time for connecting a customer to an expert is just 12 seconds — well within the acceptable window for “hot” leads.
Since its launch two years ago, Remark has partnered with 45 online stores, onboarded 50,000 experts, and facilitated 100,000 customer interactions. The startup recently raised $10.3 million in its first funding round.
What’s the Gist?
“Shoppers who chat buy more”, the startup claims — and the data supports this assertion.
Engaging in a conversation acts as a qualifying step in the sales funnel. Shoppers who opt out of such interactions are less likely to convert, saving resources for those who are more interested.
This insight has spurred the rise of “conversational commerce,” where online retailers use chatbots to handle customer queries. Recently, these bots have started leveraging AI algorithms. However, chatbots often lack the nuance to recognize emotions or adapt their tone to match the customer’s mood. Most shoppers aren’t just looking for factual answers —t hey value the personal experiences and recommendations of a real consultant.
As a result, chatbots frequently frustrate users and have limited success in driving sales. In contrast, human consultants deliver significantly better results, albeit at a higher cost.
Remark aims to combine the best of both worlds. It’s developing AI bots trained on successful expert conversations, mimicking not just the arguments and examples but also the communication style. The goal is to create digital twins of real people rather than generalized chatbots.
Interactions on Remark’s platform are designed to transition seamlessly between AI agents and live experts. When the AI identifies that it’s not advancing the sale effectively, a human expert steps in.
However, this raises questions about compensating experts. Without AI, an expert earns commissions for sales generated through their advice. But if an AI agent trained on expert data closes the sale, should the expert still receive a commission? If so, how should it be divided among all contributors to the AI’s training? Over time, as more experts contribute, individual shares would inevitably shrink.
Alternatively, could each AI agent function as the exact digital twin of a single expert, allowing online store owners to hire or dismiss specific “experts”? This approach has some intriguing possibilities.
Resolving these questions will require industry-wide standards.
Remark isn’t alone in exploring ways to integrate human experts into online sales and customer support funnels.
• Experify connects potential buyers with customers who have already purchased similar products, allowing them to share personal opinions. Since its launch, Experify has added a feature to generate reviews from these conversations for publishing on review sites, much like Remark’s auto-generated landing pages. Experify has raised $4.03 million in funding.
• Limitless offers a tech support platform that connects experienced users with newcomers, using intelligent routing algorithms to match questions with the best-equipped responders. Limitless has raised $20.5 million in funding.
Key Takeaways
If we’re going to simplify things, Remark’s model can be roughly described as an “Uber for customer support”.
Just as AI is expected to replace drivers in ridesharing services eventually, AI agents in these support platforms may complement or replace human experts after sufficient training on real interactions. But as with ridesharing, the journey starts with humans, who currently outperform algorithms in these roles.
This points to the development of hybrid human-machine platforms for customer and user support. The necessary technologies and workflows for such platforms are already taking shape, though the specifics of compensating live experts for training AI agents remain unresolved.
Such platforms could find applications beyond e-commerce and customer support. Imagine a programming assistant that handles most queries with AI but quickly connects users to live experts for complex issues.
The potential for human-machine platforms spans many industries. What other areas do you see benefiting from this approach? Or do you believe AI will soon solve all these challenges on its own?
Company info
Remark
Website: https://www.withremark.com/
Last funding round: $10.3 million, 23.05.2024
Total funds raised: $10.3 million over 1 round