Jonathan Corbin, is the Founder & CEO of Maven AGI. Previously, as the Global Vice President of Customer Success & Strategy at HubSpot, Jonathan led a team of approximately 1,000 customer success, partner success, and contract managers across multiple regions and verticals. His responsibilities included driving customer retention, revenue growth, and value realization for over 200,000 customers worldwide, ranging from startups to enterprises.
Maven AGI is a comprehensive Generative AI native solution designed to transform the customer support landscape – without the headache. While in stealth mode, Maven’s technology autonomously resolved over 93% of customer inquiries, cutting support costs by 81%, enhancing the overall customer experience, at scale, after resolving millions of interactions in over 50 languages for early customers.
You were previously the global Vice President of Customer Success & Strategy at HubSpot, where you led a team of about 1,000 customer success, partner success, and contract managers across multiple regions and verticals. What were some highlights and key takeaways from this period in your life?
During that period of time, Hubspot was one of the five fastest-growing B2B SaaS companies with over a billion dollars in revenue. There are very few people who have had the opportunity to build, grow, and manage at the scale that we were operating at. Companies that grow at this speed aren’t usually that size, and companies our size didn’t grow at that speed. I spent a lot of time focusing on creating scalable approaches to planning and growth, making sure that we were setting very clear objectives, aligning incentives across multiple organizations to create the outcomes that we were looking for as an organization, ensuring we had the systems to create visibility to what was happening in the organization, and planning over multiple horizons. Anything that we rolled out had to work not just for our current customers but had to have the ability to maintain continuity at exponential growth.
Can you share some insights on what inspired you to launch Maven AGI, and how long you have been in stealth mode?
I’ve been obsessed with customer experience since very early on in my career and that’s why I’ve spent so much time at industry-leading companies in this space (Adobe, Marketo, Sprinklr, Hubspot, etc). Back in 2017, I was coming back from a West Coast swing, meeting some great customers like Apple and Nike, and we had these incredibly in-depth conversations about the potential to unlock siloed data and create these very personalized experiences down to the individual user level. I’m not talking about the segmented approach of you falling into this age category or demographic. No, this is the ability to fully deploy all the information that you have shared with us to anticipate customer expectations and proactively engage with them. There was massive excitement from the customers but the technology didn’t really exist at the time.
My co-founders – Sami Shalabi, Eugene Mann, and I have always chatted about personalization at scale and the potential that transformers could have since the research first came out of Google. Sami built one of the largest personalization engines in the world at Google News (1B+ users) and Eugene led personalization for it so we’ve always had deep, insightful conversations about the possibilities that we could unlock as technology evolved. The application of this to what we were doing at the time is that I was struggling with being able to create a great experience at scale for our Hubspot users, Eugene was looking at how to productize LLM capabilities at Stripe, and Sami was sharing his insights on what worked well at Google.
When we first heard about what OpenAI was doing and started using some of the LLMs that had become available, we realized that we were at the point where the technology now existed for us to create the perfect customer experience at scale. Companies have had to choose between cost efficiencies and good customer experience resulting in all kinds of things like complex segmentation strategies designed to limit customer interactions, creating things that are essentially roadblocks that they called self-serve, or burying your support contact information somewhere that it can’t be found.
We started Maven AGI about a year ago in stealth mode because what we prioritize at Maven is impact – and when we announced what we were doing we wanted to give real examples of our impact and metrics, not just that we existed and had raised some money. We’re incredibly grateful for our early customers who believed in us enough to work with us in rolling out cutting-edge technology and pushing the limits to develop a better customer experience.
Can you define for us what AGI is in the context of Maven AGI?
AGI is really well defined from a language perspective – it’s artificial general intelligence. What does that actually mean in the business sense? We’re focusing on something that we’re calling business AGI and define it as the ability to handle complex tasks using functional AI agents that are specially trained for specific responsibilities with an orchestration layer that allows them to work together.
An example of this might be a bank account user engaging with their bank and asking if their deposit has cleared – what we know from account history is that they need a small bridge loan to to gap their bills and check cashing. Maven will understand the historical context and offer the loan while handling all of the paperwork that might be associated with it such as background checks, credit checks, filling in loan paperwork, understanding the risks, approval, and a specific amount that falls within the risk profile, approving the loan, and moving the money to the person’s account.
Another example would be someone going to their CRM support team and asking how to deploy a campaign. What we would understand from that is they don’t want to know how to create a campaign, but they want a certain number of leads by a certain date. Users would have the ability to say, “Give me 100 leads next month” and Maven would go through the incredibly complex task of delivering those.
What are some of the biggest problems with how AI has historically been integrated in customer support?
Historically, AI in customer support used machine learning models that were highly deterministic and took months to train. These models worked on a basic if-then logic: if a user chose X, they would be given the Y option. This simplistic approach fell short of expectations, resulting in disappointing outcomes and leaving many CX professionals skeptical of AI’s potential. True success in AI-driven customer support hinges on dynamic personalization, the ability to reason, and take meaningful actions.
What are the key steps involved in training Maven AGI to handle customer support inquiries?
It’s really simple. . . just give us access to any information that you would use to train humans on. We can have it up and running for you with a high degree of accuracy within days– not weeks or months. It will use your specific tone of voice, vernacular, and whatever emojis you want.
How does Maven AGI help in reducing customer support costs and improving overall customer satisfaction?
Companies deploy Maven AGI in a variety of different fashions but the best way to have the fastest impact is to insert Maven at the head of your support queue at the endpoints or channels that your customers want to use (chat, web, search, Slack, in product, SMS, etc). That allows us to provide instant, personalized results + actions to customers with no wait time while ensuring that those amazing support agents are doing what they do best, working with customers who really need human interactions to solve their problems.
What technological advancements have enabled Maven AGI to achieve such high rates of autonomous issue resolution?
I believe we have recruited one of the best engineering teams in the world to solve that comes down to a data problem. Brilliant folks who have worked on challenges like search at Google, and personalization at scale at Meta and Amazon, and have been thinking about solving these sorts of problems for years. Data is fragmented and siloed, and in order for us to answer customers’ questions and take actions we needed to be able to ingest more data than anyone else. The second part is the ability to take actions and build our action engine because we know that just answering questions isn’t enough. In order for us to achieve business AGI we need to be able to anticipate users’ needs and engage them with intention.
Can you provide more details about the recent $20M Series A funding and how it will be utilized?
We were fortunate to be hitting on all cylinders in what we wanted to achieve with our seed round: build a great engineering team, a product that solves real problems, and have customers who were getting value out of our product. We raised our seed round less than a year ago but had some really great investors who wanted to be part of the journey with us. After spending time with M13 we were really excited to continue to build the future of Maven AGI together with them. The $28M that we’ve raised over the last year will be used to build out our GTM team, invest in building out the partner ecosystem, and continue to hire engineers as we expand our action engine () and platform capabilities.
How do you see the role of AI evolving in the customer support industry over the next five years?
The future won’t be divided into support, services, sales, and various functions. Instead, customer support will become part of a seamless, unified customer experience without messy handoffs and siloed data. As customer expectations evolve, so will the ways we serve them.
Today’s customers needs fall into 3 categories:
- Those who want to self-serve – the ability to find the solution or answer to a question.
- Those who want access to self-service but need validation that they’re taking the correct action.
- Customers who demand white glove service and need human assistance.
The future also has 3 categories but expectations from customers will be far different:
- Expecting instant answers to their questions.
- Anticipate their needs and questions with personalisation, usage data, full historic context, and the ability to take action and engage with them at the channel of their choosing.
- The ability to engage with customer support agents without wait times and lengthy lines, who have answers available to their questions, full historic context, and the ability to instantly take actions.
Thank you for the great interview, readers who wish to learn more should visit Maven AGI.
Credit: Source link
