S4 E6: Featuring Dilip Dand

Tracie:

Hello everyone and welcome to Traceability podcast. I'm so excited today to have my friend and former boss and mentor, Dilip Dand with us today. Super excited to hear from him and all the things he has going on. Dilip is an accomplished entrepreneur and visionary at the forefront of AI and digital transformation. He is the founder of Lighthouse3, which is a pioneering software consulting business based in the San Francisco Bay Area. Dilip has dedicated over 20 years to empowering organizations with cutting-edge technology solutions, and Lighthouse3 has a reputation for delivering complex and innovative solutions such as AI-driven chatbots, and voicebots for many industries including automotive, retail, and FinTech. Dilip has always had a relentless drive to push the boundaries of what technology can achieve. This has led to the development of groundbreaking AI systems, including IoT and AI-based gas leak detection systems and a multi-agent inventory management system. In addition to Lighthouse3, Dilip is focused on advancing the field of AI performance management and I am excited to talk to him about those things today. So Dilip, thank you so much for being here today. I'm so happy.

Dilip:

Thank you Tracie, and I'm excited to be here. It's been a while and it was fortunate just to meet you at TechCrunch Disrupt, so I'm really looking forward to our conversation and connecting again with you is pretty good. So thank you for having me.

Tracie:

To catch everybody up, I was at a conference in the Bay Area a couple of weeks ago and happened to look up and there was Dilip. And we hadn't seen each other for a few years, but it was fortuitous and gave us a chance to reconnect and also make time to have Dilip on a podcast with us. So fun little story there. So Dilip, when we typically begin the conversation, I like to sort of walk through the career journey that sort of led to your... I know you had a long career in FinTech and contact center operations and now you've transitioned into entrepreneurship and AI, so I wanted to sort of get some background on how you have arrived here.

Dilip:

Thanks, Tracie. My background in both AI and entrepreneurship has been a long one. I started AI back when it was still called expert systems and neural networks, big in the early nineties, and from there I've been watching this field, but it was pretty dead for quite a few years, a couple of decades actually really, before machine learning started picking up and then slowly conversational AI came in and not generative AI had given us a lot of opportunities. So I've been always engaged with this field, but I just was observing to see what comes out of it because it's taken so long to get here. Now that it's here, it was a perfect time for me to sort of leverage my experiences and my knowledge from the past to help companies successfully implement AI. As for my entrepreneurial aspects, I come from a family of entrepreneurs. My parents have businesses, so does my brother, and so it was sort of in my blood. And so when I came to US in '95, of course I couldn't start my company right away, but as soon as I could I launched what has now become Lighthouse3. So it's been around for 20 plus years. We've been helping clients, as you mentioned, build out some of these technologies and implement solutions on the cutting edge, basically help leverage innovation and technology to solve business problems. That's our focus. We don't do innovation or technology just for its sake. We try to solve business problems and really measure and manage those. So that's been my journey with AI and entrepreneurship.

Tracie:

I love the history there. I was not aware of some of the history, that the family background and the years that you've been engaged in entrepreneurship. I am interested to know what led you into the corporate roles and how that has informed your entrepreneurship.

Dilip:

Yeah. I mean, corporate, as I said, I came to the US in '95, didn't have any family here, so I had to build my life here. So that's what led me into the corporate world. And all those years, I've been fortunate enough to work with smart people like you and others in the companies that I worked with and to be at the leading edge of some of these things that has happened. So there are three things basically that I've learned from my corporate life that I think have helped me with my entrepreneurship. First one is leadership basically, and corporate life helped me figure out how to lead a team, how to set goals, really manage and bring together and motivate teams successfully. And so that's the first thing. The second thing is really the cross-functional team collaboration, because without that, it becomes hard. And especially in entrepreneurship, you have to do so many things to start a company and to run a company that you need experts to help you. And so bringing all of those right experts to help me with building this company, continuing it on for so long, clearly that has helped me build this company to where it's now. And the third thing I would say is the data-centric and customer-centric mindset that's needed to be successful. A lot of these successful companies like Visa, like JP Morgan Chase or some of the other companies that I worked with, they have heavy focus on customers, making sure the customers are taken care of and then measuring and managing the performance based on data and making data-based decisions. So all of those I've tried to incorporate into my entrepreneurial journey and make sure we as a company are performing, making sure our customers are taken care of first of all, and then we are measuring ourselves and making sure we are constantly delivering to our goals and we are exceeding our goals.

Tracie:

You make such a great point. I think that as we go through our work lives and in the corporate roles that we have, we don't necessarily understand that we're in sort of a training ground of how to work with different types of people and how to identify experts and how to network within organizations and such. So I do really appreciate what you said there as far as learnings from those roles because certainly that's something that you take with you into entrepreneurship. You have to become even better at the networking and the cross-functional work and that kind of thing.

Dilip:

You essentially have to be a jack-of-all-trades and in order to be that, you need to know what your strengths and limitations are and then sort of bring in the right help you need.

Tracie:

How we connected originally was working in the contact center operations space, and there is much that is going on in that domain with AI and so I wanted to get your take on how that space is going to change over the next several years, especially in regards to AI.

Dilip:

Absolutely. I mean, AI is going to help first customer service as well as contact centers change drastically over the next 5 to 10 years. Now before AI, what we used to do with our customers was either we put on a simple structured chatbot that could only understand a few phrases and we would use RPA for efficiency purposes to help automate some of the routine tasks. Now, those were systems that were not very customer-friendly and they were helping improve some efficiency, but they were very rigid so they were not very easy, flexible to adapt to changing environments. With AI, now we have that ability to provide customers with a very personalized experience that brings in the real-time context of what's happening in the customer's journey with the company, as well as building out efficient solution that even the agents have the right information at the right time in front of them and allows them to focus on solving the customer's needs while not having to worry about the mundane task of entering the list ticket in the CRM or making sure the notes are properly captured. That's where AI can really help and I see that happening over the next 5, 10 years where we'll see sort of a hybrid model of AI and humans working hand-in-hand to become smarter and providing customer service that helps companies build their brand loyalty as well as to help them stand this apart from their competitors. Ultimately, I feel this hybrid model of AI and human touch will go hand-in-hand and will be a tremendous improvement and boost to customer service and contact centers.

Tracie:

Much of the conversations with AI these days is it makes people pretty nervous for lots of reasons, and I think that nervousness maybe leads us to underestimate that we still have a role and that we won't be replaced and that kind of thing. So do you want to speak to that a little bit?

Dilip:

Yeah, absolutely. Recently I was working with a retail customer, helping them build out their AI virtual assistant and that was a very real issue that we saw with the contact centers' stakeholders. So we had to really educate them and say, "This is not an opportunity to where we are reducing stuff. What we are trying to do is help you make more efficient, take away the mundane," really educate them about what the AI can do, where their focus should be and how they would be adding value to the business. Once we got them on board and helped them educate and understand what was the issue, then implementing some of these things became much, much easier and helped really drive the benefits. And we'll talk a little bit about that as we go along, some of the benefits that we saw with these particular customers. So yeah, it is a big concern that we're seeing with our clients. That human change management is a big piece of it.

Tracie:

As companies are looking to get involved in AI, and everybody is looking to do that right now and they're sort of a frenzy with it to identify use cases that make AI useful, are there some best practices and tools that you bring to companies that you consult with for things to consider with implementation?

Dilip:

Yeah, what I advise companies that are trying to get into the AI game is to really start small and make sure they have a very clearly defined goal of what they want to achieve. It does not have to be AI for AI's sake because your competitor is [inaudible 00:13:43]. It has to be targeted towards solving a particular business problem. It doesn't have to be customer service, it could be marketing, it can be in your customer journey and the personalization along that front, or it could be also in your back-end office services, like in retail it might be to do with your fulfillment area. So whatever project you decide, just make sure you have a clear vision for it. From there, I would say make sure you engage all the different stakeholders, and there are tons of different stakeholders. It's not just the immediate adjacent stakeholders that you have to that specific business problem that you're addressing, but also things like legal because now there's a whole different question of ethics and bias that we can talk about that really make this a different conversation than what we are typically used to having when we implement traditional software solutions. The third thing is I would also make sure that you will have to upskill and reskill some of the staff that are going to be involved in it, because there are some new things that you will end up having to do as part of this AI implementation that you've never done before. So for example, one of the things when we were implementing the chatbot I talked about, there was more active participation from content manager. Now traditionally, companies have had content manager but they've been project-based. So we were implementing a website. If [inaudible 00:15:17] went in there, wrote up the content, we'd move on. But with AI, it is not the same. You have to constantly tweak and make sure the content is up-to-date. And so for that, you need a dedicated contact manager that can help you continuously look at what's resonating with the customers, how to say something differently so that you can make your point more effectively. And that's a full-time job now as you implement. So there's some new skills you'll have to adapt and adopt, so be prepared for that. And then lastly, measure and do test tools because without measurement, you're not going to be able to know whether you're making progress towards your goal or should you be continuing to invest in this particular technology for the specific use case. So those are the big things that we advise customers as we build the solution.

Tracie:

So you touched on a couple of things. One was ethics, which I know is also a big concern in AI projects right now, the speed of the AI bots learning and that kind of thing. So how do you really approach ethics with AI in organizations?

Dilip:

Yeah. I mean, if you're starting small, make sure whoever your sponsor and your key business stakeholders are, are aware of what ethics mean in this context and AI ethics, and make sure they are accountable and responsible for that. What that means is you need to make sure your system AI solutions are being developed and tested for being not biased against certain race or gender or groups of people, but then also making sure that you're using the data responsibility, that the data is still being kept private, privacy is still being adhered to. And finally, I would say from an ethical use perspective, making sure not just your own handling and solutions that you're building, but if you are partnering with vendors that are helping you either build solutions or even if you're a SaaS platform that has AI in it, make sure those vendors have ethical processes in place and they have a way to monitor and manage ethics and bias issues. Now, one thing that is where I'm focusing my time right now on as it ties to ethics is that we talked, you mentioned in my introduction the performance management of AI agents and AI [inaudible 00:18:10] and that's one of the big areas that we are focusing on and tying ethics into that. So what we are doing is we are building a solution that help clients monitor their AI interactions that are happening either at the customer or even with their employee and ensure that not only is the response from AI relevant and accurate, because as we know AI still tends to hallucinate. Even though we try our best, there's still some hallucination. And thirdly, we also measure fluency, like conversation is moving forward. And as part of that, we are looking at ethics as well. So what we do is we flag any conversations that we think is biased or unethical so then somebody in the customer service department can, for example, can go in and look at those conversations and say, "Oh is that is really an issue with the system, something needs to be tweaked or changed," so you have eyes on it and constantly you have a human in the loop for managing ethics. So that's where what we've been helping our clients adopt ethical and responsible AI in their practice.

Tracie:

Yeah, I love that. Sometimes we don't think about our vendors and partners and the influence that they are having on our organizations, and so I appreciate that there's lots of checks and balances when it comes to AI implementation and that kind of thing.

Dilip:

But it has to be an ongoing effort, right? Because you cannot just say, "Okay, now that I have it in production, I'm done. I don't need to worry about ethics." And it's an ongoing thing because human bias, I mean, we all are biased by nature. It's not something that we can avoid. We are biased towards our own self, protecting ourselves. So we make sure our systems that we implement don't distort the bias towards it more significantly than it needs to.

Tracie:

So do you envision a time when really just about any role is influenced by AI? Any role in an organization.

Dilip:

Absolutely. Influenced? Most definitely. Every role from the CEO and the board on down to the last contact center person and even people, maybe facilities, will be impacted by AI in different ways. What I don't foresee is the need for humans to be still involved. We will still need that. I don't see that going away anytime soon. What I see is we leveraging AI more effectively and helping us focus on more value-add tasks that helps build trust with the business as well as build brand loyalty and improve the customer journey across the board. So really that's where I feel AI is going to help and influence all the roles. So think about things like even in facilities. You don't need a after office hours person to go around turning off lights [inaudible 00:21:36] with the smart lights and even programming how to manage the temperature in the building some of those mundane tasks. But you still need facility staff to make sure the area is secure, that the human interactions are properly taking place and we have what they need in the facility. So that's just one example. But on the customer service side, I don't foresee AI totally replacing humans at all. Recently there have been a lot of publicity around companies that have sort of laid off their customer service staff in favor of an AI bot. And to my mind, that company does not value their customers because if you really valued your customers, you would not make it harder for them to talk to somebody in your company if they have an issue. By eliminating a customer service department, obviously you impact on the customer's experience, but then from a business point of view, you also sacrifice quite a lot because now you've lost that human touch that you could have got from talking to your customers and finding out what are the actual pain points are, what are the features they're looking for, et cetera. So it's a very hard trade-off where I don't foresee AI agents completely replacing humans at all.

Tracie:

Interesting. So then as employees, do you have some tips for how we can dip our toes into AI and if not embrace it, but not be quite so stressed by it?

Dilip:

Absolutely. I think there are some very good news articles and publications that employees can actually use to learn about AI. Yesterday I saw an article from MIT which was talking about everything about just the basics of AI. It's an introduction to AI. I would encourage everybody to sort of read that so they get an idea of what the different terms and what are the different things that go on in AI. From there, I would say start interacting and using some of the tools that are now out or either integrated either with AI or are AI-based and play with it, and play with it both for your own personal stuff, but then also for small products that are not immediate impacting your customers or your immediate bottom line. So be very judicious, but at the same time expand. And I would encourage leaders, business leaders to encourage their teams to do the same thing because unless and until we all educate ourselves on how AI works, there will be a lot of disinformation. There will be misconceptions and myths and legends floating around what AI can do and that will just make it harder for people to just do it.

Tracie:

That makes a lot of sense. We want to be as educated and as knowledgeable as we can so that we don't end up in a situation where we're nervous or where we think something else is going on that's not, so that's a great point about dealing with truth and misinformation. AI is also a field that is just changing exponentially it feels like every day. Are there things that organizations need to do to stay up to date or is the AI sort of learning to keep itself up to date?

Dilip:

No, AI is not learning yet to keep itself up-to-date. It's definitely organizations have to be actively involved in learning that and making sure they are up-to-date on what it can do. A lot of the development that's happening right now is, if you think about it, it is from we got the tools, we got the technology going, we started building tools. Thinking back to when internet came up and cloud computing was [inaudible 00:25:57], how did that begin? We first started slowly. The technology team started adopting cloud technology for one or two use cases, and then over time, business use cases went onto cloud as well. And then once that was in there, then it became accepted and adopted across the board. Similarly, what we're seeing with AI is that we are in that state where the technology has been proven, there is now resources or technology available to use AI. What's happening now is we are seeing developers build out business use cases based on AI and new use cases that we had not thought possible before. They're slowly being integrated into their existing processes, existing tools and some new technology tools are coming up. All of that is happening now between, I would say, from now on to the next 18, 24 months is where we're going to see a lot of that happening. From there on, businesses will be using AI without a second thought because it will be part of their whatever tool that they're using or consciously make a choice to use a tool that is ultimately AI-based because it helps solve a business problem that they have been tackling for a while. So I feel like in order for that to happen, businesses need to keep themselves up to date. In the case of CRMs, what is Salesforce doing with [inaudible 00:27:33]? What is Zendesk doing on that side? What are some of the technology providers like [inaudible 00:27:41] or Ciscos or Avayas or others, Amazons and Google doing in that space? So there's a lot happening, but you need to keep yourself educated and be very cautious about what the vendors want to do and what they're promising because they will promise a lot. So you have to be educated in order to listen what is true and what is not.

Tracie:

So if we can backtrack just a little bit, so now you're a founder and you're out there pitching AI to different companies and that kind of thing. Can you talk about the founder's journey a little bit?

Dilip:

Yeah, sure. It's fascinating. My founder journey personally started for a consulting company. So that's we providing services. So Mike was trying to solve a business problem and trying to bring in the right team. So that's where my journey started. But a product founder, which is what I'm experiencing now and right now that's where I am because I'm building up a product, has been very interesting as well, because the first thing that, why we got the idea and we built some custom solutions for our clients, productizing that into a product is a different bucket. And for a startup founder, you really have to think about it from that perspective, is that you can be agile and nimble because you're small so you can build and test out a lot of things quickly and put features out there. But the big thing that we have to understand is the product market fit, making sure that your product has a demand for it and that customers are willing to pay for and that's where you end up spending a lot of that time that you need to learn about. And so in my case, I'm going through that right now where I'm trying to validate my product market fit. As I said, we build this solution as a custom solution first, we're productizing it so we're in a space where we want to do a product market fit and make sure there is customers that are ready. What we are discovering obviously is what we talked a little bit ahead just on a previous topic, where the customers and the business processes as well as the tools are not yet ready for what we're trying to do. So we are moving with them. As these tools are coming up, we are realizing that we need to pace ourselves and go hand-in-hand. So that's where we are focusing our effort on. And so I would encourage founders to keep trying new things because right now is a good time to do this. AI can help you speed up your path to product market fit faster. What used to take 6, 12 months to get there, now you can achieve that in less than six months. So anybody who's interested in that entrepreneur journey, it is the right time to do it and really get going. And if anybody needs help or advice, I'm happy to help [inaudible 00:30:57]

Tracie:

So one thing that I have noticed is as I've been going through a little bit of this process myself, is the need for agility and flexibility and moving quickly, but not in large ways, in small.

Dilip:

Absolutely. You have to be very targeted and focus on what you want to grow because as we said, being small, we can be agile and nimble and do several more things that we thought of, but is it really useful in the market? As product manager, I have had to say no to a lot of features because of that. You realize that maybe it's not yet time, maybe it's not what the market needs. So a lot of time, you end up taking time to say no because that you think are useful. So from that perspective, it's really a very interesting journey because as I said earlier, you end up spending time on so many different things as an entrepreneur that you need to make sure you are focusing your product on the right solution, not forgetting that vision because you tend to end up having... As you talk to more customers, you'll end up having more features, demands and you then try to chase those, which is not a good place to be.

Tracie:

Yeah, I've got so many ideas, lots of times, so many ways to expand and that kind of thing, but you sort of need to have a little stability just in one spot before you sort of branch out.

Dilip:

Exactly. Exactly. And you need to make sure customers are willing to pay for what you are doing in that space, in narrow niche. Once you get that, then you can think of branching out and building.

Tracie:

Are there things that you do to sort of build that demand, maybe customers don't know what it is that they need?

Dilip:

Yes. We're spending a lot of time educating the customers right now, especially in the space we are in, because customers are not used to measuring performance or AI or any sort of automated solution in the past. Nobody cared about whether an RPA solution was doing 10,000 things or one thing and whether it was doing it right or wrong. With AI, we have had to spend time educating the customers on that because it's a black box [inaudible 00:33:43] "thinking black box." You need to be treating them as your human workforce. They are actually out there representing the business and if you don't treat them like that, have set performance goals, make sure you measure them and then tweak them as you go along to make sure you're still working towards delivering your business goals, you are not going to have a successful AI solution and you're going to end up spending a lot more money, time, and effort that would have not showed the results that you're expecting. You need to treat them just like we do with human [inaudible 00:34:26]. You hire humans because of that. You motivate them and make sure they're still delivering what you need. Otherwise, same thing AI [inaudible 00:34:35]

Tracie:

Are there certain metrics that organizations should be looking for?

Dilip:

Yeah, so I mean, there are definitely the traditional metrics that we end up doing that most of the AI tools provide these days. But what we are finding is that as new use cases, especially as AI takes on new use cases, we have to define some new metrics and that's what we've done with our solution. We've identified least three new metrics that we are measuring AI on, and then there are the new metrics that come with ethical use cases like how many times did you end up violating your own policies. Not even regulations, but even your own policies and whether you are still tracking towards being an responsible AI solution or not. Where I see the other set of metrics that are coming in that probably doesn't impact it too much for companies that are in US just yet, but especially that are working in Europe and other parts of the world, is the regulations and regulatory compliance needs that are going to come in. And that will require things like measuring incidence of bias or incidence of breaking the regulatory guidelines or things like that, and then what action companies have taken to mitigate those. So those are some of the new metrics that we are seeing coming into this space.

Tracie:

Interesting. Very interesting. I never would've thought of some of those things, but it's a new world and-

Dilip:

Exactly. It's a new world and it's a new way of doing things, so we need a new set of metrics for this.

Tracie:

Yeah, for sure. So I have not asked you about this previously, but going back to career stuff, I know that you've had some mentors in your career that have been super impactful for you. I wondered if you could just take a minute and talk about mentorship and the impact on your career and how you've tried to pay that forward.

Dilip:

Yeah, absolutely. I mean, mentors have been critical for me throughout my life. Even when I was in my master's program, I had a couple of good mentors that helped me guide in my thesis work and my counselor on the thesis was really helpful and mentoring me in how to deploy it. Fast-forward from there to even some people you know and we know in common, like Chris Simenio who was my boss, we were working together, he has been immensely impactful in my growth, especially in corporate world, but that's in my entrepreneurial journey as well. My parents and my brother has been significant influence, so I'm really honored to have some of these mentors in my life that have helped me throughout the career and they've helped me grow and I'm looking to mentor people now. In my life, I try to spend time these days helping youngsters come up and help them with either product market fit or whatever issues they're dealing with entrepreneurship these days or technologically, from an AI perspective as well. So I am open to helping people who are looking for internships like that and will have questions, happy to help.

Tracie:

For sure. So as I mentioned at the beginning, we were both at the TechCrunch conference a couple of weeks ago. I was wondering if there were any ideas or things in particular that you came away with from that conference that seemed most interesting or most valuable for you?

Dilip:

Yeah, I think what came to me were two things that were coming up in this TechCrunch event. One was that small language models or tailored language models that were tailored to a specific domain or organization is going to be sort of where the next set of innovations is going to happen. And that I've seen quite a few companies since then that have also pitched, talked about LLMs or generative AI on a device like a phone or a laptop, so smaller versions of it. And so that's something that I was very interested in and exciting to see that happen because I think that will open up to a lot more business use cases that now businesses can feel a little bit more comfortable about their own data and data privacy and integrating their data into AI. So that's a good thing. The second piece that was along those lines was that whole agent agentic AI as they say, or AI agents, which to my mind is a whole new way of doing things. And I feel like that area is going to grow significantly in a lot of ways. I mean, we've seen Salesforce announce the agent falls, but where I see really the power of agents coming through is when disparate systems from different vendors are able to communicate and interact and complete that entire workflow without too much of a human intervention where appropriate is what the power of AI agents or multi-agent frameworks is going to deliver. And I'm excited about that. We're helping clients build some of those things. We built one for a retail client recently and we are already seeing some results from that, where we are seeing an improvement in terms of our contact center we built and we are seeing an improvement is about 30% improvement in their handle time. So we've eliminated what used to take them 30% of that average handle time to just do the grant work. So by building out these multi-agent solution, we've been able to eliminate that. So that's a fantastic use case. And we are seeing that also on the inventory management side where we can now tie in your in-store inventory, have an ML model that is constantly looking at your inventory levels and figuring out where they are and are they going along. And once that happens, they can trigger off agents that help clients place that order either in the warehouse or to their supplier to backfill the inventory. So those are some of the exciting new use cases that I'm looking forward to. And we've already started helping clients build some of these things, so I'm very excited that we can build and showcase some of these things as we go along.

Tracie:

Yeah, I'm looking forward to seeing that. So as we wrap up today, do you have any final thoughts that you'd like to share?

Dilip:

Yeah. I think we are at an exciting time in our journey in technology. I mean, we've said that in various times in the past, whether it was with the internet or with mobile phones or with e-commerce or when ML was popular. But now with AI, I think it's really a different type of change that we are going to see, and it's going to be a long-lasting event of change, meaning this change is going to take us 20, 25 years to really get and see the benefits of it. We have to be prepared for that. So I'm excited how we as humans can partner with these AI solutions and really help improve our lives and work and deliver some valuable benefits out of it.

Tracie:

That's great. And that's a very hopeful note to remind folks that we're sort of in this together with AI and there's going to be a lot of benefit for years to come from it.

Dilip:

Absolutely, and I'm looking forward to things like climate change. How are we going to leverage AI to help improve and reduce the impact of the climate change that we are experiencing right now? So something is a passion project that I'm doing on the side, just if anybody's interested. Talk to me about that.

Tracie:

I love that. Yeah. I love that there's the granular and then there's the strategic and forward-thinking side of AI.

Dilip:

Absolutely.

Tracie:

Great. Yeah.

Dilip:

Well, Tracie, it's been great talking to you.

Tracie:

Thanks so much for being with me today.