AiThority Interview with Hamza Waqas, VP of Engineering at LivePerson

Hamza Waqas, VP of Engineering at LivePerson, chats about key architectural and operational principles, top trends shaping the future of AI-powered conversational experiences, common struggles with AI adoption and more in this Q&A:

——–

Hi Hamza. You have a strong track record of delivering scalable SaaS solutions globally. How do you approach aligning multi-year technical strategies with business outcomes?

First of all, thank you for the opportunity to share my perspective. The market is shifting quickly, and times are changing like never before. My approach to aligning multi-year technical strategies with business outcomes begins with a farsighted perspective, establishing a clear and direct line of sight between technology investments and key business objectives. I start by deeply aligning with business strategy, including revenue targets, customer metrics (acquisition and retention), and operational (engineering) metrics. Once foundational goals are clear, I work with stakeholders and teams to translate them into actionable OKRs. The OKRs allow us to track the progress against metrics while working on a unified strategy. I firmly believe that in order to make everyone speak and understand a consistent message, we must have a clear understanding of “Why” – why are we doing this, why is it important, and why this approach over others, considering the broader market landscape.

And it is important to build a culture of autonomy and accountability, where everyone understands the common goal and contributes towards achieving it. In a sizable organization, micromanagement just could not work; instead, a favorable culture drive is important where everyone understands their role and, most importantly, the impact of their contribution. I always tell my leadership team to fully understand how individuals produce an output, and it is our job to stitch those outputs into an outcome. While individuals may feel satisfied with what they contribute, our collective success depends on an outcome. The outcome must contribute towards achieving our key goals that we measure through key metrics.

Previously, strategic planning prioritized competitive analysis. However, a revised approach necessitates that the strategy be rooted in a fundamental business value proposition and iteratively adapted to evolving market dynamics. Achieving sustained vision in conjunction with organizational agility requires the segmentation of long-term strategic plans into discrete short-term (three-month), mid-term (six to twelve-month), and long-term (twenty-four to thirty-six-month) cadences, facilitating adjustments in response to market see-sawing. The contemporary proliferation of Generative AI exemplifies the imperative for such adaptability.

The Generative AI wave has recently forced many companies to discard and rebuild their strategies. This reactive approach to strategy can be chaotic. Multi-year strategies with time-based goals provide a better long-term vision. This allows companies to adapt to trends without compromising their unique offerings.

Also Read: AiThority Interview with Vishal Joshi, Head of Engineering at TabSquare.AI

LivePerson operates at an immense scale, handling billions of transactions. What are the key architectural and operational principles that enable this level of scalability?

Customer obsession sits as a core principle of LivePerson’s engineering culture. LivePerson’s ability to operate at an immense scale, handling billions of transactions at planetary scale, is underpinned by several key architectural and operational principles.

These include a Distributed Architecture that allows for horizontal scaling, and Statelessness in service design ensures that any instance can handle requests, improving resilience and scalability. We also leverage Polyglot Persistence, choosing the right data store for specific needs to optimize performance and manage diverse data types. Intensive platform monitoring and observability are crucial, providing real-time insights into system health and enabling proactive issue resolution.

AI-driven customer interactions are rapidly evolving. In your view, what are the top trends shaping the future of AI-powered conversational experiences?

It is clear that embedded AI is our future, as it simplifies and empowers tasks efficiently. LivePerson has been the market leader in Conversational AI, servicing conversations since 1995, and the new Generative AI trend reflects exactly what we have been offering to our customers. The benefits of intelligent AI-based conversations include faster turnaround times – customers no longer have to wait for human agents to be available during “working hours”, or navigate through preset options to identify their pain points. Context awareness in AI has eliminated wait times, enabling faster resolutions. We have noticed a surge in positive experiences when humans interact with conversational AI, compared to traditional chatbots. Looking ahead, several key trends will fundamentally reshape how humans interact with technology in conversational settings. Firstly, we are seeing a move towards

Hyper-Personalized AI, or what I think of as User-Aware AI. This AI will become a companion and, over time, will learn our patterns to provide increasingly tailored service. Personalization has started to make an impact with targeted ads (AdTech) where customers are shown the relevant ads, but it wouldn’t just end here. It would go a few steps further, mirroring our personality, choices, and preferences, reflecting them in tailored responses. Secondly, Emotionally Intelligent AI represents a significant leap. Today, generative AI can still feel artificial in its responses and is not sensitive to human emotions. Textual data alone does not translate into how humans feel in a given moment, where words and expressions can have nuanced meanings. Empathy within Emotionally Intelligent AI will be a game-changing factor, unlocking doors for therapies and emotionally available companions. While this topic itself could be controversial, we are already seeing a surge of AI platforms enabling AI as a romantic partner. This will certainly not be the only use case.

Once AI is empathetic and can distinguish emotions, it unlocks the potential for therapeutic practices and emotional support. Thirdly, Seamless omnichannel experiences are paramount, enabling AI to cater to broader, consistent experiences across different mediums. Screentime has increased in recent years, making it extremely important to achieve consistency as users shift their attention span across channels. Practically, moving from laptop to phone while walking towards your car, then changing to CarPlay while driving, and then shifting back to a tablet — all these transitions must be catered to with AI in order to provide effective service.

Lastly, our SaaS industry would be changed forever. Looking back, I remember the dot-com period when a wide range of software started to be widely available as one-time purchase fees and run it yourself. Then, we shifted towards SaaS (post 2008 crash), and software was available as a service, and on a subscription basis. Now, AI will obsolete the SaaS model, and initiate a new possibility of AI Agents who could connect with various sources, systems, and even other AI Agents to perform the job. The typical SaaS may exist to an extent, but would no longer be how we would interact with systems. The AI Agents would play a vital

role in orchestrating the services we use today (our bank, health, travel, etc.), interfacing them in a single pane that is context-aware, and of course, emotionally intelligent. Do we want our AI companion to stop us from making a transaction? That’s a point to ponder.

The increasing capabilities of embedded and generative AI are driving these trends, as exemplified by LivePerson’s extensive experience in this area. This is leading to a focus on interactions with technology that are more human-like, personalized, and efficient, ultimately transforming the way we engage with it.

Many enterprises struggle with AI adoption due to legacy systems and data silos. How can engineering leaders drive seamless AI integration into existing workflows?

It’s a key problem that I am sure every engineering leader is currently facing in this recent paradigm shift. There are a few generalized approaches, but I take the liberty of exemplifying how to decide to seamlessly integrate AI into existing products. First, we must have the baseline metrics on our non-AI workflows (or features) – that most of the sophisticated systems already possess. Then, based upon those metrics, we must classify our product into tiers (1-3), where higher tiers have more user touchpoints. There is no urgency in revolutionizing the email content if it is sent too often. AI is integrated into the optimization of the journey itself once the journeys have been tiered. Like in our case, Smart Suggestions (an NLP use case), or fewer clicks to finish the transaction (predictive analytics), etc., and then validate with a small set of users whether our hypotheses are correct. Going back to the initial point, we have baseline metrics, and post AI integration, the baseline metrics must trend up (or trend down if that is the expected outcome) to validate the hypothesis. Once we establish the success, integration should roll out to the rest of the users, and we repeat this for all of the tiered features.

Putting user journeys in tiers helps catch up with a) competitors, and b) user experience. If customer obsession is the forefront of your product, it would be natural to improve key workflows ahead. From my perspective, there is no rush to revolutionize the entire legacy system overnight. The engineering leaders must prioritize business continuity with the available resources, and innovate on key areas that could give immediate return in a) feedback, b), user experience, and c) satisfaction. I personally never felt any concern with legacy systems, they are, at the end, called legacy, because business succeeds way faster than technology can catch up. And a good engineering leader must not let data silos intercept how the business shifts, and it can easily be overcome by forming a data lake, and data governance. No matter how heterogeneous the systems are, both practices can quickly enable your engineering teams to catch on to the AI wave. In terms of execution, engineering leaders must rethink how they are organized and reconsider whether they must reorganize themselves by putting AI integration into workflows. Discover whether cross-functional teams are relevant and if they may help you achieve driven outcomes. Luckily, my leadership team has always been open to reorganizing ourselves as the business changes.

With AI reshaping customer engagement, how do you see the next wave of AI-driven transformation in contact centers unfolding?

The good thing about mature Contact Center as a Service Platforms (i.e., LivePerson) is that we were ahead in Generative AI capabilities from the start. I often say this in interviews: We were doing Conversational AI before it was cool. Yes, an increasing trend in Generative AI equally helped us to be recognized with the growing need for improved customer experience. It simplified several key touchpoints that we had to solve with only proprietary technology earlier, and I am personally an open-source enthusiast, so the transformation perfectly aligns with my idea of how the market should move. From my perspective, I can say that once Chatbots were hated to an extent, customers only wanted to connect with humans as soon as possible. This is not the case anymore, where my biased opinion thinks more than half of the consumer and brand interactions are done through Conversational AI. Yes, Chatbots are bad, but Conversational AI is not. Notably, customers no longer need to oblige with working hours or suffer with wait times to talk to a human, so yes, that’s a big win already. Once Generative AI matures into Empathic AI, I am certain the majority of the interactions would be HCI-driven (Human-Computer Interaction), and only a percentage of interactions would involve human-to-human communication.

The next wave of AI-driven transformation in contact centers will be significantly shaped by AI-powered agent augmentation. We are already observing improvements in Customer Satisfaction Scores and Turnaround Times by leveraging AI within agent workflows. While today, AI often recommends smart replies based on context, the future will likely see AI confidently taking over more routine responses autonomously. Another crucial trend is Predictive and Proactive Engagement. Although mature contact center systems connect with CRM solutions and feed AI with historical conversations, the next wave will refine this into a truly unified Conversational Experience, moving beyond time-based and siloed engagements for both human and AI interactions. Furthermore, I believe the Integration of Multimodal AI will be a defining characteristic of the future, extending beyond text to encompass media streams like voice, videos, and images. Imagine a

Speaking-Conversational AI that can hold valuable conversations, revolutionizing clunky IVR systems, and seamlessly escalating sensitive matters to human agents. This same principle applies to other media, with potential applications like AI Interviewers removing conscious bias in hiring.

Would that mean the Contact Center will become obsolete? No. I personally think they will be more relevant than ever, for exactly those subsets of interactions that require humans to intervene. I am not sure if we would still call it a Contact Center, though

Also Read: AiThority Interview with Aashima Gupta, Global Director of Healthcare Strategy and Solutions at Google Cloud

For engineers aspiring to leadership roles, what advice would you give on developing both technical and business acumen before we wrap up?

I feel proud when engineers aspire to assume leadership roles, and open to learning the trade. When I coach the next generation of leaders, here is how I break it down:

Identify your leadership style: Leadership styles vary greatly, and self-awareness is crucial for emerging leaders. It’s important to understand your personal leadership style and how it aligns with the requirements of your role. While adaptability is key in different situations, recognizing your natural leadership tendencies is essential for success.

Continuously learn: I often say this to my leaders, engineers only learn from better engineers. You can be an amazing manager, but you won’t win trust until you aspire to learn. Know what areas you need to improve, and continue to excel in them.

Be a commendable negotiator: The most important entry-level skill on the leadership track. At every step, literally every hour of the day, you are negotiating with your stakeholders, peers, customers, or the team itself.

Data removes bias: Put data at the forefront of your culture. More often than not, you overcome unconscious biases and prejudices that were developed over the years of experience. Understand that you are trusted with an important responsibility, tunneling significant investments through you to your team, and you must take responsibility to make quality decisions, which only comes with data-driven decisions.

You don’t know a thing: As much as you know almost “everything,” it is time to let it go and trust your team. Do not rush into solutions immediately, and be open to hearing everyone out. Understand that creative ideas only come without biases.

[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]