How GenAI Transformed Me

It's been approximately six months since I've had the opportunity to change functional areas within Google Cloud. I have spent so much time talking about Telecom Transformation, I thought it was a good moment to reflect on my own personal transformation.

For those who are unfamiliar with my journey, I will not-so-quickly summarize it, then you can find out about how this journey has transformed me, and what I've learned so far. (Or, you can choose not to read any further and go about your business, which won't offend me.)

Telecom Data Before Data on Telecom Was a Thing

I had a long-standing (23 years!) history in the Telecommunications Industry focused on network transformation, and dabbling in how data, AI, automation and cloud native technologies impact overall network operations. I launched data over cellular (2G 40kbps) before data over cellular was a thing, and built IP-based Telecom networks before packetized Telecom was a thing. I'm definitely no Marconi, but I guess I could claim I was there "at the beginning" of cellular data.

I launched network-based billing when "on-net" and "off-net" (you know, a precursor to "OTT" when carriers had "walled garden" services) were things. I worked ten stories below ground in a Telecom CO in Paris. I launched circuit over packet (CEoPS), SS7 over IP, and 3G PDSNs and GGSNs with carriers, and was contributing to the very first ETSI NFV Workshop in Cannes when we were fortunate to have an ice storm freeze the hilly roads). I wrote two books on mobile networking - both from an end-to-end system perspective and from a protocol stack perspective, and hold just about 50 patents in mobile technology (depending on how you choose to count). I've been to 21 Mobile World Congresses. I was a "first five" hire for Telecom / Service Provider Lines of Business at 4 different companies. To say "been there, done that" would be a good way to summarize my journey.

More importantly than what I knew was who I knew. I was part of the Telecom community. Many of the engineers who I worked with to build 2.5G and 3G networks grew their careers in similar ways as I did, and as I achieved Executive levels in my career, they were leading Network Development or Technology Development or Network Operations. The Director/VP-level leadership teams I partnered with were off as CTOs at carriers or network vendors or CEOs at startups. Analysts in the Telecom industry stay fairly consistent, and I bantered with them regularly - sell-side, buy-side, independent, contrarian - made no difference. The Telecom Network WAS my Social Network.

A Unique Opportunity, 8 Months in the Making

Then, I was asked to take on an additional charter at Google Cloud. I was asked to take on our Contact Center AI solution space in mid-2022 because of such significant traction with the Telecom Industry. Nothing unique about the product to make it well-suited for Telecom - just a timely solution addressing one of the industry's biggest pain points - Customer Experience (and cNPS). Knowing HOW to work with a Telecom, and the nuances of how Telecoms work with vendors/partners, how executives think, and how things get done was more important than knowing the technology. So I did it, and as a byproduct of leading, I learned the technology.

Then something funny happened at the end of 2022 - #GenerativeAI came to the forefront, with an overwhelming percentage of use-cases oriented around using Large Language Models (#LLMs) to understand understand human #intent, and using that #intent to trigger #generative actions. And I was fortunate to have been leading a team that had deep technical and operational experience at building #conversationalai agents.

So - the opportunity that presented itself was to step away from #telecom (largely, maybe not completely) and expand across more industries, more domains, and more customers. An opportunity to partner more closely with product engineering to define and iterate on GenAI products and solutions, and partner more closely with customers to define, develop, pilot, and launch #GenAI use-cases. Scary? Yes. Unknown? Yes. Uncomfortable? Yes.

Worth taking? Most definitely!

Learning About Myself By Venturing

I've moved roles within companies. I've moved across companies. I've been a building of tech, of organizations, of people. But I've never strayed far from the domain I knew best - where people in the room knew and and, quite candidly, I could feel confident I was the most knowledgeable in the room.

This time it was different.

So, how can I best summarize what I've learned and apply this to every day life?

1) Listening and Learning is Energizing (and Tiring!). Listening is easier when you are forced to listen. When you are not the most knowledgeable and most experienced, you have no choice but to defer. You find yourself in listening and learning mode more often than not, and absorbing factual information and viewpoints/opinions that you find yourself down rabbits-holes trying to validate. It turns out to be energizing to learn once again, and it's exhausting trying to absorb new information at a rapid rate (especially in GenAI, with a high rate of change). Learning to listen feels like an acquired trait now - I am confident that I am able to continue to listen even when the knowledge gap closes.

2) Soft-skills are highly transferrable. As important as WHAT I knew and WHO I knew in Telecom, it turns out the skills I developed as a critical thinker, and an explorer, as an innovator, and as a leader are highly transferrable. I had thought that my success in leading large organizations was because they trusted my vision for the future of telecommunications. Maybe it was, but there is more. I have a newfound confidence to apply my understanding of (and empathy for) people, to think at the systems (design) level, to mix consultative leadership and collaborative engagement, to apply technology to business problems, and to challenge the "status quo" with simple questions or observations, to be foundational leadership attributes. It's a lesson for me to never stop developing myself around the edges (no, not the Telco Edge!).

3) True Leaders recognize leadership. This was important for me, as I am currently surrounded with a new set of leaders who aren't as interested in tapping into my 20ish years of Telecom domain knowledge. And, they understand that my domain knowledge in CAI is not yet on part with others. Yet - this new professional community still intuitively see the transferrable soft-skills mentioned in #2 above, and find ways to tap into those skills to further our strategy, products, solutions, and customer relationships. This is something that I've internalized quite a bit - historically, as I've moved into more senior positions in Telecom, I've pulled known entities with proven domain-specific experience to help build. It's a reminder to me to look for the intangibles when developing high-performing teams.

4) Hindsight is MORE than 20/20. Turns out, taking a new role does not take away those 20+ years of experience and knowledge. As a lurker and an observer over the last 6 months, including in/around MWC 2023, I have a higher degree of confidence (freedom?) in expressing contrarian opinions. To note, I've always expressed pragmatic opinions and tried to venture away from pure corporate marketing and product positioning, but, when you are close to an industry, and the team you've built, the products and solutions that team has built, and your company is reliant on strong relationships within that industry, there is a clear boundary that you opt not to cross, even in your own personal interactions on social media (and the "opinions I share here are my own" is not sufficient). It's top of mind for me to strive for comfort in challenging status-quo, not just hedging bets.

5) I still have a heart for Telecom (and Networking), but with a new lens. I can't help it - I find myself returning to Telecom networking and continually thinking about how I would apply generative technologies to telecom networks and operations. There are so many ideas I have, many of which are highly disruptive to some of the directions we've pursued to-date (Example: Do we really NEED multi-vendor network abstractions or common data models, or can I generate vendor-specific templates and config files by indexing configuration guides and using existing configs in few-shot prompting?). I will be watching anxiously to see state-of-the-art, from Google and others, emerge here. So far, all I've seen is either (a) misunderstanding of what "Generative" in GenAI means, or (b) a miscalibration on the sophistication of LLMs leading to a perceived need for industry-specific models. Maybe it means I didn't say goodbye to Telecom networking, but instead more of a "see you soon".