Behind the Scenes - How We're Driving AI Innovation within Google Cloud Consulting

Having the opportunity to speak with customers every day, we hear a common challenge: how to maintain agility and innovation within mature, complex organizations.?

How can we innovate from within to drive evolution and change??

Today I am sharing a little of how we’re incorporating AI across our business.

How we started: Earlier this year, like many of you, we made a strategic investment to enhance our business with AI. We operated under a simple global remit: activate our teams of experts to drive rapid development of AI use cases - both internally in how we run Google Cloud Consulting and externally to enable the services we provide to customers.? I also took a senior leader, Bob Greenlees , out of their day job to lead the AI efforts.

Every team within Google Cloud Consulting has been learning and implementing AI - we’re walking our own talk and eating our own dog food (although I have to say I prefer “drinking our own champagne”). To bring these collective efforts into focus and bring the outputs of our innovation into practice faster, we created an internal effort called LEAP to act as a catalyst for accelerated impact.??

For more on the importance of speed in getting started with Generative AI, which has been core to our approach.

Here’s the process we created and followed. We invite you to create AI Innovation following these steps within your organization:?

Three Steps to Drive AI Innovation within Your Teams

1. Apps vs Platform: When we started, we first focused on individual use cases, but quickly saw that many shared common challenges. Most use cases had the same data source challenges and similar infrastructure needs. This is why we’ve decided to build LEAP as a platform, rather than a single app. With this approach, we’ll address common data and architecture issues, making it easier to expand and add solutions over time.

2. Pilot & Iterate: AI technology is evolving at such a rapid pace, you need to quickly validate use cases and be ready to modify your approach. From hackathon POCs to pilots to production tools, at each stage we have iterated to integrate product enhancements and evolving architecture best practices. We have also refined UX aspects as users have provided feedback.?

3. Align with Priorities that AI can Impact: We’re focused on ensuring GCC is at the forefront of being an AI-powered services business, so the core priorities for us are increasing productivity and delivering enhanced value for customers and our partners. Key to aligning with these priorities was going bottom up, top down, and working across teams to break down silos.

I recently read Boundless, where Ben Greenfield states “Silos risk their own downfall by ignoring the outside world.” I couldn’t agree more.

AI can be a distractingly exciting technology, so it is essential to establish priorities to focus your teams for maximum impact.?

We asked Bob for tips we’ve learned along the way. He had five recommendations for leaders seeking to drive AI innovation:?

• Lean into the unknown: One of the reasons you are embarking on this journey is your curiosity, your desire to take on a new challenge of innovation. But the unknown can be intimidating and fraught with fears of failure and not keeping up with technical change. That is exactly how I felt at the beginning of this journey, and still feel every day: uncomfortably excited about both the opportunity and the challenge. Conquering the unknown is why you are here.
• Learn and adapt: Our first chat application isn’t perfect, but it exists. And it is a proving ground for what we need to do in the second version. We found that focusing on high quality experiences for fewer use cases is far more valuable than a general purpose assistant trying to solve many problems at an average level. We’re also adapting architectural decisions based on new best practices being developed every day in this space. We’ve learned, for example, that a retrieval augmented generation (RAG) approach with LangChain would be essential for our next iteration. Our first step, although imperfect, led to all these learnings.
• Establish a core group: When we started, priority one was finding people excited and motivated to execute. We brought together folks with a mix of product, engineering and program expertise. Many were not AI experts, but shared a common mindset: a determination to learn and a positive mindset.??
• Enjoy the journey: If you are reading this, you are likely thinking about how to implement AI for your organization. That is an exciting and fortunate opportunity - despite the scale of the challenge ahead, it is important to appreciate the journey. Your enthusiasm and belief in what’s possible will be infectious - it will motivate others to join your mission, it will help align leaders across your organization and it will be critical to work through challenges as they arise.
• AI Principles: First, Second & Last: Google’s AI Principles are a guiding factor in all that we do, and can provide guidance for other companies as well.?

We’re focused on ensuring GCC is at the forefront of being an AI-powered services business

We’ll be sharing more in future updates on the pilots we’re running, but here are three that are core to our business and are being piloted today:?

1. AI Powered Technical Account Management (TAM): We’re developing tools like Hive (below) to enable TAMs to help customers in providing product guidance, managing cases and escalations, leveraging playbooks and delivery guides and identifying customer health risks and opportunities.?

2. Professional Services Delivery: Tools like Delivery Navigator (below) have added AI functionality to help our engineers and delivery teams reference implementation best practices, past delivery scenarios and delivery methodology in order to provide exceptional service. We recently announced that we’re making this tool available to Partners.

3. SoW Generation: We’re leveraging insights and context from our past contracts and our contracting standards to draft SoWs(statements of work) based on customer use cases and objectives. Over time this will reduce the time to get started while also improving the overall consistency of contracts.

Are your priorities driving AI adoption?? Reach out to [email protected] to talk to one of GCC’s AI experts.?