Dalle-2 prompt: a person in a collaborative and creative office space using technology in the style of a cubist illustration emphasizing primary colors
Disrupt Avanade was established to explore how Generative AI could be implemented to transform work at Avanade. The initiative focused on establishing a strategy for the successful implementation of AI through continuous ideation, experimentation, and learning, with the goal of reimagining how work is done. The framework included equipping employees with necessary resources to build a culture of innovation, and prioritizing ideas for experimentation based on business value. Early outcomes showed a shift towards more innovative ideation and rapid experimentation enabling continuous learning and agile decision making.
As part of the small and innovative Disrupt team, I find myself wearing a few different hats to fit the need at hand. Depending on the day I may describe myself as a business analyst, project manager, strategist, UX researcher, or product designer. In this work I am most often collaborating closely with a strategist, Engineering team, functional and business experts, and IT Portfolio leads. I am often responsible for workshop facilitation, supporting storytelling and visualization, interviewing, analysis, experience strategy, requirement gathering, process flow diagramming, journey mapping, and UX design.
In November of 2022 OpenAI announced ChatGPT, and just two months later they garnered an estimated 100 million monthly users, setting a record for the fastest growing user base in history (Source: Techcrunch). In May 2023 at the Microsoft Build conference Satya Nadella announced GPT powered Copilots would be integrated into nearly every Microsoft software or service from Edge and Bing to Windows, Teams, and the Office suite. The pace of development in Generative AI has been incredible, and with this has come much speculation and promise of what these agents could mean for productivity and the future of work.
In the wake of the abrupt shift to remote work in 2020, the IT and Professional Services industry experienced a significant boom due to the heightened demand for accelerated digital transformation. However, after two years of growth, this boom began to slow down. Excess spending finally caught up, concerns of an impending recession spread, and market pressure grew to cut costs and achieve more with less. Coupled with the unprecedented level of tech innovation and development in AI during this time, an opportunity emerged for Avanade. How might we take advantage of this transformational technology to disrupt Avanade and lead by example for our clients?
How do you do this in a way that is employee centric so that we are not just creating solutions for temporary productivity boosts, but building a resilient and future-ready workforce that can continuously innovate and evolve? How do you ensure that the solutions you introduce to the workforce are in service of the humans doing the work, and enable them to spend more time focused on the work they do best?
In July of 2023 Disrupt Avanade was established, creating a task force dedicated to exploring how [Generative] AI could enable us to be more effective, productive and reimagine how we do work. The mandate of our team was to:
The behavioral implications of generative AI were clear, what we identify as work today, will be fundamentally different from what we see as work in the future. You cannot reimagine the way you do work without having a workforce that is constantly experimenting with technology and thinking of better ways to do things. A significant cultural shift would only be possible if we placed people at the forefront with a strategy founded on the principle of:
AI Everything, Everywhere, All At Once through empowered curiosity
For this human-centered strategy to be successful it needed to outline the support structures, resources, and systems necessary to activate a culture of continuous, iterative, and compounding innovation, creating an environment where all employees are driven to continuously ideate and experiment on how to improve everything they do, certain resources and support structures are necessary.
The framework objectives were to:
Despite being in the first quarter of execution, one key outcome to highlight is the improvement in ideation quality. Prior to the establishment of the ideation support structures and resources, most submitted ideas for AI were found to be relatively basic in nature, focusing on the automation of existing tasks and processes. Following the introduction of trainings and ideation frameworks including the ideation workshop series, we’ve seen our coaching on innovative thinking push concepts further helping employees to reimagine value chains and core competencies, driving greater value potential for the business.
A great example of this can be seen in one of the ideas submitted for the sales team. An early idea was to use Generative AI to generate statements of work based on estimation tool inputs. Upon a deeper reflection on the sales process and what is possible with AI during the workshop series, this evolved further into the concept of a “Sellers Agent”; a tool to collect all communications and materials related to a client to generate a summary of the client opportunity and provide coaching to support them in their conversations. This would help sellers ensure they’re understanding the core client need and considering all Avanade offerings, leaving nothing on the table.
It’s an exciting yet daunting assignment, shaping a strategy for the enterprise use of an emergent and rapidly changing technology like Generative AI. This initiative was far larger than building out a few interesting generative AI POCs. We had to define a human-centered strategy capable of activating the cultural shift necessary to drive continuous, iterative, and compounding innovation, reimagining the way we work. In order to do so we identified the two foundational pillars of execution to be Continuous Innovation and Value Realization. My contributions as a part of the Strategy and Innovation team were centered on Continuous Innovation and the definition and execution of the Ideate, Experiment, and Learn components.
At a professional services firm like Avanade, it is no small request asking all employees to stay curious, experiment, and continuously ideate while delivering on their standard day to day client responsibilities. Knowing the success of our strategy hinged on a significant culture shift, one of our key objectives was the establishment of a steady pipeline of ideas from across the organization oh how employees would use AI to reimagine how we work. In order to facilitate something like this you have to equip them with the right resources, structures, and incentives to make creative and innovative ideation a core role expectation.
Technology Capability: To imagine what can be possible, you have to know what the bounds of your sandbox are and what’s available to you within it. This requires establishing an accessible knowledge pipeline of what AI discoveries are being made, how the technology is developing, and what others are doing with it. We consider this one of the fundamental building block or lego sets necessary to provide our employees with to enable them to think creatively.
Business Capability: Beginning by looking inward. At a time when there is so much uncertainty as we look to assess where there can be tangible value from these technologies, we found it vital to establish an approach for continuous self-reflection through a critical and inward looking eye. By being critical of what we do, why we do it, and what we want to be the best at, we are able to stay laser focused on our priorities and strategically direct our efforts and investments. This is the foundation of our assessment framework for measuring value and we even found it to be a helpful framework when introducing ideation efforts to those that had less experience flexing their creative thinking muscles.
Mastery: When we assess skillsets we look at mastery as not just having a knowledge of the skillsets and tools available, but having a command of that knowledge that enables an individual to take that foundation and meld, modify and adjust to fit a specific situation or need. To build a mastery of AI among our employees it is more than just making a knowledge pipeline available. This is where tailored training and learning opportunities are vital to drive quick and impactful up skilling. One of the first initiatives launched this fall by the L&D team was the Avanade School of AI, a tailored foundations course rolled out to all employees.
People: You can bring a horse to water but you cannot make it drink. You can make a training mandatory but it’s much harder to ensure those learnings translate into changed behaviors and outcomes. To take mastery and apply it to drive continuous innovation among the workforce, you have to provide the people with the proper resources, time, and incentives for practical experimentation and the psychological safety to sometimes fail. This began with top-down leadership endorsement, protected time for participation, and thoughtful incentive structures prioritizing recognition, the use of peer networks, open information sharing, and radical visibility.
Ideation Methods: Providing structured ideation frameworks, materials, events and channels for submission opened the pipeline of ideation across the organization. We quickly found there to be a few tiers of creative ideation that could be clustered by the type of value they drive for the business, covering a spectrum from productivity centric “optimize for today” ideas through to entrepreneurial “reimagine for tomorrow” ideas. Regardless of the type of value generated, all ideas were important to capture, share, and develop to reinforce the culture we wanted to build. This led to the establishment of a few types of ideation methods and channels including a white-glove facilitated workshop series, discrete hackathons and ideation events, and asynchronous idea submission with the custom Innovation Station tool.
Evaluation and Prioritization: Establishing a steady pipeline of ideas was a core objective of our work. This meant that if things were working, there would always be a surplus of ideas that could not all be developed. It was critical to have a light yet well-defined evaluation framework to quickly assess and prioritize ideas to ensure we were best allocating our efforts in service of our strategic objectives and producing outcomes. Our value framework takes into account a spectrum of factors including growth potential, cost reduction, risk mitigation, and strategic objectives. Ultimately, the effectiveness of this framework depends upon accountability and trust among the team to move quickly, make decisions, build, and iterate.
The experimentation process takes prioritized ideas and progresses them through rapid prototyping to test, iterate, and scale. Having an approach that moves fast to define, experiment, learn, and iterate ensures compounding progress can be made and there are tangible results to share and learn from. One of our first experiments was the creation of a custom tool for AI idea submission called the Innovation Station which is what some of the featured visuals are highlighting. For all experiments the primary objective is identifying opportunities to test the limits of what we can do with generative AI and for this one we used semantic search and novel human-AI collaboration patterns to drive knowledge content organization and creative ideation.
Define: The experimentation process is rough, iterative, and highly collaborative between a strategist, designer, data scientist and developer. Experiments are defined by identifying:
Develop: Following the definition of an experiment, wireframes and requirements are handed off the the AI lab development team. Collaboration continues over 1-2 sprints as the developer, strategist, and designer meet to discuss blockers, questions, and iterations. This often includes conversations on meta-prompting and model structures in order to fine tune outputs to meet user expectations.
Test: After 1-2 sprints a baseline functioning PoC is built and end users are brought in for rapid testing. This process uses light moderation while they interact with the tool to assess user expectations, tolerance for error, and value potential. If a certain threshold is met, the experiment will continue with iterative cycles to refine the experience and progress towards MVP and scale. If the threshold is not met, the experiment is retired and recycled for learnings.
Learn exists at the heart of the overall framework. The collection, organization, and use of what we learn from these efforts is essential to continued development, growth and progress. Knowledge management is a challenge faced by most enterprises as we have seen information and data availability multiply exponentially, but our efficient use of the information has been unable to keep pace. Creating clear structures around organization, storage, and sharing help us to learn from both failures and successes as we move forward.
To see the experimentation approach in action review the Innovation Station case study: Human-AI collaboration for creative ideation