Innovation can take many forms in the enterprise. The latest insights from MIT Sloan Management Review offer leaders three approaches to increasing innovation: enabling organizational agility, pursuing deep-tech ventures, and committing to best-in-class artificial intelligence products.
4 guardrails to empower agility in the enterprise
Many leaders would like their large firm to be as nimble as a startup. But many also face the challenges of risk aversion, shared decision-making, and the need to align innovation efforts with strategic objectives or shareholder expectations.
a research scientist at the MIT Center for Information Systems Research, explains how enterprises can encourage organizational agility, proactivity, and adaptability by putting in place four guardrails that define how employees should pursue innovation.
- Put purpose into action. It’s difficult to promote agility when strategic plans take so long to write that market forces render them obsolete before they’re finished. Reframing the strategic plan as an “engaging and relatable narrative” allows for flexible road maps while limiting the impact of failed experiments.
- Democratize data. When data is siloed, only those at the top have enough information to make evidence-based decisions. Data that’s accessible, understandable, and actionable for all teams is necessary to encourage collaboration and creative problem-solving.
- Establish minimum viable policies. Applying high-level principles across the organization strikes a balance between managing expectations — such as which components to use or which assets to purchase — and enabling flexibility and reducing complexity. Teams that know what they can and cannot do are able to get to work quickly.
- Provide the required resources. Agility means moving quickly when an experiment shows promise. That’s difficult if teams need to wait for the next annual budgeting cycle. A venture capital type of funding approach lets innovators unlock off-cycle funding if their prototype shows potential — and again if it succeeds as a minimum viable product.
Read: The four guardrails that enable agility
Deep tech can catalyze long-term innovation
Many of the problems today’s enterprises face — including sustainability, labor shortages, and ongoing market shifts — can’t be solved with a simple digital offering. These challenges need an infusion of deep tech, broadly defined as “a category of solutions rooted in atoms rather than bits,” such as quantum computing, synthetic biology, or other innovations that emerge from the research lab.
MIT Sloan associate dean writes that deep-tech startups hold promise for industries that tread lightly in R&D, such as financial services, retail, and infrastructure. But these sectors, along with health care delivery and government services, aren’t always accustomed to working with emerging technologies. Along with co-authors Martin Murmann and Stefan Raff (both of the Bern University of Applied Sciences), Murray identifies three challenges for deep-tech ventures, along with best practices for managing them.
- Commercialization risk. Deep tech is complex tech. Prototypes are expensive, and market reception is hard to predict. Enterprises need to build in-house expertise to identify the partnerships that make the most sense.
- Capital investment. Deep tech is expensive tech. Early-stage financing rounds often exceed $20 million. Shared-risk agreements between enterprises and entrepreneurs can provide necessary resources or supply chain access.
- Extensive timelines. Because they’re complex and expensive, deep-tech ventures often take at least a decade to materialize — longer than the tenure of most executives. Clearly defined milestones demonstrate progress and longevity.
Read: Why you should tap innovation at deep tech startups
Innovative, responsible AI tools at Scotiabank
Scotiabank, like many financial institutions, has embraced generative AI and built a chatbot for its contact center. But few others have received the same level of recognition as the Canadian bank, which has won awards for innovation and AI ethics. MIT Initiative on the Digital Economy fellow Thomas H. Davenport and co-author Randy Bean explain how Scotiabank built its chatbot and made it stand out.
The bank took a collaborative approach to developing the chatbot, with designers working alongside engineers and data scientists. The model was trained using an “AI-for-AI” strategy, with automated workflows for curating its knowledge base and identifying training topics. Developers did their best to curate unstructured data before feeding it into the model. Scotiabank also worked with Deloitte Canada to develop an application that assesses the ethical impact of an AI use case early on; the bank has since required anyone doing advanced analytics work to participate in data ethics education.
While Accenture and DataIQ have recognized the bank’s efforts, the chatbot’s results speak for themselves. Since the chatbot’s debut in late 2022, its accuracy level has increased from 35% to 90%. More than 40% of customers get their questions answered without human intervention. And when human agents need to step in, the chatbot’s automatically generated summary of the online conversation reduces resolution times as much as 70%.
