When David and Goliath Work Together | SGInnovate

When David and Goliath Work Together

Monday, April 29, 2019

Topics: A.I., Startups, Investments, Others

Delegates at the invitation-only AI roundtable organised by SGInnovate and ADB-DutchCham discussed strategies to encourage corporate-startup collaborations in the field of AI.

At the roundtable, we heard a range of views on the enablers and barriers to an effective corporate-startup collaboration as well as the benefits that may come from working closely together as part of an ecosystem. We agreed that startups and corporates operate on a very different time scales: a few months may be considered a very short duration for corporates but could represent a lifetime for startups. The attendees also validated how entities such as chambers of commerce, SGInnovate and AI Singapore could help bring corporates and startups closer together.

The sharing was done under Chatham house rules so we will not identify the invited speakers in this article.



The heart of a hummingbird beats at an incredibly rapid rate, exceeding one thousand beats per minute in some species. In contrast, the heart of an elephant beats at a meagre rate of approximately thirty times per minute. This disparity in the animal kingdom is a fitting analogy for the difference in pace at which large corporates and startups operate.

“A few months may be considered a very short duration for corporates but could represent a lifetime for startups… Startups are worried about survival every day, every week,” said Mr Steve Leonard, founding CEO of SGInnovate, at the ‘Artificial Intelligence (AI) Roundtable: Corporate-Startup Collaboration’ discussion held at The Netherlands Embassy in Singapore in March.

Jointly organised by SGInnovate and ADB-DutchCham, the invitation-only event brought together startup leaders and senior executives of conglomerates to explore opportunities for innovative partnerships in the field of AI.


Agreeing on the Destination 

For startups and their much-larger counterparts to work together successfully, it pays to begin with the end in mind—that is, to start each collaborative project with a clear use-case, said a director at an insurance and asset management company. Instead of adopting AI for its own sake, it is recommended that leaders at large enterprises clearly define the business problem that they hope to solve with AI, then engage with startups to build a bespoke solution. 

Some large companies may therefore elect to experiment with the use of AI through proof-of-concept (POC) trials with a startup, said a partner and co-founder of a business development firm. While this allows the corporate to “hedge its bets”, multiple POC trials without follow-up implementation can take a severe toll on startups, which may not have much time or capital to experiment with, he said.  

Hence, one of the delegates suggested having collaboration frameworks wherein large corporates bear part of the costs associated with POC development. With this structure, corporates tend to be more cautious with their requests and this helps them prioritise and more clearly define what they want. 

Taking a Top-Down Approach

Further dissecting the thorny issue of driving an AI innovation from concept to product, the delegates noted that it was more “efficient” for startup founders to engage directly with the leadership at large corporates. They agreed that if there are no business owner involved from the start, the collaboration usually ends with POC without going anywhere. 

Having a senior corporate executive on board also helps to pull the project through the different stages of development, such as navigating legal processes, negotiating commercial contracts and—very importantly—setting the tone for data sharing, said a CEO and co-founder of a logistics startup. 

Workarounds to Make Things Work

Before AI can perform complex functions, it needs to be trained on data. However, some corporate leaders expressed reservations about making data freely available to startups, citing privacy concerns, adding that “the consequences of data being misused or mishandled is an order of magnitude greater than using it well”. 

To overcome this barrier to collaboration, startups need to be sensitive to the cybersecurity posture of corporates and come prepared with some idea of the data security processes that need to be put in place. 

One of the startups explained how his company does not even upload data to the cloud during the early stages of engaging with a large company. Therefore, rather than forego collaboration altogether because of fears over data security, startups and corporates should work together to find an acceptable middle ground.

“Having worked with a large enterprise before, I can vouch that we performed way better as a company because our partner pushed us. We couldn't have done it by simply raising funds from investors and sitting in a lab,” a CEO of customer engagement solutions firm said. 

At SGInnovate, we believe that by combining the agility of startups with the resources and reputation of corporates, AI solutions can be developed and scaled more rapidly for real-world deployment. Hence, we are always interested to hear from our startup and corporate communities their insights and perspectives, and possibly areas of collaborations. As such, we host such events regularly to discuss, and inspire our deep tech community.

Li Lidao
Topics: A.I., Startups, Investments, Others

Share this with your network!

    You may also like the following: