Artificial Intelligence
Nov 26, 2017

Artificial intelligence’s nemesis: Big legacy

“God created the world in seven days, because he didn’t have to port anything from legacy systems” – this is an apt quote I recently heard from the CEO of a leading software company.

The same sentiment was echoed in a recent MIT Technology Review article, “Seven Deadly sins of AI Predictions,” written by Rodney Brooks, MIT’s former director of computer science and artificial intelligence laboratory. In the article, Brooks notes that the seventh major sin of AI is the speed of deployment is heavily influenced by the underlying refresh rate of technology.

A four-year lag

Unfortunately, businesses tend to trail behind in the adoption of new technologies – especially in comparison to the consumer space. My team has researched this over the years and found that it is hard to accurately predict when new technology will be productively embedded, at scale, in enterprise operations. And, a four-year lag is common between the initial prediction and actual large-scale adoption. The path between the two is often erratic. This is especially true for business-to-business (B2B) technologies that transform operations with deeply entrenched ways of doing things (not just systems).

A lean digital approach

Much of Silicon Valley – especially the startup scene – doesn’t appreciate the lag in innovation. That’s why methods like “lean digital” are particularly useful since they blend the appropriate set of technologies, including nimble systems of engagement that sit atop systems of records, such as data reconciliation of dynamic workflows. And, yes, they can be applied in Silicon Valley too.

The big problem

But more often than not, the big obstacle to overcome is legacy systems, namely processes and ways of working that are embedded in an organization’s culture and people. Think of why bank branches are so hard to reform despite many trying. Or, why healthcare providers are so slow to generate enough of the right data to feed algorithms that could train their machines.

The implication isn’t that enterprises should wait until technology fully proves itself, as the best competitors are increasingly harnessing these technologies to secure early advantages. However, what is needed is a deeper understanding of what hinders the embedding of technology into the fabric of enterprise business processes.

What should you do? 

Companies should: 

  • Identify a range of potential projects that vary not only in technological complexity but also in organizational readiness.
  • Use human-centered design methods, like design thinking, that enable the thorough understanding of the sources of value in order to deliberately narrow the scope of intervention.
  • Explore options to rejuvenate old applications through “systems of engagement” that surround the pre-existing technology assets and complement them with a thin layer of cloud-based applications.
  • Pinpoint the datasets that require extraction from legacy databases – through a process that doesn’t only consist of identification of data sources. Instead, the best approach is called “persona-based analytics.”
  • Choose a balanced portfolio that doesn’t rely only on superior organizational transformation.
  • And, ensure that they have a team with a mix of digital experts, human-centered design thinkers, lean practitioners, and business domain experts. 

About the author

Gianni Giacomelli

Gianni Giacomelli

Chief Innovation Leader

Gianni serves as Chief Innovation Leader where he drives and sponsors Genpact’s strategic initiatives aimed at sustaining clients’ transformation into digitally-enabled companies. He also co-leads the Massachusetts Institute of Technology (MIT) efforts to set up a Collective Intelligence Design Lab.