What’s missing from your company’s A.I. transformation is a transformer

Incumbents like CBA are increasingly looking to A.I. technology to solve their business problems and are eyeing external tech partners to source those A.I. solutions. But these traditional companies have faced challenges nurturing meaningful collaborations that maximize the support they get from A.I. players. Only 1 in 5 incumbents found the right kind of A.I. player, like H20.ai is for CBA, that offers access to custom technology, as well as support for talent, training, and change management, prompting the incumbent to overhaul its processes. We call these A.I. players that provide such support transformers.

For industry incumbents that are able to identify and effectively collaborate with a transformer, the value is clear. When the BCG Henderson Institute surveyed 600 leading companies, we found that those incumbents that successfully fostered meaningful collaborations on customized A.I. solutions were three times as likely to derive a high (positive) financial impact from A.I. as those that did not. 

Incumbents currently adopting A.I. should aim to find their transformers to maximize their chances of deriving value from the technology. But there are numerous barriers to a meaningful incumbent-transformer partnership. To overcome these hurdles, incumbents need to recognize and change preconceived notions and ingrained behaviors that are holding them back.

What type of support should incumbents seek?

Incumbents should target A.I. players that help them eliminate the three key barriers that usually inhibit them from adopting A.I.: technology, talent, and change management.

Technology: Bridging legacy gaps to customized A.I. Three-quarters of incumbents we surveyed said they were challenged by a lack of tools necessary to build their own A.I. solutions. And when they looked externally for partners, 80% still had compatibility issues with their legacy IT systems. Transformers, in many instances, specialize in a particular vertical or function, allowing them to bridge an incumbent’s technology gaps by developing an A.I. solution tailored to the incumbent’s needs. Mature off-the-shelf products may be able to be more quickly adopted, but they might not fully solve the business problem, and when the technology becomes standardized, adjacent support on talent and change management also tends to fade.

Talent: Overcoming the A.I. skills deficit. Incumbents, almost universally, report facing challenges in sourcing tech talents (83%) and providing the necessary A.I. training (85%) to their current employees. Transformers eliminate this talent deficit because they work on cutting-edge A.I., which attracts top talent. Because transformer firms already specialize in the incumbents’ industry, these workers already speak the same language as the adopting incumbent workforce, facilitating the upskilling of the organization’s non-A.I. workers. 

Change management: Reinventing ways of working. The complexity of upending existing processes with new tech was a challenge for 83% of incumbents, and 76% of incumbents were challenged by their employees’ lack of trust or understanding of A.I. technology. Transformers can act as a change agent, smoothing an incumbent’s transition by charting A.I.-specific strategies, reinventing existing processes to harness the value of A.I., and establishing A.I. governance to prevent risks from materializing, safeguard responsible practices, and establish trust for users and consumers. Meaningful transformer engagement can also help fight cultural resistance, by shifting employee mindsets from viewing A.I. as a threat to seeing it as an opportunity and supporting the redefinition of job descriptions with A.I., as well as establishing trust in A.I. insights—all of which help the A.I. tools put down deep roots inside organizations. 

Yet the journey to finding—and engaging—with transformers can include pitfalls.

Navigating the A.I. marketplace itself is a fundamental obstacle for all incumbents, because A.I. falls outside incumbents’ experience and expertise; the navigation process slows down 83% of incumbents in their A.I. adoption journey. To find the right partner, incumbents need to devise a clear A.I. partnership strategy, something our survey found only a third of incumbents currently had. Additional challenges materialize depending on an incumbent’s experience with A.I. adoption, and at each stage incumbents need to change organizational behaviors to overcome new hurdles to collaboration. 

At the beginning of your A.I. transformation

When incumbents are at the early stages of their A.I. transformation (i.e., they have not yet adopted or are adopting A.I. in some processes), they have an intrinsic apprehension about working with A.I. startups or scale-ups, with half of incumbents surveyed saying such apprehension hindered collaboration. Incumbents also demonstrate a preference for mature A.I. products rather than experimentation, as 43% of early-stage incumbents cite lack of product readiness as a roadblock to engaging with A.I. partners. 

In order to foster meaningful collaborations, incumbents need to change their organizational behavior in two ways. They must see transformers as allies rather than adversaries and prioritize tailored solutions over mature products.

From fearing competition to collaborating. Incumbents must radically shift their mindset recasting transformers as strategic allies instead of adversaries. “We viewed tech companies as owners of product solutions, not as true collaborators,” an executive at a European car manufacturer said of the company’s evolution from change-resistant to embracing A.I. collaboration. “We changed and decided to set up partnerships. The tech company invested a lot into the partnership—headcounts, trainings, discounts, a lot of manpower. After two years, it became a win-win partnership. We used it not only for product consumption, but also for building up a product.”

From firewalls to open doors. This shift in mindset frees incumbents to be transparent with their data and their industry intelligence, which, in turn, enables transformers to better customize A.I. technology. When leading Norwegian drilling equipment and service provider MHWirth, now HMH, needed to conduct data-driven maintenance on its offshore drilling rigs, the incumbent gave its A.I. partner Cognite full access to its data via API key and free rein to deploy its solution. This approach helped HMH keep costs in check, extend the lifespan of equipment, and decrease downtime via customized predictive models. 

From ‘ready-made products only’ to welcomed experimentation. Instead of purchasing off-the-shelf solutions, incumbents need to embrace experimentation, particularly in industries or use cases where easily adopted mature products do not exist or don’t create a competitive advantage. Incumbents must instead place their bets on the transformers to provide customization, which takes time and requires process—and cultural—changes. There weren’t any mature A.I. products, for instance, that met Brazilian aircraft manufacturer Embraer’s needs for autonomous flight. So the incumbent sought a specialized player to develop new products, like electric vertical landing and takeoff aircraft that require technologies incorporating visual traffic detection and camera navigation. That led the company to Daedalean, an A.I. startup that possessed a wealth of knowledge and experience in autonomous flight.

From traditional pricing models to novel distribution of value. Incumbents also need to understand that experimentation and innovation might confound traditional pricing models. With a bespoke A.I. solution that creates a potentially novel distribution of value, incumbents will have to work closely with transformers to establish coherent pricing that takes into account the true value generated by A.I. A.I. tech company Bluecore did this by setting up a new monetization model when pitching its Multi-Channel Marketing Platform analyzing consumer behavior and personalizing retail marketing. The company established a pricing model based on success instead of volume—in this case, in the form of customer conversion rate or repeat purchases. That prompted incumbent retailers Foot Locker, Sephora, and Tommy Hilfiger to partner with Bluecore, embracing innovative pricing that rewarded experimentation. 

In the advanced stages of A.I. transformation

When incumbents are in more advanced stages of their transformation (e.g., they are beginning to deploy A.I. at scale), new behavioral requirements emerge that are necessary for successfully building meaningful collaborations. At this stage, incumbents will need to push beyond product scaling and embrace the organizational reinvention. In the late stages of adoption, one-third of incumbents still expressed concerns about the pricing of A.I. products, underscoring the pervasive incumbent concern over how value is distributed in the partnership. 

From product scaling to organizational change. Incumbents at advanced stages of A.I. adoption need to shift their attention to structuring collaborations to accompany scaling of A.I. That shift means involving transformers in identifying and prioritizing use cases to diffuse across the business. It also requires setting up an appropriate IT infrastructure to deploy these use cases— all as part of defining A.I. strategy at an organizational level. When Shell, for example, enlisted C3.ai to set up predictive maintenance programs for 10,000 pieces of its gas equipment, the oil incumbent empowered the startup not just to scale its insights across its business units, but explore additional use cases in production optimization, system optimization, and safety, as well as to expand into new business units such as Shell’s renewables vertical. The partnership demonstrates how incumbents can embrace a new way of thinking in their collaboration with transformers.

From legacy distribution of value to continuous redefinition. When deploying A.I. at scale with a transformer, the data and knowledge required increases—as do the potential incumbent benefits generated by A.I. This process can create data asymmetries and uneven financial benefits across A.I. use cases that can spark conflict in the partnership. To avoid this, incumbents should redefine, with their transformer partners, new ways of sharing and monetizing the value generated by A.I. at scale. An incumbent and its A.I. partners can create a new value pool by commercializing the solution originally developed for internal use, as was the case with the collaboration between C3.ai and Shell (i.e., OA.I.). 

*** 

For industry incumbents to get more out of A.I., they need to change their mindset and their behaviors to allow for meaningful collaboration with transformers.  Incumbents need to find their transformer to maximize the A.I. transformation through customized A.I. support—leaning in to experimentation, changing organizational mindsets, overcoming cultural resistance, and opening up to the uncertainty and potential of creating tailored solutions. The hurdles, as we show, are surmountable and the payoff in value is clear.

Read the study ‘What’s Missing from Your A.I. Transformation is a Transformer’.

Read other Fortune columns by François Candelon. 

François Candelon is a managing director and senior partner in the Paris office of Boston Consulting Group and the global director of the BCG Henderson Institute (BHI). You can reach him by email at candelon.francois@bcg.com.

Rémi Lanne is a project leader in BCG’s Paris office and a BHI ambassador. You can reach him by email at lanne.remi@bcg.com.

Clément Dumas is a BHI ambassador. You can reach him by email at dumas.clement@bcg.com.

Some companies featured in this column are past or current clients of BCG.