EthosPowerAIA tender arrives with your client's own IEEE 485 battery sizing, schematics and BOM attached. They authored almost everything they need themselves, using their favourite LLM, thus eroding the value your firm once added.
This is the new reality: your customers have changed and so must your strategy. Today your customers know a lot more than they used to, tomorrow even more; so they may not need your firm. The knowledge-gap that gave your firm its margin is closing and will not reopen.
Elevate the value
Because basic services are becoming commoditised by AI, organisations must elevate their value — from supplier to trusted consultant and stakeholder.
AI by itself, for its own sake, actually drains productivity. Targeting AI at low hanging fruit works for a while. Instead, a firm should use its vast data access to actively consult the client — such as acting as a "QA as a service" to assess the reliability of a client's supply chain — thereby actively ensuring the client's business thrives.
Let the business owners build the prototype
A highly effective adoption strategy is empowering business owners to build their own software prototypes (MVPs) using AI. This allows business leaders to define exactly how the business logic should work before handing it over to engineers, effectively removing the historical translation barrier between business needs and technical execution.
While applying technology is only the "third and minor part" of the transformation process, it should be highly iterative. Playing with tools shouldn't be done just to automate a task, but to help business leaders crystallise their understanding of what the industry will actually look like in five years.
Screening the expert
If a firm is seeking to hire an expert to guide this transformation, they must screen candidates against a highly specific and rigorous set of criteria. Afanasyev warns there are only about 200 mature "applied AI experts" globally.
The "anti-fungible" requirement — decades of experience. The right consultant cannot be a professional who simply jumped on the AI bandwagon after a two-week course because their previous field (like ESG or blockchain) cooled down. You must seek individuals with years or decades of accumulated experience, as mastering this space takes immense time and cannot be mass-produced.
The uncompromisable "triad" in a single mind. The consultant must possess deep expertise across three domains: business operations, organisational transformation, and broad technology. A firm cannot simply hire three different specialists and put them in a room; the interconnections are so complex that the knowledge must be unified within a single person's head.
Demands a CEO-level mandate. A true applied AI expert will not accept a mid-level implementation role. To be effective, they must insist on reporting directly to the Chief Executive Officer or sitting on the board. Furthermore, they must have the explicit mandate to command IT and Operations resources, as they cannot execute structural change without operational authority.
Focuses on mechanism design over code. The right consultant will explicitly view the actual AI technology as the "minor part" of the journey. Their primary focus will be on "mechanism design" (macroeconomics) — redesigning the firm's core value proposition to survive in an ecosystem where clients and adversaries are now smarter than the firm itself.
Structures business-led trial and error. Because mature talent is scarce, a great consultant will establish internal incubators to nurture future AI translators from within the company. Crucially, they will ensure these experiments are led by strategic business thinkers looking five years ahead, explicitly rejecting the traditional model of letting isolated engineering teams randomly test tools.
The argument is Maxim Afanasyev's, Financial Services Industry Head for Asia Pacific and Japan at Google Cloud. His interviews, the sources for the quoted phrases, and my own reading of what they mean for battery-backed power are set out on the Approach page.