Observe ZDNET: Add us as a most well-liked supply on Google.
ZDNET’s key takeaways
- Profitable firms refine their methods to information AI explorations.
- You can begin tons of of initiatives, however amount should develop into high quality.
- Contemplate key parts, equivalent to structure, redundancy, and targets.
With many enterprise leaders struggling to create a aggressive benefit from AI, your organization will want a well-thought-out technique to beat the remainder.
For Artwork Hu, international CIO at tech large Lenovo, there aren’t any half measures. “We wish AI to penetrate all features of our enterprise,” he informed ZDNET in a one-to-one dialog at a resort in London, suggesting his firm has created a top-down and bottom-up dedication, the place workers are inspired to discover AI in a tightly ruled and safe method.
Additionally: 5 methods to forestall your AI technique from going bust
Key use instances embrace dialog summarization to help help specialists, making use of agentic AI to enterprise-grade software program engineering, and utilizing gen AI to create efficient advertising collateral.
Hu and his workforce have constructed a cautious technique to form their AI explorations — and listed below are 5 methods you are able to do the identical.
1. Undertake a portfolio strategy
Hu mentioned Lenovo takes a portfolio-based strategy to rising expertise throughout the broad AI product lifecycle.
“That strategy stretches all the best way from ‘I’ve heard of AI, and I am enthusiastic about it,’ via to ‘I am taking part in with it in a sandbox,’ to ‘I’ve deployed it to my division,’ to ‘Hey, tens of 1000’s of individuals on the firm use this device,” he mentioned.
This portfolio strategy means the corporate has greater than 1,000 registered initiatives throughout all enterprise areas.
Additionally: This firm’s AI success was constructed on 5 important steps – see how they be just right for you
Hu has partnered with the agency’s chief safety officer and chief AI officer to ascertain a coverage for AI exploration that has shifted from letting 1000’s of flowers bloom to offering cautious steering.
“We wish individuals to bloom and discover, however we have to management the chance, as a result of there’s fairly a protracted tail of sudden outcomes, particularly if you happen to’re not cautious,” he mentioned.
“You really wish to be able the place individuals are very excited, the place they’re pushing you on approving extra initiatives.”
2. Hone your working mannequin
Hu mentioned the important thing implication of encouraging individuals to discover AI is that IT should be run in a different way.
“Up to now, individuals used to return to IT, and we would run a extremely centralized structure,” he mentioned. “The best way it interprets for us in an age of AI is that we have needed to rethink how we work with the enterprise.”
Hu: “We wish AI to penetrate all features of our enterprise.”
Lenovo
Quite than his workforce taking non-IT professionals’ necessities and turning them into methods, whether or not via off-the-shelf or self-developed functions, Hu’s expertise group should now fastidiously handle a brand new space of demand: generative and agentic AI.
Additionally: 3 methods AI brokers will make your job unrecognizable within the subsequent few years
“As an alternative of a small subset of workers working in IT transformation and constructing methods, the barrier is decrease, and there is potential for the whole firm to contribute to digitalization and clever transformation,” he mentioned.
“If the entire firm desires to get entangled in AI and transformation, how can we handle that? It is a very completely different working mannequin for the guardrails and pointers.”
3. Embrace spare capability
Hu mentioned it is also vital to think about the idea of redundancy in an age of AI. Beforehand, and doubtless for a lot of the previous twenty years, enterprise leaders prioritized globalization and centralization.
Now, volatility and uncertainty characterize fashionable financial and technological practices, and that new regular means CIOs should undertake a recent strategy.
“We’re architecting and pivoting our methods so that they will be rather more regional, and that is due to elevated concentrate on information sovereignty and information privateness by the essential regulatory regimes from China to the EU to the Americas,” he mentioned.
This new strategy to enterprise structure begins as a enterprise technique, with in-built shock absorbers that present redundancy throughout regional variabilities.
Additionally: Dreading AI job cuts? 5 methods to future-proof your profession – earlier than it is too late
In brief, spare capability which may as soon as appeared wasteful is now a smart technique for firms working in an AI-enabled financial system.
“With redundancy, if one thing occurs in a single half, your entire system does not get paralyzed or develop into unable to function,” he mentioned.
“Usually, within the hyper-efficient world, redundancy is unhealthy. However I believe via COVID and present volatility, the pure response is, ‘Properly, perhaps buffers are good,’ as a result of you do not know what is going on to occur essentially sooner or later.”
4. Create a scoreboard
Hu mentioned one other key factor of Lenovo’s AI technique is that every govt committee member has AI targets.
“So, we’re all dedicated to adopting AI,” he mentioned. “That has an attention-grabbing competitors impact, as a result of we put up a scoreboard: ‘How is everybody’s space doing?’ And you do not wish to be slower than the particular person subsequent to you on the manager workers.”
Additionally: I take a look at AI for a residing, and these 3 free instruments are those I used most in 2025
Hu mentioned the competitors is wholesome as a result of the ensuing dynamic is that professionals in capabilities, equivalent to advertising, gross sales, or HR, are all enthusiastic about the right way to use AI productively.
“We take into consideration our enterprise in a really structured, systematic, and detailed approach. We have now quantitative targets mapped very particularly to degree one, degree two, and degree three achievements,” he mentioned.
“From a lifecycle and value-chain perspective, we’re making an attempt to supply as a lot transparency as we are able to. And it is enjoyable if you happen to play the time lapse of lighting up this map of the corporate and seeing the way it’s going.”
5. Whitelist nice instruments
After all, tempo is essential with regards to growing AI initiatives, however so is high quality. Hu mentioned that Lenovo tries to not over-prioritize high quality initially.
“I am not going to fret about small quantities of funding on the early stage as a result of we wish to construct a funnel,” he mentioned.
“Early on, an important factor is, ‘Did we study one thing?’ High quality issues much more if you wish to productionize AI for tons of or 1000’s of individuals. For 5 or 10 individuals, if it really works and also you study one thing, that is good.”
Additionally: Gemini vs. Copilot: I in contrast the AI instruments on 7 on a regular basis duties, and there is a clear winner
Hu mentioned transferring to a concentrate on high quality requires clear course on necessities and instruments. One essential factor for fulfillment has been establishing governance to chop via a few of the huge uncertainty and noise round AI instruments.
“What that strategy interprets into for us is that we have ended up whitelisting a set of instruments to say, ‘We expect these will meet 85% to 90% of your necessities if you happen to’re beginning to discover AI,'” he mentioned.
“However there is not any assure these instruments do every little thing, so we even have an ongoing course of to introduce instruments. Individuals can say, ‘Hey, I’ve a use case, and not one of the instruments that you simply’re offering are well-suited. Can we get this device on the whitelist?’ After which we’ll do the suitable critiques and assessments to have the ability to introduce that device.”
