Five effective plays to become AI - first
This year’s IBM annual CEO study entitled Rewiring the C-suite: The fast track to 2030 developed in partnership with Oxford Economics makes five predictions for the future of an organization. In effect, these are five plays CEOs must make to become AI-first.
- Rethink the C-suite for speed and clarity
- Create an AI-agent flywheel
- Customize your AI mix, not just your AI models
- Orchestrate intelligence – human and artificial
- Expect unpredictable futures
Of the five, we will take a closer look at the two in bold above, both of which directly tie into the work we do – building systems for clients with a human-in-the-loop approach that customizes the AI mix to meet these challenges.
But before we do, a bit on background.
Who participated in the survey?
2000+ CEOs I 33 Geographies I 21 Industries
The report surveyed 2,000 global CEOs from 33 geographies and 21 industries between February and April 2026. As well as the survey, the report draws insights from multiple client conversations, which include a series of CEO interviews conducted between October 2025 and April 2026.
Now to the meat of it.
Customize your AI mix, not just your AI models
On the front foot
The study shows that CEOs who systematically incorporate proprietary data and IP into custom AI models and agents expect 13% more of their 2030 revenue to come from products and services not offered today.
This means adopting a hybrid strategy that in many ways differs from today’s current AI strategy. It is a multi-modal strategy that creates AI agents by combining large language models (LLMs) for reasoning with two other types – task-specific small language models (SLMs) and ultra-specialized language models (ULMs) for speed and precision.
Below is an image which shows how the current approach is expected to change between 2026 and 2030 to favor hybrids. In short, we see an increase in hybrid strategies from 13% to 50%, while pre-trained models fall to 13% from its current 39%.
The crux of the matter
Once this new strategy has been adopted, in theory at least, what steps can companies make to implement them? The study highlights four areas:
1. Draw a hardline between commodity AI and competitive AI
2. Tailor your AI portfolio
3. Make your agents unmistakably your own
4. Assign ownership for differentiated AI
The image above shows which steps need to be taken to articulate a hybrid AI strategy that delivers outcomes. It requires leadership in put in the form of sign offs and course-correcting to stop investment in the event value does not materialize.
For a hybrid approach to be successful, it also requires a complete redesign of how people cross-functional teams work together.
Orchestrate intelligence - human and artificial
The study argues in favor of redesigning how cross-functional teams work together, stating that those who have done so are more than twice as likely to have delivered on their business objectives. This is because when specialized tasks are handled by AI ‘with speed and precision, the human advantage shifts to those who can see patterns across functions.’ improvement compounds.
Our view
In our case, we see this often. Manual admin tasks become part of an orchestrated workflow that unearths information that can be used not just for communications and marketing teams but for sales teams and even HR—important information that would have nonetheless gone to waste. In fact, we will shortly be publishing a case study about how the information Briefing AI was able to support an agency with not only unearthing new clients but also supporting its sales and HR teams. Click the sign up button below and simply let us know you'd like to receive the case study using our contact form.
The crux of the matter
The study goes on to say that CEOs should focus on the interlocks that connect functions because when functions evolve independently, there are only incremental improvements. However, when redesigned as an integrated system, performance compounds. This can be achieved in the following ways.
1. Redesign workflows before redesigning jobs
2. Take an end-to-end approach to human-AI collaboration
3. Reskill for orchestration, not replacement
4. Turn AI usage into a key operating metric
