“Zero to one” can mean making a first mark on a blank page—or getting an organisation from AI experiments to its first useful deployment. In both cases, starting is only the beginning: the harder work is shaping an initial result into something that meets a real need and can be trusted. For a business, that means fitting AI to a workflow, measuring whether it helps and deciding where people must remain involved.
What “zero to one” means for AI
There are two distinct starting points. For an individual, it is overcoming the blank page: finding a first idea, draft or visual direction. For an organisation, it is moving from no practical AI use to a first use case. The second does not mean that an organisation has successfully adopted AI at scale; a promising trial and a dependable production workflow are different achievements.
Neither sense has one universally hardest step. AI can help create a starting point, but the person still has to judge and develop it. At work, a tool may produce a useful demonstration quickly, while adapting it to actual conditions, establishing oversight and showing that it improves the process takes more effort.
How AI can help with a creative start
Accenture’s Life Trends 2023 describes “the hardest part of the creative process” as “going from zero to one”—making the first mark on a blank page or canvas. The report argues that neural networks can help people get started and that creators can build layers on top of the initial output. That is a perspective on AI’s creative role, not a guarantee that a generated draft will be accurate, original or good enough to use. Read Accenture’s Life Trends 2023.
#1 Best Overall
The practical value of a generated first pass is that it gives a person something to react to: an outline to revise, a set of options to compare or a rough draft to reshape. Treat it as material for human judgment, not a finished answer. Accenture’s report also discusses faster content creation and adaptive content, but it was published in 2023 and does not establish that every tool or workflow produces better creative work.
Why a first AI use case is not the same as successful adoption
For organisations, initial use cases such as internal chatbots, coding assistants, content-generation tools and data analysis can be stood up relatively quickly, according to the UK government’s Digital and Technologies sector plan. The bigger challenge is making a promising demonstration dependable in production. The plan describes that work as moving to “a reliable production use case with clear success metrics, process changes and human oversight.” Read the UK Digital and Technologies sector plan.
Rank #2
This transition is not simply a matter of switching on a tool. The OECD, BCG and INSEAD’s 2025 report says companies need to invest time and resources to tailor each AI use case to their needs and conditions. Projects also involve experimentation, so return on investment is uncertain. A pilot can reveal whether an idea is promising, but it cannot by itself establish that the use case will deliver lasting value. Read the OECD, BCG and INSEAD report.
The UK plan also identifies skills and management capability as commonly cited barriers, and points to experimentation and rapid learning as useful responses. That makes the first use case partly an organisational learning exercise: people need to understand the tool, the workflow it affects and the conditions under which its output needs review.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
What adoption figures do—and do not—show
Different reports use different populations and measures, so their figures should not be combined into a single measure of AI adoption. OpenAI’s 2025 enterprise report draws on de-identified enterprise usage data from OpenAI and a survey of 9,000 workers across almost 100 enterprises. It identifies organisational readiness and implementation as primary constraints in its analysis. The report also says more than 1 million business customers used OpenAI’s tools, that ChatGPT message volume grew 8x, and that API reasoning-token consumption per organisation increased 320x year over year. These are OpenAI-specific measures, not counts or growth rates for all businesses using AI. Read OpenAI’s 2025 enterprise report.
Separately, McKinsey & Company’s 2025 global survey found that nearly two-thirds of surveyed respondents said their organisations had not begun scaling AI across the enterprise. That is a survey finding shaped by its sample and question wording, not a census of organisations or a universal rate. It does, however, illustrate why trying AI and scaling it across an organisation should not be treated as the same milestone. Read McKinsey’s 2025 State of AI survey.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to choose and test a first workplace use case
The following is a practical way to apply the issues raised in the OECD report and UK government plan; it is not an official scoring framework. Start with a bounded task in an existing workflow, then test whether the tool helps under realistic conditions.
- Choose a specific task. Define the work to be improved rather than adopting AI as a goal in itself. Identify who does the task, what information it uses and what a successful result must do.
- Check workflow fit and tailoring effort. Consider whether the tool can work with the task’s inputs and constraints, and estimate the time, resources and expertise needed to adapt it.
- Define success before testing. Choose a measure connected to the actual workflow, such as whether the task is completed to the required standard or whether the process improves. A demonstration that looks impressive is not a substitute for a clear measure.
- Set human review and process changes. Decide which outputs require checking, who is responsible for that review and how the workflow changes when AI is introduced.
- Run a limited experiment and record results. Test the use case on appropriate work, have people assess the output and note the effect on the process, including the effort required to tailor and review it.
- Decide whether to adapt, stop or expand. Use the results to determine whether the use case meets its criteria and is worth further investment. Do not assume that a successful trial will transfer unchanged to other teams or tasks.
When comparing candidate tasks, weigh workflow fit, tailoring effort, measurable success, potential value, available skills and management capacity, required process changes, and the level of human oversight. A task with high potential value may still be a poor first choice if it is hard to measure, requires extensive adaptation or leaves no clear way to review the output.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Best Value
When to get help building AI capability
If a team lacks the skills or management capacity to choose, test and oversee a use case, relevant AI adoption training or implementation support may help address that gap. The UK government plan identifies skills and management capability as barriers, but it does not establish that every organisation needs an outside provider. Choose support according to the capability the team is missing and the work it needs to do.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
