In short: FYAdvisers reviews processes, data, tools, people and risk, then hands you a prioritised list of AI use cases, each marked build, buy or skip, with the reasoning, what it needs, the main risks and a suggested first step. A use case with no clear reviewer is treated as not ready.
An AI readiness assessment tells you where AI is worth using in your business, where it is not, and what has to be in place first. FYAdvisers runs a fixed-scope assessment for small and mid-sized B2B businesses in Singapore. We review five areas (processes, data, tools, people and risk) and finish with a prioritised list of use cases, each marked build, buy or skip. It is the first step of our wider AI consulting work, but the output stands on its own.
Why assess before building
Most AI projects that disappoint were aimed at the wrong task, or at a task whose data was not ready. An assessment checks the basics before money goes into a build: is the task frequent enough to matter, is the input available in a usable form, can a person realistically review the output, and does a tool you already pay for cover it?
Many general readiness indexes and surveys score a company’s maturity. That can be useful for large organisations. A smaller business usually needs something more concrete: which tasks, in what order, and what to do about each one.
What we assess
Processes
We map the work where people spend time reading, sorting, searching, checking or drafting: enquiries, quotations, tenders, invoices, reports, follow-ups. For each, we note how often it happens, how long it takes, how much the input varies, and what a mistake would cost. Predictable steps are usually better served by rules or workflow automation than by AI.
Data
AI is only as useful as the information it can reach. We look at where the relevant data lives (email, WhatsApp, shared drives, a CRM, accounting software, paper), how consistent it is, whether past examples exist to test against, and whether it contains personal data.
Tools
We list the software you already use and check which AI features are already included or available, such as those in Microsoft 365 Copilot, Gemini in Google Workspace, ChatGPT’s business plans, or your CRM and helpdesk. We also check whether your systems can be connected, since a use case that needs data from three systems without APIs is a bigger job than it first looks.
People
We talk to the people who do the work and the people who would check AI output. Who would review drafts, how much time they have, how comfortable they are with the tools, and who owns the process once it changes. A use case with no clear reviewer is not ready.
Risk
We look at what could go wrong: wrong information reaching a customer, personal data sent to the wrong place, an AI suggestion treated as a decision. We note which steps must stay with a person, such as prices, commitments and advice, and what logging is needed so mistakes can be traced.
What you receive
A prioritised list of AI use cases for your business. For each one:
- A recommendation: build, buy or skip. Build where the task depends on your own process or data. Buy where an off-the-shelf tool already fits. Skip where the effort, data gaps or risk outweigh the benefit, or where plain rules would do.
- Why. The reasoning, in plain language.
- What it needs. The data, integrations and review step required.
- Main risks and how to manage them.
- A suggested first step, usually a small prototype tested on your real examples.
We walk through the list with you so it can be acted on, whether by us, another provider or your own team.
How the assessment runs
- Scoping call. We agree which teams and processes are in scope and who we should speak to.
- Interviews and walkthroughs. We watch how the work is done and collect a few real examples, with personal details removed where possible.
- Systems and data review. We look at the tools involved, how data moves between them and what can be connected.
- Quick tests where useful. For promising use cases, we may run an AI model on a small set of your examples to check the idea holds up before recommending it.
- Findings and walkthrough. We deliver the prioritised list and go through it with you.
Data and governance
Where personal data is involved, we check each use case against the PDPA. PDPC’s Advisory Guidelines on use of Personal Data in AI Recommendation and Decision Systems, issued on 1 March 2024, cover consent, notification and accountability when personal data is used in AI systems, and how service providers should support the organisations they work for.
For generative AI, IMDA and the AI Verify Foundation published the Model AI Governance Framework for Generative AI on 30 May 2024, which sets out nine dimensions including accountability, data, testing and assurance, and security. The AI Verify Foundation also maintains the AI Verify testing framework, which sets out 11 principles for assessing whether AI systems are implemented responsibly. These frameworks matter most once a system is built and running. In an assessment we use them as a reference for the risk questions to ask early.
What drives the scope
- Number of teams and processes reviewed.
- Number of systems to examine and how well documented they are.
- Whether quick tests on real data are included.
- Sensitivity of the data involved, which affects how much risk review is needed.
Who this suits
- Owners and managers who suspect AI could help but do not know where to start.
- Businesses already paying for AI tools who want to know if they are being used well.
- Teams that have been pitched AI products and want an independent view on build, buy or skip.
It suits less well if you already know the task and want it built. In that case, start with AI integration or AI document processing.
Examples from our work
Our past builds show the kind of use case an assessment tends to surface. For Chromatic Media, a tender preparation workflow reads uploaded tender files and produces a compliance matrix and proposal drafts for the team to check. For an individual PropNex property agent, a workspace groups WhatsApp chats into deals and prepares reply drafts and translations the agent reviews.
What to prepare before we talk
- The processes or teams you want looked at.
- The tools and systems you use today.
- Any AI tools already in use, and how.
- A few examples of the work, with personal details removed.
- Any rules about where your data can be processed.
Get in touch and we will suggest how to scope the assessment.
Frequently asked questions
What is an AI readiness assessment?
It is a structured review of whether a business is ready to use AI, and where. It looks at how work is done, what data exists, which tools are in use, who would use and check the output, and what could go wrong. The useful result is a short, ranked list of specific use cases rather than a general score.
What do we get at the end?
A prioritised list of use cases, each with a recommendation to build, buy an off-the-shelf tool or skip it, the reasons, what data and integrations it would need, the main risks, and a suggested first step.
Do we have to hire you to build afterwards?
No. The list is yours to use with any provider or in-house team. Where an off-the-shelf tool is the better answer, we say so.
Is this the same as AI Verify?
No. AI Verify is a testing framework maintained by the AI Verify Foundation for assessing whether AI systems are implemented responsibly. Our assessment comes earlier: it decides which AI use cases are worth pursuing. Governance frameworks such as AI Verify become relevant once a system is being built and deployed.
How long does it take and what does it cost?
It depends on how many teams and processes are in scope and how many systems we need to look at. We agree the scope and outline pricing after a first conversation.