Ting Guides · For the CEO
How do you identify and prioritize AI use cases?
The short answer
Start from the work, not the technology. Trace a few high-volume processes end to end, mark where work is retyped, waits or gets sent back, and write each candidate as a specific task with an owner. Then score every candidate on value, feasibility and risk, and fund the one that holds up on all three.
Reviewed by Badr Al-RashedUpdated
What it depends on
Volume and repetition
AI pays back where the same judgment is made many times: reading the same kinds of document, answering the same questions, routing the same requests. A task that happens twice a month rarely justifies a build.
Whether the data exists
A candidate that needs data you don't capture is a data project first. RAND's interviews with 65 experienced data scientists and engineers list lacking the data needed to train an effective model among the leading root causes of AI project failure.
How exceptions are handled today
If the rules live in one experienced person's head, write them down before scoring. Feasibility depends on how many exceptions there are and who resolves them.
Tolerance for error
Drafting a reply that a person checks tolerates mistakes; posting to the ledger doesn't. The risk decides the design, from fully automatic to a person reviewing every item.
How many systems it crosses
A candidate inside one system is cheaper than one that crosses email, WhatsApp, the ERP and a shared drive. Count the systems before you estimate anything.
Where Saudi operations tend to find candidates
General information, not legal advice. Each regulatory point summarises the official source cited beside it, as checked on the date shown under Sources; confirm how it applies to your company with your legal adviser.
Arabic documents and phone photos
Contracts, permits, delivery notes and supplier documents arrive as Arabic PDFs, scans and phone photos. Extracting and checking them is a frequent and well-bounded candidate. Judge feasibility on your worst scans, not your cleanest.
WhatsApp as the real front office
Orders, complaints and approvals often arrive on WhatsApp and get retyped into a system by hand. That retyping is frequently the clearest place the work gets stuck in the whole process, and the easiest to measure.
The order SDAIA describes
SDAIA's AI Adoption Framework describes effective adoption as identifying opportunities and challenges, designing solutions, implementing them, and continuously evaluating outcomes. Prioritization is the first step's output, not a step you can skip.[2]
Five candidates from one after-sales team, scored
ExhibitWorked example
Rasikh Motors, the car dealer of the fictional Rasikh Holding, with nine service centres, scored five after-sales candidates from 1 to 3 on impact, ease and low risk, so more dots always mean better. Fund what holds up on all three; park the rest with a written reason.
Impact
Ease
- Low risk
- Medium risk
- High risk
| Candidate | Impact | Ease | Low risk | Decision |
|---|---|---|---|---|
| Route WhatsApp service requests to the right centre and job card | Fund first | |||
| Answer "is my car ready?" from the job-card status | Next in line | |||
| Draft the repair explanation a service advisor sends the customer | Small trial, every draft reviewed | |||
| Predict which parts each service centre will need next month | Park: parts usage isn't recorded the same way in every centre yet | |||
| Approve warranty claims automatically | Redesign: a person approves; AI pre-checks the claim file |
The decisionFund the WhatsApp routing first: high impact, easy to build, moderate risk. Warranty approvals stay with a person; AI only checks the claim file before them.
What Ting does and doesn't do here
What we do
- We trace how work actually moves through documents, approvals and systems, and find where it repeats, waits or gets sent back.
- We write each candidate as a specific task with its data, systems and owner, and rank the candidates by impact and ease with every assumption written down.
- We tell you plainly which candidates should not become AI projects at all.
Where Ting helps: AI Transformational Audit
What we don't do
- We don't audit your finances or your AI models. This studies how your operation runs.
- We don't make the audit a condition. A single, clear use case can go straight to a Solution Blueprint.
- We don't promise savings figures. The expected return comes with every assumption written down, so you can check it.
Sources
Primary sources only, each read in full on the date shown.
- The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed (RR-A2680-1, 2024)RAND Corporation · checked
- AI Adoption Framework, version 2 (May 2025)SDAIA · checked
