AI services, indexed by the problem you have
You know the problem — leads vanish, invoices wait, calls ring out. You do not know which service fixes it, what to ask, or who sells it here. Twelve sheets, one matcher, sixty questions, a sourced provider list.
| Code | Problem | Symptom | Effort |
|---|---|---|---|
| P-01 | Customer service | The same questions arrive on WhatsApp all day; after hours nobody answers. | Low–Medium |
| P-02 | Lead generation | Enquiries arrive from Facebook, WhatsApp and walk-ins and vanish before anyone follows up. | Medium |
| P-03 | Marketing | Posts go out when someone has time; the brochure is from 2023; nobody measures anything. | Low |
| P-04 | Sales | Quotes take two days, follow-ups depend on memory, and the pipeline is a group chat. | Medium |
| P-05 | Operations | Dispatch runs on phone calls, stock is counted on paper, and the daily plan lives in one head. | Medium–High |
| P-06 | Finance | Invoices wait for one person, EcoCash and bank lines are matched by hand, month-end takes a week. | Medium |
| P-07 | Administration | The inbox is the to-do list, forms are retyped three times, and finding last year's contract takes an afternoon. | Low |
| P-08 | HR and recruitment | Two hundred CVs for one vacancy, the same leave question every week, and onboarding by word of mouth. | Medium |
| P-09 | Reporting | The monthly report takes a week, three spreadsheets disagree, and management decides on last quarter's numbers. | Medium |
| P-10 | Document processing | Invoices, delivery notes, IDs and forms arrive as photos and are typed into systems by hand. | Medium |
| P-11 | Voice | Calls go unanswered at peak, the IVR is a maze, and customers send voice notes nobody transcribes. | Medium–High |
| P-12 | Computer vision | Counting, grading and gate checks are done by eye; incidents are found on the CCTV afterwards, if at all. | High |
Read, match, ask, shortlist
| Step | What you do | Where |
|---|---|---|
| 1 | Find the sheet for your problem. Each one lists symptoms, what AI can and cannot do, service shapes, vendor questions and an effort band. | /problems/ |
| 2 | Match problem, company size and budget band to the service shapes that fit — and the ones that do not. | /match/ |
| 3 | Take the questions to the vendor meeting. Sixty, in eight risk areas, copyable. | /vendor-questions/ |
| 4 | Shortlist from providers who actually sell that shape in Zimbabwe — sourced and dated, ownership disclosed. | /providers/ |
| 5 | Choose with a weighted scoring matrix and a pricing model that matches the risk. | /how-to-choose/ · /pricing-models/ |
Six things you can actually buy
Audit
Inspect the process, the data and the constraints before anyone builds anything.
Low–Medium · 1–3 weeks
Pilot
Prove one process on your real data in weeks, with a stop-or-scale decision at the end.
Medium · 3–8 weeks
Implementation
Build, configure and hand over a working system your staff run without the vendor.
Medium–High · 4–16 weeks
Integration
Connect the AI to the systems that hold your stock, customers, money and invoices.
Medium–High · 2–10 weeks per system
Training
Teach the people who will run the system — often the cheapest fix on this catalogue.
Low · 1–2 days per group, plus follow-up
Managed support
Keep the system working after go-live — monitoring, fixes, retraining, changes.
Low (monthly) · Ongoing; 3–12 month terms
Who sells what, sourced
12 providers listed across 5 categories, each with the services they publish, listed prices where they exist, a source URL and the date we checked. Two are related to this site's publisher and say so on the listing.
- Off-the-shelf platform — 2
- Implementation and automation — 5
- Compute, tooling and telco AI — 1
- Compute and hosting infrastructure — 3
- Language tooling — 1
Vendor-neutral by construction
The problem sheets do not recommend a company. The publisher's own implementation company appears in the listing on the same terms as everyone else. Effort bands are ranges, not quotes. Every Zimbabwean fact on this site carries a source; every price is the provider's own listed price on the date shown.
For a directory of AI companies in Zimbabwe, see the Zimbabwe AI directory. For what automation could save you, use the ROI calculator at businessai.co.zw. This site starts after those questions.
Sourced, dated, Zimbabwe-specific
Buying a WhatsApp customer-service assistant in Zimbabwe: a walkthrough
The channel, the official-API question, four ways to buy, the licence you must hold, and a four-week pilot plan.
What an AI service actually costs in Zimbabwe: the six-layer cost stack
Quotes hide six layers. Here they are, with the local prices that are actually published, and a worked twelve-month example.
SI 155 of 2024 and your AI vendor: who is the controller, who is the processor, and what goes in the contract
Licensing in one page, then the controller/processor split and the nine clauses your AI contract needs.
Document AI and ZIMRA fiscalisation: where automation fits and where it must stop
FDMS in plain language, the compliant flow drawn out, and how to test extraction on your own invoices.
Where should your AI run? Power, connectivity and hosting choices for Zimbabwean businesses
Grid, fibre and forex decide more than the model does. A four-question decision table for edge, local cloud or foreign cloud.