Frame & Field Studios Business Plan — Executive Summary
A commercial photography and content studio: R11.86m funding, 7.5 photographers, 1,195 billable days and R27.96m Year 5 revenue.
Executive Summary
Jump to section
- Overview & contents
- i. Important Notice and Basis of Preparation
- 1. What Artificial Intelligence Has Taken, and What It Has Not
- 2. Executive Summary
- 3. Service Lines, Clients and Pricing
- 4. Capacity: The Shooting Day and the Editing Day
- 5. SWOT and Competitive Position
- 6. Organisation, Rights and Compliance
- 7. Financial Plan
- 8. Break-Even and Debt Service
- 9. Investment Analysis
- 10. Sensitivity and Scenario Analysis
- 11. Risk Analysis
- 12. Implementation Roadmap
- 13. Key Performance Indicators
- 14. Key Assumptions
- 15. Conclusion and Recommendation
- A. Appendix A: Consolidated Financial Summary
- B. Appendix B: Capacity, Mix and Cost Schedules
- C. Appendix C: Funding, Debt and Working Capital Schedules
- D. Appendix D: Risk Register
- E. Appendix E: Glossary
- 2.1 What changed
- 2.2 The exposure framework
- 2.3 Using the technology rather than competing with it
2.1 What changed
Two things happened quickly. Generative image tools became good enough for commercial catalogue use, and the platforms that host most product listings began offering those tools directly to sellers. The economics are not close: a generated image costs a few rand where a photographed one costs hundreds or thousands.
|
Measure |
Figure |
Implication |
|---|---|---|
|
AI-generated image cost |
US$3 to US$12 |
R54 to R214 an image; a few cents at subscription scale |
|
Traditional photography cost |
US$85 to US$250 per stock keeping unit |
R1 517 to R4 463, including retouching |
|
Traditional footwear shoot |
US$5 000 to US$15 000 |
R89 250 to R267 750 for a single production |
|
Creative professionals using generative AI |
83% |
Adobe; 20% say employers or clients require it |
|
Marketers using AI for image assets |
62% |
Salesforce |
|
E-commerce operators budgeting for AI imaging |
67% |
JungleScout; adoption is mainstream, not experimental |
|
Apparel listings AI-generated by end 2026 |
40% |
And virtual model adoption forecast at 89% in fashion e-commerce |
|
In-house teams retaining photographer-led shoots |
30% |
The remainder moves to AI; the photographer directs rather than produces volume |
|
South African e-commerce market |
US$41.86 billion in 2026 |
Growing at 8.54% a year |
|
Marketplace sellers |
18 000 Takealot, 10 000+ Amazon SA |
The trap: a large addressable market at an unsurvivable price |
2.2 The exposure framework
|
Service line |
Share of days |
AI exposure |
Why |
|---|---|---|---|
|
Corporate and executive portraiture |
26% |
10% |
Real, identifiable people who must be photographed |
|
Property and architectural |
22% |
15% |
A listing image must depict the actual property |
|
Events, conferences and documentary |
20% |
5% |
A moment that happened cannot be generated afterwards |
|
Brand, editorial and campaign |
17% |
45% |
The creative-direction work that survives; partly AI-assisted |
|
Motion and video content |
11% |
35% |
Editing-intensive; the fastest-growing client request |
|
Catalogue and product |
4% |
85% |
Deliberately minimal; this is the work AI is taking |
|
Weighted exposure, day-weighted |
100% |
21.8% |
Against 54.0% on a conventional mix |
|
Weighted exposure, revenue-weighted |
+25.7% |
R7 185 528 of mature revenue at risk |
Two ways of weighting the exposure are shown deliberately. Day-weighted exposure of 21.8 per cent measures how much of the diary is at risk; revenue-weighted exposure of 25.7 per cent measures how much income is. The second is higher because the exposed lines — brand at R33 500 and motion at R38 500 — carry above-average day rates, while the protected lines of portraiture and property sit below the blended rate. An investor should use the revenue-weighted figure when sizing the exposure, and the day-weighted one when thinking about how much of the team’s time is affected.
2.3 Using the technology rather than competing with it
|
Application |
What it does |
Why it matters here |
|---|---|---|
|
Retouching and clean-up |
AI-assisted masking, background work and object removal |
Reduces editing time, which is the constraint identified in Section 4 |
|
Variant generation |
One photographed scene extended into seasonal, format and market variants |
Sold as a package rather than as additional shoot days |
|
Culling and selection |
Automated first-pass selection from a large event take |
Pure time saving on the highest-volume line |
|
Creative direction of generated assets |
Directing and quality-controlling generated imagery for clients |
Keeps the brief rather than losing it entirely |
The plan budgets R268 000 a year for AI tooling and licences. A studio that refuses to use these tools will be undercut by one that does; a studio that relies on them for its whole offering has no defensible position. The intended posture is to use AI to compress cost on the work it keeps — and specifically to attack the editing ratio, which Section 4 identifies as the binding constraint on output.
One governance point belongs here. Where generated imagery is supplied to a client, ownership, training-data provenance and indemnity should be addressed in writing, and emerging disclosure requirements for generated imagery in advertising should be tracked. Section 6 carries this in the compliance table.