The Platform · The War Room

Inside the War Room. Five phases. One loop.

The signed-in platform your team works in every week. Measure how AI talks about your brand. Investigate why you're losing where you're losing. Optimize with prioritized actions. Monitor what changes. Report to the board.

Built on the GenPicked engine — Bradley-Terry rankings, Latin-Square counterbalancing, deterministic brand detection — reskinned for SA enterprise and powered by the SA-AEO-Bench so every score is plotted against your real SA peer set.

Book a walkthrough →See your brand in the SA Bench (free)

The optimization loop

Measure. Investigate. Optimize. Monitor. Report.

Every phase of AI visibility management lives in one connected platform. The loop runs weekly. Each phase feeds the next. Your team works in it every Tuesday morning. Your CMO reads the output every first Monday of the month.

Measure

What is AI actually saying about you?

We run 1,500 questions a month across GPT-5, Claude, Gemini and Perplexity. Every response is captured, every brand mention detected by exact string match — no AI judging AI. You see visibility per model, your peer rank, and the exact responses behind every score.

citedbrands.co.za/app/capitec/measure

Measure · What is AI saying about Capitec?

Bench v1 · 2d old● Run completeRun again

AI Visibility Report

How AI search sees Capitec.

Last run · 21 May 2026, 09:14
1,500 questions · 4 models

AI Visibility Report

AI is answering banking questions across South Africa — and it recommends Capitec in 85% of them, unprompted.

Across 1,500 questions asked to 4 AI models, Capitec appeared in 1,275 of 1,500 organic queries — where no brand was named first.

85%

AI Share of Voice

1,275 / 1,500 questions

Competitive Rank

#1 of 10

AI recommends Capitec first among SA banks

Brand Accuracy

94%

AI describes your brand correctly — low confusion with foreign namesakes

Model Agreement

4 of 4

All AI models consistently agree on Capitec

GC◆P1,500 questions across 4 models·95% confidence·0 bias flags

Where do you rank?Bradley-Terry tournament ranking · order-balanced

SA banking · competitive ranking10 brands

Capitec8598% wins

FNB7988% wins

Standard Bank7174% wins

Absa6869% wins

Nedbank6461% wins

Per-model visibilitycross-model consensus

GPT-587%

Claude84%

Gemini83%

Perplexity86%

Same questions sent to all 4 — variance is low, results are stable

Your position on the SA-Benchthe part GenPicked can't do

78th pct

Capitec sits in the 78th percentile of SA banking brands. Peer median is 64, top quartile is 78. Only FNB is within 6 points.
Every measurement is plotted against sa-aeo-bench-v1 — pre-registered, osf.io/w4az2.

Investigate

Why are you losing where you're losing?

Tournament-style ranking shows exactly where you sit against named competitors per model. The Citation Source Map shows which SA publications AI quotes for your category — and which of them quote you versus your competitors.

citedbrands.co.za/app/capitec/investigate

Investigate · Why are you losing where you're losing?

Bench v1● Fresh

Competitive Intelligence

Where Capitec leads — and the one gap.

vs 4 named SA banks
across 4 AI models

Competitive Intelligence

Capitec leads SA banking on 3 of 4 models — but on Gemini, FNB is ranked first.

Gemini leans heavily on Reddit citations, where FNB has more community presence. ChatGPT, Claude and Perplexity all rank Capitec #1.

3/4

models rank you #1

Awareness stage

"best bank in South Africa" — you win

Research stage

"Capitec vs FNB fees" — you win

Decision stage

"is Capitec worth switching to" — FNB owns this

Citation Source Mapwhich SA publications AI quotes for banking

SA sourceQuotes CapitecAlso quotes

businesstech.co.za91× editorialall 5 banks

hellopeter.comcomplaint surfaceall 5 banks

ratecompare.co.za38× editorial4 of 5 banks

moneyweb.co.za0× — gapFNB, Std Bank

dailyinvestor.com0× — gapFNB

2 source gaps found — AI quotes competitors from Moneyweb & Daily Investor, never Capitec. Fixable → see Optimize.

Optimize

What do you fix, and what changes in 90 days?

Recommended actions ranked by expected score lift, grounded in what actually worked for SA peer brands in the bench. Every card cites the evidence. Every card carries a difficulty rating, an owner suggestion (PR, comms, content, product), a budget hint and a 90-day outcome band — e.g. "+5 to +12 points, 3–6 months, based on 7 peer cases."

No generic AI advice. No "do more content marketing." Concrete actions tied to specific gaps in your specific industry.

citedbrands.co.za/app/capitec/optimize

Optimize · What do you fix this month?

Bench v1● 3 actions queued

Priority actions · this monthranked by expected lift

Pitch BusinessTech on Capitec digital banking case study

+5 to +9 ptsLow effortPR

Bench evidence: n=4 SA banking peers, 6 months

Wikipedia entity-stack refresh + 2026 leadership update

+2 to +5 ptsLow effortContent

Bench evidence: n=11 SA peers, 3 months

HelloPeter sub-1hr response on top 12 open complaints

+3 to +7 ptsMedium effortComms

Bench evidence: Defends negative-polarity SA-share

Monitor

Are things getting better, or worse?

Visibility, sentiment, citation counts and peer position tracked every week. Real change vs random variance distinguished by 95% credible bands so you don't celebrate a +2 swing that's noise or panic about a -1 dip that isn't. Automated alerts the moment something needs attention.

citedbrands.co.za/app/capitec/monitor

Monitor · Are things getting better, or worse?

Bench v1● Tracking weekly

Visibility · last 12 weeks+8 pts

12 weeks ago: 63Now: 78 (95% CI 74–82)

Report

What does the CMO take to the board?

Every measurement feeds an 8-page board pack — live URL, monthly email digest, printable JSE/King-IV-formatted PDF. Editable in place. Cryptographically signed when the CMO freezes a version for the board. Every number cites the AI response behind it.

citedbrands.co.za/boardpack/capitec

Report · What does the CMO take to the board?

May 2026● Signed

Board Pack8 pages · audit-committee ready

PDF

Capitec · Board Pack · May 2026

Signed snapshot · for audit committee

Exec summaryVisibility trendCompetitor rankCitation mapTrigger watchValue proofRecommended actionMethodology

SHA-256: 4f2e…a91c · POPIA-compliant · FSCA-aligned

Why the loop works

Decision-grade scoring, not screenshots and guesses.

Every score in the War Room is built on methods you can defend to a data-science team. Pre-registered methodology, open dataset, every number traceable to the AI response behind it.

Bradley-Terry · 1952

Tournament-style ranking

The same maths used to rank chess players and AI chatbots against each other. Every pairwise comparison feeds a maximum-likelihood estimate of true brand strength.

Latin Square

Order-balanced testing

We ask "Brand A vs Brand B" AND "Brand B vs Brand A." Whichever brand we name first gets no free boost. Position bias cancels out.

Blind vs named delta

Name-cue lift detection

AI tells you what you want to hear when you name yourself. We catch it by running each query blind AND named. If the gap is >20%, we adjust. Your score reflects reality, not flattery.

String match, not LLM judge

Deterministic detection

We do not use AI to judge AI. Brand mentions are detected via exact string matching. "Capitec" found = counted. "A major SA bank" = not counted. Zero ambiguity, fully reproducible.

Every score in the platform is plotted against the SA peer set.

100 SA brands · 10 industries · 3 frontier LLMs · 14,826 responses · 188,877 citations. Pre-registered on the Open Science Framework before data collection. Updated quarterly. Public bench is free to browse, forever.

SA brands measured

100

Industries

AI responses analysed

14,826

Citations classified

188,877

Browse the SA-AEO Bench →Read the methodology

The executive surface

What the CMO sees on login: four answers.

Operators work the five phases. Executives get the Outcomes view — a 90-second read of where the brand stands in AI search, and the one thing to act on this month. Every number traces back to the War Room.

citedbrands.co.za/app/capitec/outcomes

Outcomes · How is Capitec doing in AI search?

May 2026Open Board Pack

Your brand, in four answers

Capitec in AI search.

90-second read
Updated 21 May 2026

Strong

Outcome 01

Cited

85 / 100

AI brings Capitec up 85 times in 100 banking queries. #1 of 10 SA banks. Peer median 64.

Strong

Outcome 02

Perceived

94% accurate

AI describes you accurately and positively. Low name-cue lift — you appear on your own, not just when named.

Watch

Outcome 03

Repaired

61% SA-held

Reputation surface 61% SA-controlled — healthy. Watch: 39% sits on foreign complaint sites you can't moderate.

On track

Outcome 04

Forecast

+15 / 12 wks

Visibility up +15 over 12 weeks. Branded search rose 19% in the same window — strong correlation.

The one thing to act on this month

FNB closed the gap by 3 points. A Moneyweb contributor placement is projected to add +5 to +9. Your board pack is signed and ready.

Open Board Pack →

See the War Room live with your brand's data. Book a walkthrough.

30 minutes. We pre-load your brand and your three closest SA competitors before the call. You see exactly what AI says about you and where the gaps are. No deck. No pitch. Just the platform.

Book the walkthrough →Or browse your brand in the bench first

Pricing is on the /pricing page. This page is about the product.