TQ Group AI Adoption Showcase

TQ Group · AI Adoption Showcase

A living, interactive view of how AI can enhance creative production, strategy, and in-branch experiences — with evidence-based benchmarks and a pragmatic roadmap. This is a space for quick sharing and review of the main project deliverables.

TQ Group

A) Full Content Snapshot Comparison

A side-by-side comparison of selected original assets vs AI-generated counterparts covering TQ’s core content styles and formats. Establishes a clear cost/quality baseline.

Side-by-side comparison overview

1) Penflex · Rank TV

Key Details
Hours 11 (incl. ~10 minutes for music)
AI/compute spend R250 (±R180 video, R20 music, R50 images)
Estimated total spend ~R5,000. Note: This was a reproduction of half the video, so using AI tools it wuld cost ±R10,000 for a full-length full quality match of this kind of content.
Tool stack Compositing in Adobe (no traditional 3D modelling); low-cost AI music.
Notes Good parity via compositing; music was fast and inexpensive (~R20).
Recommendation Adopt upscale + frame interpolation as a standard finishing step.

2) Standard Bank · Credit Score

Key Details
Hours 8 (music + VO ≈ 30 minutes)
AI/compute spend R300
Estimated total spend ~R2,700 (±R5,400 for a full-length equivalent)
Tool stack Stills: Midjourney; Audio: Suno (music) + ElevenLabs (VO); AE components substituted with AI where possible.
Notes With this format, because it's relatively inexpensive to produce in the traditional way, I'd estimate a 2x price reduction, which is not as prominent as with other types, where we see a much bigger cost reduction. VO/music remain quick, low-cost wins. No need to settle for stills any longer. Animate the on-screen characters for better quality at similar budget.
Recommendation As with the 'SB Complaints Process' example, there is no need to stop at stills stock photos for tis kind of content - with AI Video generation as a tool in the stack, the cost is the same whether you bring the characters to life, or keep them static. Using AI for this content format is only half cheaper, but there is potential to save, while making the creative 'pop' even more with the new tools at our disposal.

3) Standard Bank · Prestige

Key Details
Hours 9 (labour ≈ R3,600)
AI/compute spend R600 (mostly Veo-3)
Estimated total spend R4,200 (Full AI reproduction est.: ~R8,000)
Tool stack Animation: Veo-3; Post: traditional finishing.
Notes Balanced cost vs. quality; strong candidate for AI-assisted production. This content type is similar to the format used in the 'Credit Score' and 'Complaints Process' examples - with one key differece: The original uses stock library footage clips (instead of stills). Given the nature of AI video generation tools, animating the stills is the same cost as keeping them still. Hence it makes sense to opt for this content in the AI paradigm. Stills / static posts will certainly always have a place, but animated content for small brands will be normalised.
Recommendation Bake in upscale + frame interpolation for cleaner motion at minimal extra cost.

4) African Bank · Tech Deals

Key Details
Hours 5.5
AI/compute spend R400
Estimated total spend R2,600
Tool stack Images: Midjourney, Sora, Ideogram · Audio: Suno (music), ElevenLabs (VO/SFX) · Animation: Veo-3 · Post: After Effects, Premiere Pro.
Notes Efficient AI-first pipeline across image, audio, and animation; smooth handoff to traditional post.
Recommendation Adopt upscale + frame interpolation as part of the standard pipeline.

5) SB Insurance Conversations

Key Details
Labour cost ~R4,400
AI/compute spend R600
Estimated total spend ~R5,000 (Note: half-length reproduction. A full-length version would be R8,000–R10,000).
Notes
  • Traditional tools are still required for brand elements like logos, chyrons, and captions.
  • It is more effective to generate fresh audio than to use the original to manage lip-sync and character consistency.
Challenges
  • Continuity over longer shots requires using the last frame of one clip as the start frame for the next.
  • Voice consistency requires audio processing: separate speaker tracks, process each with a voice changer (e.g., ElevenLabs), then recomposite the audio.

6) Standard Bank · Financial Services

Key Details
Hours 8
AI/compute spend R500
Estimated total spend R3,700
Workflow
  • Narrator in empty room · character consistency (MJ, Veo, Ideogram; edit in Premiere Pro)
  • Voice consistency via voice-changer (ElevenLabs, FL Studio)
  • B-roll memories (Veo-3)
  • Final composition & finishing (After Effects)
Notes Beautiful original; script carries through, but the actor’s performance is hard for AI to match today. Expect stronger parity within 6–12 months.

7) Standard Bank · TV Complaints Process

Key Details
Hours 5
AI/compute spend R150
Estimated total spend R2,250 (this is ~⅓ length; ~R7,500 for a full production in this style)
Notes
  • As noted with the Credit Score example, this type of content is not significantly improved by AI, and the savings potential is only around 2x as per my best estimations after doing the recreations. For other content formats we can expect much bigger cost savings. Chyrons and text elements are best handled with traditional tools for brand consistency; AI is unreliable for this at strict brand spec.
  • Fast and economical to build with AI, especially when paired with strong Adobe skills (no need to be a full specialist to see gains).

Premiere-baked A/B comparison reels

Note: NB: The Premiere-baked A/B reels below are for quick visual comparison only — they play with the original audio. To experience the full AI versions (including AI-generated music/VO/SFX), please watch the individual original vs AI pairs above. You can also open each in Google Drive to download in Full HD (large files) for the cleanest playback.

Links: Insurance: open, Prestige: open, Complaints: open, Credit Score: open, African Bank Tech Deals: open, SB FC: open, Penflex Rank TV: open.

Penflex · A/B composite (Premiere) Open full quality
Credit Score · A/B composite (Premiere) Open full quality
Prestige · A/B composite (Premiere) Open full quality
African Bank · Tech Deals · A/B composite (Premiere) Open full quality
Insurance Conversation · A/B composite (Premiere) Open full quality
Financial Services · A/B composite (Premiere) Open full quality
Complaints Process · A/B composite (Premiere) Open full quality

Cost comparison: Traditional toolstack vs AI-augmented toolstack

Asset
(All videos in Full HD or higher)
Length Approximate Cost · Traditional toolstack Approximate Cost · AI-augmented toolstack
(same length, brief & script)
6758 Penflex Rank TV 0:30 R40 000 R10 000
6728 SBTV Credit Score (Stock library still) 1:30 R10 000 (per video) R5 500 (per video)
7208 SBTV Prestige Banking (Stock library footage) 1:19 R28 000 R8 000
African Bank · Tech Deals 1:18 R24 000 R5 500
6419 SBTV Insurance Conversation 1:59 Possibly R100 000+ (exact budget uncertain) R8 000
6390 SBTV SBFC Video 2:00 R50 000 R7 500
7080 SBTV Complaints Process 2:40 ± R25 000 R12 000
Notes & Assumptions

At a high level, there is a trade-off between creative imagination and fine-grained control. Traditional tools still allow for more precise control over the final output. With AI, you hand over some of that control, but in return you gain creative freedom on a blank canvas where imagination is the primary limit, rather than budget.


It’s important to note that these reproductions were not crafted to optimise for similarity to the original, nor for quality per se. Rather, they were designed to showcase how current AI tools can be used in place of traditional ones. As such, I opted for an AI-first approach and only used traditional tools where absolutely necessary for brand-specific elements where precision is paramount.


Assumptions: Assumed rate for creative labour at R400 per hour. The above is calculated with that figure as the baseline labour cost. The figures above assume no client reverts (additional time would be required for revisions), that all original assets are provided, and that the length, brief, and script are the same as the original production.

General notes arising from content re-creations

  • Occasionally, company logos are overlaid to mask AI watermarks in generated footage.
  • Standardise an upscale + frame-interpolation pass in finishing — low effort, strong perceived quality boost. These models are lightweight, and can even be run locally on a decent computer (Graphics card with 8 - 12 MB VRAM recommended)
  • Image model selection benefits from testing various models; I compared outputs from state-of-the-art models like Imagen4, Nano Banana; Midjourney 7, GPT-Image-1, Wan 2.2 and Ideogram, then curated and manually polished the strongest results in traditional tools.
  • LLMs help accelerate ideation and prompt development, and can assist with rapid iteration on visual directions.
  • “Compute” includes model credits, API usage and tool subscriptions. Example: Google charged ~R2,600 for 12,500 AI credits; an 8-second clip costs ≈R20 in “Quality” mode or ≈R4 in “Fast” mode. Similar costs apply across other providers (e.g., Kling, Higgsfield, Replicate).

B) The AI Opportunity Report

A curated view of the AI content stack for video/animation, imagery, audio, copy, and final production — plus a concise section on the strategic shift from SEO to AEO within the LLM era.

  • Tool landscape & licensing
  • Workflow patterns & automations
  • Considerations, limitations & workarounds

Strategic Reports

Click below to access full PDF reports.

New Adobe Boards + Google DeepMind Suite Creative Workflow examples - Penflex social media ads.

C) Strategic Roadmap Document

This document provides a strategic framework with targeted suggestions and recommendations for AI adoption tailored for TQ Group's unique situation. Moving beyond the hype, it establishes first principles for a responsible, human-centric approach focused on scaling quality, not just quantity. It uses scenario planning to map potential futures based on low, medium, and high rates of AI adoption, outlining a clear strategic response for each. Finally, it analyses the key economic and market vectors—from compute costs to the rise of data moats—that will shape the landscape. The goal is to equip TQ Group with the foresight to navigate the AI shift and turn disruption into a durable competitive advantage.

OpenAI have recently released a very comprehensive report that I wholeheartedly recommend reading. It outlines their Align, Activate, Amplify, Accelerate, and Govern framework for organisational adoption. OpenAI: Staying Ahead in the Age of AI (PDF)

Strategic foresight & roadmap document

Click below to read the full PDF.

D) The Creative Potential Showcase:

A TQ-Tailored AI-first Content Ideation Springboard

The initial benchmark provides a crucial cost and quality baseline. However, the true creative opportunity of AI is not in replicating existing work for less, but in new forms of visual storytelling that were previously unfeasible due to budget, time, or technical constraints. This showcase will feature a selection of new, innovative concepts designed to demonstrate this power.


For TQ Group, the strategic application of AI can be simplified into two core functions: the ability to SCALE and the ability to TAILOR.


The power to TAILOR content unlocks a new level of personalisation. This could include; dynamic mini-campaigns for smaller cities that were previously not cost-effective to target, generating beautiful e-cards or thank-you notes with personalised name typography for individual customers, creating custom emails sent directly to a CRM via APIs,


This in turn creates an AI data flywheel, where more campaigns generate more data, enabling even more effective personalisation in the future.


Creators like The Dor Brothers and GerderGotit exemplify these principles of AI-first storytelling and provide inspiration for what is now possible.


Some fun ideas for TQ to propose to clients to help them embrace AI:


Standard Bank: "The Blueprint of Your Future"


Overarching Idea: To visually represent how Standard Bank helps build a customer's financial future by blending three distinct life stages into a single, seamless portrait.


AI Ingredients & Execution:
Image 1 (The Foundation): A close-up, aspirational portrait of a young professional.
Image 2 (The Growth): A beautiful photo of their dream family home.
Image 3 (The Legacy): A stylised architectural blueprint or a financial growth chart.


The "Magic" Step: Use the new blending tool to seamlessly merge these three images. The blueprint/chart would form a semi-transparent, textural layer over the entire image. The dream home would be blended into the upper part of the portrait, emerging from the person's mind like a tangible thought. This creates a single, powerful image that tells a complete story of financial growth.


Tagline: "Build your tomorrow, today."


Kimberley Anne Hotel: "The Taste of a Memory"


Overarching Idea: To evoke the rich, multi-sensory experience of dining at the hotel by blending a signature dish with the raw ingredients and the luxurious ambiance of the restaurant.


AI Ingredients & Execution:
Image 1 (The Dish): A stunning, professionally shot photo of the restaurant's signature winter dish.
Image 2 (The Ingredients): A dark, moody photo of the raw, key ingredients (e.g., rich red wine, herbs, a cut of venison).
Image 3 (The Ambiance): A photo of the "Copper & Coal" restaurant's interior, focusing on a warm leather chair and the glow of the fireplace.


The "Magic" Step: Blend the images so the final dish sits in the foreground, but the textures of the raw ingredients are subtly visible within the food itself. The warm, out-of-focus restaurant ambiance would form the background, creating a deep, layered image that conveys taste, quality, and atmosphere all at once.


Tagline: "Every ingredient tells a story."


TQ Group: "The Storyteller's DNA"


Overarching Idea: A powerful self-promotional piece for TQ Group that visually represents their brand promise: "Storytellers with a knack for technology." It will blend the human element of creativity with the organic nature of stories and the structure of technology.


AI Ingredients & Execution:
Image 1 (The Storyteller): A dramatic, black and white portrait of a creative person.
Image 2 (The Story - Organic): A macro shot of the intricate rings of an ancient tree trunk.
Image 3 (The Technology - Structure): A clean, abstract image of a glowing geometric data network.


The "Magic" Step: Blend the three images into a sophisticated visual metaphor. The tree rings would be mapped to the contours of the person's face, representing the deep, organic nature of storytelling. The glowing data network would then be subtly overlaid, tracing the lines of the tree rings, representing the technological layer that enhances the story.


Tagline: "Storytellers with a knack for technology."


African Bank: "The Audacity to Ask"


Concept: A series of ultra-short, powerful social videos featuring real South Africans. Each person looks directly at the camera and asks a bold question about their future, which then transforms into a beautiful, surreal visual representation of their dream.


Visuals: A young entrepreneur in a small workshop looks at the camera and asks, "What if my workshop could reach the world?" As she speaks, the background behind her dissolves and morphs into a hyper-lapse of her products being shipped across a glowing map of Africa. A student in a library asks, "What if this degree could build a hospital?" The library shelves behind her transform into the gleaming, modern facade of a hospital.


Kimberly Anne Hotel: "A Memory in Every Detail"


Concept: A slow, luxurious, and artistic video where the camera glides through beautifully detailed, dreamlike still-lifes of the hotel experience. Each object subtly tells a story of a guest's memory.


Visuals: Instead of showing people, we see the aftermath of their perfect stay. The camera focuses on a beautifully rumpled bedsheet, which subtly animates to show the ghost-like impression of a couple sleeping. We see two half-empty champagne glasses where the bubbles are tiny, shimmering diamonds. The camera pans across a guest book where the ink of a signature slowly animates into a stylised map of the guest's journey to get there.


Penflex: "The Unfinished Masterpiece"


Concept: A viral-style video that starts with a famous, unfinished work of art (like da Vinci's "Adoration of the Magi" or an unfinished symphony score) which is then magically and colourfully completed using a storm of animated Penflex stationery.


Visuals: The camera pushes in on the blank, unpainted section of a famous historical painting. Suddenly, a swarm of brightly coloured Penflex pens, markers, and highlighters fly into the frame. They act like a painter's brush, filling in the missing details with vibrant, modern colours and textures. The final shot reveals the entire masterpiece, now fully complete, with a Penflex pen resting at the bottom of the frame like an artist's signature.

Exploding pencil case · back-to-school concept

This short spot explores a playful, high-energy reveal as stationery bursts from a pencil case in slow motion. It's also seamlessly loopable (good for TikTok and YT Shorts).

It demonstrates how GenAI video can create eye-catching, physics-inspired moments quickly — ideal for promos and social content where surprise and colour drive recall.

The music, created in Suno 4.5+, is designed to be a trendy urban re-imagining of a popular isiXhosa nursery rhyme, commonly sung by school children in SA. Note: This is a very powerful music engine, which supports detailed prompting. You can talk to it like it’s a room full of musicians. Remarkably, it will at times even adhere to very specific phrases like “Exultant semiquaver triplets in the myxolydian mode”. As with other AI models, ‘meta-prompting’ is a good idea; Often,. ChatGPT and Gemini understand the quirks and nuances of prompting the various models, better than we human creatives do.


ISuno prompt: “Award winning hip hop urban South african trap/ hip hop smash hit - contrast from nursery rhyme and childish melodies to dark gritty bass, and growling 808s with gnarly urban rhythms - grime and trap style beat spiced up with many layers and found sounds - traditional griot call and response structure."


Suno Lyrics prompt. Note,as with ElevenLabs’ Voice generator, these prompt fields support [direction tags] for control over articulations, emotion, accent, intonation, which adds a lot of richness, and can even ‘humanize’ the AI output somewhat:


“[Intro - mysterious, subminimal whispers, barely audible] Capha capha capha…Imanzi ilokwe yam / Capha capha capha / Imanzi ilokqe yam / Gqum gqum gqum / kuyaduduma / ...”


INote the big international models tend to be under-exposed to South Africa content in their training data, and as such certain phonemes in isiXhosa, especially plosive ‘click’ sounds like c, x, q that are unique to Southern Africa are difficult to achieve using current gen AI.

The direction also shows how AI footage, Adobe finishing, and light compositing can combine for fast iteration.

Kimberley Anne · Social ad gallery

The Kimberley Anne Hotel has a beautiful, consistent brand identity that’s ideal for elevation with AI. The point isn’t to replicate what’s already possible — it’s to unlock visuals that would be impossible or prohibitively expensive with traditional photography, while staying true to the brand’s soul.

Practically, the workflow remains familiar whether AI is used or not: brief, iterate, art direct, and finish — AI simply expands the creative surface area. Rapid ideation for social ads using AI image generation: start with LLM prompts to explore angles, headlines and tonal directions, then generate base images (e.g., Midjourney / Imagen), and composite brand elements in Photoshop for on-brand outputs. This accelerates testing — produce several concepts in an afternoon, review with stakeholders, then promote the strongest into motion or print.

TQ Brand Concepts · Motion & Stills

Inspired by TQ's exquisite website hero banners, this ad was designed to be quick and cost-effective to produce while staying true to the visual language of the brand. The seamlessly loopable sequence pairs a short motion piece with four stills to explore connected metaphors - ideas, orchestration, organic growth, and weaving - staying tru to the TQ brand's visual langauge language.

Because this kind of video content follows a familiar creative cadence, teams can seamlessly incporporate AI into traditional workflows, through integrated platformms, such as Adobe's 'FireFly Boards', to test ideas and mock up concepts faster. The workflow mirrors our other pieces in the TQ AI showcase: LLM-assisted ideation and image generation (various models, curating and manually polishing the best results); these are then animated in Veo-3 and Sora, Kling etc. and final composition in Premiere Pro (or After Effects if needed). The sequence pairs a short motion piece with four stills to explore connected metaphors — ideas, orchestration, organic growth, and weaving — all in a single brand language. This example also showcases the power of precision prompting for unique, creative-specific music scoring.

The final stills were made in Google's new "Nano Banana" image model (Gemini Image 2.5 Flash) which was released towards the end of August. I have included a little more info about this important new technology and it's implications in Deliverable B - under the AI tools and model ecosystem section.

E) Executive Summary -
A Strategic Framework for AI Adoption at TQ Group

Artificial intelligence is rapidly becoming a cornerstone of the modern "knowledge economy." As with previous disruptive technologies like digital synthesizers or cameras, which were initially dismissed as "cheating" but quickly became industry standards, AI will normalise. The central strategic challenge is no longer if AI will be used, but how to leverage it to maintain a competitive edge and sustain relevance.

This challenge presents several potential pathways for an agency like TQ Group:

  • Conservative Niche: Capitalise on AI's divisive nature by positioning TQ as a premium, non-AI boutique for clients who value purely human-led creative processes.
  • Moderate Integration: The most likely path. Integrate AI for improved efficiency and enhanced capabilities, adopting a "centaur" model where human operators direct AI tools to achieve superior results.
  • Radical Pivot: Use TQ Group as an internal case study to build and productise proprietary AI tool stacks. In this model, the business pivots from selling creative assets to selling the (GenAI-powered) means of production. For example, pivot to providing clients with custom-built agentic creative asset engines tailored to their brand, or “meta-campaigns”.

Whichever path is chosen, the strategic application of AI can be simplified into two core functions: the ability to SCALE (producing more, faster) and the ability to TAILOR (creating highly personalised content for specific audience segments). Successfully combining these functions creates an AI data flywheel, where more campaigns generate more data, leading to even more effective personalisation.

To succeed, particularly in the moderate integration scenario, requires a shift in focus from technical execution to human-led judgment. The key differentiators in an AI-augmented workflow become:

  • Good Taste: When generation is easy, curation and artistic direction become paramount.
  • Being Relentlessly In Touch: Audience connection, and the ability to tell stories that resonate on a human level is still a quality AI cannot currently mimic, and perhaps never will be able to.
  • Strong Organisational Skills: AI can be overwhelming. It presents a blank canvas of possibilities where imagination is the limit! The new tools require new management skills to guide teams, maintain quality control, and avoid overwhelming workflows. This means actively adopting a "centaur" model, where a skilled human is always in charge, directing the AI to do the heavy lifting.

This project provides TQ Group with a comprehensive strategic package to navigate these choices. The deliverables are designed to work together, moving from a broad understanding of the landscape to specific actions and strategic foresight.

Deliverable A: The Snapshot Comparison provides the evidence-based benchmark. It offers a direct, side-by-side comparison of original creative assets against their AI-generated counterparts, establishing a clear baseline for cost, time, and quality savings. The key finding from this benchmark is an average 5x cost reduction (around 80% cheaper) across TQ Group’s main content formats.

Deliverable B: The AI Opportunity Report provides the foundational context. It is a detailed report explaining the modern creative ecosystem, practical human-in-the-loop workflows, and the strategic shift from SEO to AEO. It is supported by a comprehensive Tool Matrix that audits best-in-class applications.

Deliverable C: The Strategic Roadmap contains the high-level strategic foresight. It establishes first principles for responsible adoption, outlines a "quality-first" mindset, and maps three potential futures through low, medium, and high adoption scenarios, each with a clear strategic response for TQ Group.

Deliverable D: The Creative Potential Showcase serves as an ideation springboard. It moves beyond replication to demonstrate storytelling possibilities that were not previously feasible, featuring a range of new, tailored concepts for TQ Group and its clients.

Collectively, these deliverables provide TQ Group with a robust, clear-eyed framework for turning the disruption of AI into a durable competitive advantage.

Listen: TQ AI Adoption Overview – NotebookLM AI-generated podcast