The generative AI landscape, explained for professional firms
Generative AI is no longer one tool. It is a fast-moving ecosystem of assistants, media generators and creative platforms, each with different strengths, different risks and different data handling practices.
This page gives you a plain-English overview of the tools your staff are most likely using, or asking to use. For each one, we explain what it does and what a professional firm should think about before approving it.
KiwiGen.AI is vendor-neutral. We do not sell or resell any of these tools. We help firms choose, govern and safely adopt the right ones for their work.
AI Assistants
These are the general-purpose tools most firms encounter first. Staff use them to draft, summarise, research, analyse and brainstorm. They are also the tools most likely to be used informally, on personal accounts, without oversight.
ChatGPT — OpenAI
ChatGPT is the most widely recognised AI assistant and often the first tool staff adopt on their own. It handles drafting, summarising, research, analysis and image generation, and can search the web for current information. OpenAI offers Business and Enterprise plans with stronger data protections than the free consumer version.
What firms should consider: The gap between consumer and business plans matters. Free and personal accounts may use your inputs in ways that are unsuitable for client information. If your people are using ChatGPT, the question is which version, under whose account, and with what guidance.
Gemini — Google
Gemini is Google’s AI assistant, and its natural home is inside Google Workspace. For firms already running Gmail, Docs, Drive and Meet, Gemini can draft, summarise and analyse directly within the tools staff use every day. It also powers Google’s NotebookLM research tool and connects to Google Search for up-to-date answers.
What firms should consider: Deep integration is both the appeal and the risk. When an assistant can see your email and documents, access controls, admin settings and data governance need to be configured deliberately, not left on defaults.
Claude — Anthropic
Claude is Anthropic’s AI assistant, known for strong performance on long documents, careful reasoning and nuanced drafting. This makes it popular with lawyers, advisers and analysts who work with lengthy contracts, reports and correspondence. Anthropic offers team and enterprise options, and Claude is also available through major cloud platforms.
What firms should consider: Claude’s long-document strength encourages staff to upload substantial files. That is exactly where confidentiality policies need to be clear: what categories of client material may be uploaded, on which plan, and with what review before outputs are relied on.
Copilot — Microsoft
Microsoft Copilot brings generative AI into the Microsoft 365 environment: Word, Excel, Outlook, Teams and PowerPoint. For the many New Zealand firms built on Microsoft, Copilot is often the path of least resistance because it works inside existing tenancy, identity and security arrangements.
What firms should consider: Copilot can surface anything a user already has permission to access. Firms with untidy SharePoint permissions or legacy file shares often discover that Copilot makes over-sharing visible. A permissions and information-governance review before rollout is time well spent.
Video Generation
AI video has moved quickly from novelty to production tool. Firms are beginning to use it for training content, client education and internal communications.
Seedance 2.0 — ByteDance
Seedance 2.0 is ByteDance’s video generation model. It produces cinematic, multi-shot video with native audio from text, image, video and audio inputs, and has attracted significant attention for the realism of its output. It also includes content-authenticity measures such as C2PA watermarking.
What firms should consider: Seedance’s realism cuts both ways. It is powerful for training and communications content, but the same realism raises deepfake, impersonation and copyright questions. Firms should also weigh the data jurisdiction and supplier due-diligence considerations that apply to any offshore platform.
Veo — Google
Veo is Google DeepMind’s video generation model, producing high-quality video with audio from text and image prompts. It is available through the Gemini app and Google’s enterprise cloud, which makes it a natural option for firms already inside the Google ecosystem, with enterprise controls and Google’s SynthID watermarking for AI-generated content.
What firms should consider: Enterprise access via Google Cloud gives firms clearer contractual and data-handling terms than consumer video tools. As with all generated video, firms should set rules for disclosure, brand use and approval before content is published.
Image Generation
Image tools are now genuinely useful for professional work: diagrams, infographics, training visuals, presentation graphics and marketing assets.
ChatGPT Images 2.0 — OpenAI
ChatGPT Images 2.0 is OpenAI’s image model, built directly into ChatGPT. It is a “thinking” image model: it reasons about the request before generating, which produces reliable text inside images, accurate multi-element layouts and consistent characters across a series of images. That makes it well suited to diagrams, posters, storyboards and training material.
What firms should consider: Because it lives inside ChatGPT, the same account and data-handling questions apply. Accurate-looking infographics still need human fact-checking; a polished image of wrong information is worse than no image.
Nano Banana Pro — Google
Nano Banana Pro (Gemini 3 Pro Image) is Google DeepMind’s flagship image generation and editing model. It renders legible text in multiple languages, supports up to 4K output, maintains brand and character consistency using multiple reference images, and can ground images in real information via Google Search. Every image carries an invisible SynthID watermark.
What firms should consider: Its strength in brand-consistent, text-heavy visuals makes it attractive for marketing and client-facing material. Firms should establish review steps for factual accuracy and brand approval, and decide who is authorised to publish AI-generated visuals externally.
Audio and Voice
ElevenLabs
ElevenLabs is the leading AI voice platform, producing natural-sounding speech in many languages, along with voice cloning, dubbing and conversational voice agents. Professional firms use it for training narration, e-learning modules, podcasts and accessible client communications.
What firms should consider: Voice cloning demands special care. Firms need clear consent rules before anyone’s voice is cloned, and staff should understand that realistic voice technology is also a fraud vector: voice-based impersonation scams are rising, and internal verification procedures should reflect that.
Creative Platforms
These platforms combine multiple AI models and professional editing tools in one place. They suit firms building serious content programmes, such as learning libraries, marketing campaigns or client education series.
Runway ML
Runway is one of the most established AI creative platforms, offering video generation, image tools and a suite of editing capabilities used widely in media and marketing. Its Gen series of video models pioneered much of the current AI video landscape, and the platform is built around professional creative workflows.
What firms should consider: Runway is a mature choice for firms producing regular video content. Review its plan tiers and data terms, and set internal standards for disclosure when AI-generated footage appears in client-facing material.
Higgsfield
Higgsfield is a creative platform that provides access to multiple leading generation models, including Seedance 2.0 and other video and image models, through a single interface. This “many models, one platform” approach lets teams pick the best model for each task without managing separate subscriptions.
What firms should consider: Aggregator platforms simplify access but add a layer to your supply chain: your content passes through the platform and on to third-party model providers. Due diligence should cover both the platform and the models behind it.
Figma Weave
Figma Weave (formerly Weavy, acquired by Figma in 2025) brings leading AI image and video models together with professional editing tools on a node-based canvas. Instead of one-off prompts, teams build repeatable visual workflows: generate options with one model, refine with another, then apply hands-on edits like masking and colour grading. Weave tools are now beginning to integrate into the wider Figma platform.
What firms should consider: Workflow-based platforms reward firms that treat content creation as a repeatable process, which suits brand consistency and quality control. As with any multi-model platform, understand which underlying models your content is routed to and on what terms.
Choosing the right tools for your firm
There is no single “best” AI tool. The right choice depends on your existing technology environment, the sensitivity of your client information, your regulatory obligations and what your people actually need to do.
What matters most is not which tool you pick, but how you adopt it: approved tools and plans, clear rules about client information, human review of outputs, and staff who know how to use AI well.
That is where KiwiGen.AI helps. We work with law firms, accounting firms, financial advisers and other professional practices to select, govern and safely roll out generative AI, the Kiwi way.
Start with a practical conversation about the tools your firm is using, or should be.
KiwiGen.AI 