All categories
Every top-level taxonomy in the reference.
The frontier chat products people open dozens of times a day. This is where most professionals start, and where model quality, tool use, file handling and data-retention policies actually get decided.
4 tools Coding assistantsAutocomplete, chat-in-editor and autonomous agents that plan and edit multi-file changes. Selection hinges on how you are billed (seat vs. metered tokens), which models you are allowed to send code to, and how much review overhead the agent creates.
0 tools Image generationText-to-image models, background removal, retouching, upscaling and brand-asset pipelines. The practical questions are commercial licensing, style consistency across a campaign, and whether edits are reproducible.
0 tools Video generationText-to-video, image-to-video, avatars and AI-assisted editing. Costs are usually metered by render seconds, so teams need shot-level budgeting rather than a per-seat mental model.
0 tools Audio & voiceVoice synthesis and cloning, music generation, sound design and audio cleanup. Check consent and disclosure rules before cloning anyone's voice, and read the licence terms attached to generated music.
0 tools Writing & editingDrafting, rewrite and tone-shifting tools plus grammar and style engines that plug into docs, browsers and CMSs. The useful ones preserve your voice and your house style instead of flattening both.
0 tools Meeting notesBots and local recorders that capture calls, produce structured summaries and push action items into your task tracker. The deciding factors are recording consent, retention, and whether they work on your conferencing stack.
0 tools ProductivityAssistants and agentic runners that sit across email, docs, calendars and internal systems. Look hard at audit logging, permission scoping and what happens when a run fails halfway through.
0 tools Data & BINatural-language querying, automated charting and BI copilots. The important question is whether semantic definitions are governed or invented on the fly, because that decides whether the numbers survive a board deck.
0 tools Slides & docsDeck builders, document assemblers and PDF pipelines. Most teams care most about template fidelity, on-brand fonts and colours, and whether the output stays editable.
0 tools DesignPrompt-to-UI generators, design copilots and prototyping tools that hand off to real component code. The workflow question is whether the output is a picture, a spec, or a pull request.
0 tools Marketing & SEOKeyword and content briefs, ad-copy variants, creative generation and campaign analytics. These tools move fast, so treat any ranking or reach claim as a measurement task, not a promise.
0 tools SupportPlatforms for building and operating the assistants that answer your customers. This category is about grounded answers, escalation paths, analytics and CSAT - not about general chat products.
0 tools Sales & CRMCall intelligence, lead scoring, enrichment and CRM copilots. Value usually depends on data hygiene and on how well the tool fits the CRM your team already lives in.
0 tools Research & searchAnswer engines, literature search and deep-research agents that cite sources. Verify the citations: fluent synthesis with a broken source trail is worse than no answer.
0 tools No-code buildersDescribe an app or an automation and get a working one, plus the platforms that stitch SaaS tools together. Read the hosting, SSO and data-residency terms before anything touches production.
0 tools Dev tools & MLOpsVector stores, evaluation and observability harnesses, fine-tuning and inference platforms. This is the layer that turns an impressive prototype into something you can operate and bill for.
0 tools HR & recruitingSourcing, screening, scheduling and employee-experience tools. Automated decisions about people are regulated in many jurisdictions, so bias auditing and human review are part of adoption, not optional extras.
0 tools Legal & financeContract review, research and drafting, plus finance work like reconciliation and modelling. These fields punish confident errors, so citation quality and auditability matter more than speed.
0 tools GovernanceWhere organisations inventory what AI is in use, red-team it, monitor it in production and evidence compliance to auditors. Usually bought by platform, security or risk teams rather than by individuals.