# The tools I actually use this quarter Source: https://getstartedwithai.net/blog/the-tools-i-actually-use-this-quarter Site: Get Started With AI Tagline: Build AI automations that actually run your business. Audience: solo professionals, freelancers, consultants and AI implementers in training Contact: hello@getstartedwithai.co.uk This Markdown companion is generated from the prerendered page for AI crawlers. # The tools I actually use this quarter AI tools stack 2026 (Q2): the working AI consulting stack runs Claude Pro and ChatGPT Plus as the two assistants, Cursor for code work, Notion AI for second-brain notes, Zapier and Make for automation, Perplexity Pro for cited research, Canva for design, plus a small set of per-project tools that come and go inside the quarter. A consultant who tells you about the latest tool every week is selling something. The tools I use are the same set most months because what works keeps working. This is the list for the last 90 days with the work I used each one for, the three tools that entered the stack this quarter, the two that left, and the cost trajectory across the year so far. ## Daily tools - Claude Pro for client research, document drafting and any conversation longer than three turns. Voice fidelity on long drafts is the reason it stays daily. - ChatGPT Plus for fast lookups, code snippets and the Custom GPTs I run for client work. Speed and Custom GPTs are the reason it stays daily. - Cursor for any code work that needs to land cleanly. The IDE integration and the model selection make it the right tool for build code. - Claude Code when the work is operational rather than product-shaped (one-off scripts, small migrations, audit transcripts). - Notion AI for second-brain notes, the weekly review, and the running practice handover document. ## Weekly tools - Zapier for client automations and a few of my own. The breadth of integrations keeps it the default automation tool. - Make for the visual workflows that need branching, especially for client builds with three-plus conditional paths. - Perplexity Pro when the research needs citations and the citations need to survive a peer review. - Canva for the social and one-pager work that does not need a designer. - Otter Pro for meeting recording and transcription that flows into HubSpot and Notion automatically. ## Monthly or per project - Claude Projects for large client engagements with their own context window of a hundred-page brief plus prior outputs. - Lovable when a prospect needs an interactive demo from a sales call (rare, but a high-value piece when it lands). - Retell or Vapi for voice agent prototypes on client engagements where the workflow has a phone-call component. - Whisper for transcripts when the meeting tool has not done it. - NotebookLM for the kinds of multi-document research projects where a podcast-style summary is the deliverable. ## What entered the stack this quarter Three tools entered the stack between January and April 2026. Claude Code, after the operational scripts I was running in Cursor outgrew the IDE-shaped workflow and wanted a CLI-shaped one. NotebookLM, after a multi-document client research engagement that needed a single source of truth across 14 documents and the audio summary feature turned out to be useful for the prospect-facing version of the deliverable. Otter Pro, after the previous transcription tool started auto-pushing bad data into HubSpot and the manual cleanup time was eating the saving. Each of the three replaced an existing tool or use case rather than expanding the stack; the one tool per job rule still holds. ## What left the stack this quarter Two tools left between January and April. A premium proposal-writing tool that promised AI-drafted proposals; Claude with the right Project covers the work for free. A scheduling AI that booked discovery calls automatically; Calendly covers 95 percent of the value at less than a third of the cost, and the auto-booking sometimes booked the wrong slot. Both tools were costing about £30 a month each. Cancelling them paid for the Otter Pro upgrade with change to spare. The cost of switching out a tool is small if the workflow is documented; that is the point of the WIRE Method handover. The same logic applies to the consulting practice as it does to client builds. ## Cost trajectory across the year so far Q1 2026 average monthly tool spend: £92. Q2 2026 average monthly tool spend: £104. The increase is mostly Otter Pro replacing the cheaper transcription tool plus NotebookLM coming online for the research workflow. The increase pays back inside the first two client engagements in Q2 because the time saved on transcription cleanup and document research compounded across the audit conversations. The honest practitioner read on cost: the stack should grow only when a specific workflow needs a tool the existing stack cannot cover. Tools added speculatively almost always end up in the leave-the-stack list inside two quarters. ## Tool stack mapped to use case Each tool has a clear job. Client-facing work (proposals, audits, written deliverables) runs through Claude Pro. Internal work (Custom GPTs, fast lookups, image generation) runs through ChatGPT Plus. Build code runs through Cursor. Operational scripts run through Claude Code. Automation runs through Zapier first, Make second when the flow needs branching. Research runs through Perplexity Pro for cited work and NotebookLM for multi-document synthesis. Design runs through Canva. The mapping by use case rather than by feature keeps the stack scannable; I can answer a question like which tool handles client research without opening anything. The stack is small enough to hold in my head, which is the point. ## How the quarterly review works on my own stack Two questions per tool. Have I used it weekly across the quarter, and is it earning its fee in saved time at my own rate. Tools that fail either question get cancelled by the end of the month. Most of the stack stays stable; one tool tends to get swapped each quarter as the assistants pick up capabilities the dedicated tools used to own. The review is 30 minutes on the first Monday of the quarter, with a one-page checklist. The same Monday Morning Test I run on client engagements is the diagnostic I run on the practice. Time-stamped published reviews like this one are themselves the trust signal: the consultant practising what they sell, on a public schedule, with the dropped tools named. ## Build your first AI system live with Richard A 60-minute live build with Richard Batt. Bring one repeatable task. Leave with a mapped automation, success metrics and a deploy plan you can test the next morning. Free seat, no upsell. ### Live build session The next live session covers the WIRE Method end to end on a real audience build. [Join the next session](/webinar) ## Questions people ask before they start ### undefined undefined ### undefined undefined ### undefined undefined ### undefined undefined ### undefined undefined ### undefined undefined