10 new hacks every day
110 practical AI hacks that companies and businesses use or can use, across 11 tools from ChatGPT and Claude to Higgsfield and Grok Bot. Each one says what to do, who it suits and how much effort it takes, and links to the source so you can check it yourself. Last updated 30 October 2026.
Built with love by Shushant Lakhyani. Building bestproductivityresources.com
Search by tool, tactic or tag, then narrow it down by tool, business area, effort and date.
Tool
Showing 20 of 110 matching hacks (110 in total)
ChatGPT Record on the macOS desktop app turns meetings into transcripts and a canvas of notes. Turn on record history and later ask things like what did we decide in Monday's roadmap sync. It beats scrolling through a doc nobody finished writing, and the audio is deleted after transcription.
Since a February 2026 update you can restrict deep research to trusted sites and connect it to your own apps and MCP servers. Point a competitive analysis at your industry's regulators, top trade publications and your own drive, then let it run for 5 to 30 minutes. You get a cited report without the random blog spam.
ChatGPT's data analysis runs real Python, so a messy file gives you messy answers. OpenAI's own tip is simple: headers in the first row and one row per record. Spend five minutes on that, then ask for interactive bar, line or scatter charts straight from a Drive or SharePoint file.
Agencies and consultants can open one project per client and switch on project-only memory. ChatGPT then remembers that client's tone, history and quirks inside the project, without leaking details into another client's work. It is the cheap version of a separate account per client.
On Plus, Pro, Business, Enterprise and Edu, tasks can now be triggered by events in Gmail, Slack and GitHub. Set one up so a new email from a key customer gets summarised and a reply drafted the moment it lands. You stop being the human cron job that checks the inbox every hour.
OpenAI's help center says GPTs are being retired, with standard retirement on December 11, 2026, and GPTs moving to plugins. If your team leans on internal GPTs, list them now, copy out their instructions and knowledge files, and rebuild the important ones in the new setup. Do it before a sales or support workflow breaks on a Monday.
Write one monitoring task that checks a competitor's pricing page or news every morning and only pings you when something changes. Share the task link with the team so nobody builds a duplicate. Business plans get 10 active tasks, so spend them on the checks that would actually change a decision.
Moderna
When Moderna rolled out ChatGPT Enterprise, it ran a prompt contest and used it to find its top 100 power users. Those people became AI Champions who coached everyone else. It costs a few prizes and finds the people who will drive adoption far better than a mandatory training video.
Create a shared ChatGPT project for something like Q4 planning and invite the people who own it. Add the files, the Slack channel and the Google Drive folder you keep pasting into chats as project sources. Now everyone asks questions against the same context instead of rebuilding it in twenty private chats.
Moderna
Moderna built a GPT that turns one source document into versions for different audiences, like slides for an earnings call. The same move works for any team: paste the dense original and ask for a board version, a customer version and a frontline version. Each reader gets the part they care about.
Claude can create and edit Excel files, Word docs, PowerPoint decks and PDFs. Upload raw data and ask for a financial model with scenario analysis or a project tracker with a dashboard. You get a file you can send, not a wall of text you still have to rebuild.
If your app sends the same big system prompt, product catalog or policy doc on every call, mark it with prompt caching. Cache reads cost 0.1 times the base input price, while writes cost a bit more. For support bots and document Q&A this is often the single biggest line item you can cut.
Claude's advanced Research searches the web, Google Workspace and connected apps, and can work for up to 45 minutes before handing back a cited report. Connect the tools where your real answers live, like your CRM or project tracker, before you ask. The report then mixes market data with what your own team already knows.
Claude Enterprise smart reports analyse a sample of a team's work, group it into workstreams with spend and deliverables, and flag friction like a missing connector. Each report ends with recommendations, including recurring workflows to turn into skills. Run it on your best team, then roll their habits out to everyone else.
Source
Smart reports, surveys, and pooled group budgets for Claude EnterpriseClaude · 28 October 2026
Claude Projects let you ground answers in your own knowledge, like style guides, interview transcripts and past work, plus custom instructions per project. Drop in your brand voice doc and three of your best emails. Now anyone on the team writes like marketing on the first try.
Skills are folders of instructions, scripts and resources that Claude loads only when they are relevant. Anthropic's own example is following your organization's brand guidelines. Write the skill once and it works across Claude apps, Claude Code and the API, so every deck and doc comes out on brand.
The Message Batches API processes large volumes asynchronously and cuts costs by 50 percent. Most batches finish within an hour and all within 24. Tagging thousands of support tickets, rewriting product descriptions or running evals rarely needs a live answer, so batch them.
An artifact can be a dashboard, a doc or a small interactive tool that opens beside the chat and can be shared. On Team and Enterprise plans, people in your organization can use what you share without adding cost for you. Build the quote calculator or onboarding checklist once and hand everyone the link.
Claude's memory is scoped per project, you can view and edit what it remembers, and incognito chats do not save to memory at all. Keep long-running client work in its own project so context builds up. Do salary talks, HR cases and anything sensitive in incognito.
Anthropic's prompting guide says XML tags like instructions, context and input help Claude parse complex prompts, and that 3 to 5 examples work best. Build your team's reusable prompts this way. Outputs stop drifting in format every time a different person runs them.
Every hack is written up from the linked source, checked on the day it was added. AI tools change fast, so features, plans and prices can move after that date. Numbers are as reported by the source, often by the vendor itself, so treat them as claims, not audited results. Effort and business area are our own rough judgement. Test before you roll anything out, keep a human reviewing what the AI produces, and follow each tool's terms and your own data rules. Company and product names belong to their owners, and we are not affiliated with any of them.