AI Tools for Business
We show what today's AI tools can really do and help you match them to your team's work. We test new models and features for you on your company's own tasks, and teach the team only what actually works.
WHAT CLIENTS COME TO US WITH
- The team uses AI as a better search engine and doesn't know most of the features the company already pays for.
- New models and features appear every month, and nobody has time to check which ones are worth adopting.
- It's unclear which tool and which plan to choose, or what is actually worth paying for.
- Everyone works with AI their own way, and good practices don't spread to the rest of the team.
- There are no rules on what data can go into AI tools or how to set up privacy.
WHAT AI CAN DO AND WHICH TOOLS DO IT
- 01
Writing and editing
Emails, proposals, reports and translations in the company's voice. The model writes the first draft, a person edits and approves it.
ChatGPTClaudeMicrosoft 365 Copilot - 02
Working with documents
Asking questions about contracts, procedures and notes, comparing versions and pulling the key points out of long files.
Claude ProjectsNotebookLMCustom GPTs - 03
Data analysis
Insights from spreadsheets and exports, charts and summaries without writing formulas, with every step open to review.
ChatGPTClaudeCopilot in Excel - 04
Research with sources
Market, competitor or regulatory research with links to sources you can verify.
Deep ResearchPerplexityGemini - 05
Meetings, voice and images
Meeting notes and action items, voice conversations with an assistant, draft visuals and working materials.
Copilot in TeamsGemini in MeetChatGPT - 06
Code and technical work
Coding assistants in the editor and terminal: writing, reviewing and testing code, and working with documentation.
GitHub CopilotClaude CodeCursor
SCOPE OF WORK
- A review of what today's AI tools can do, using examples from your company's work.
- Choosing tools and plans (ChatGPT, Claude, Gemini, Copilot) to fit the team's tasks and budget.
- Setup: projects with company files, custom assistants, instructions and templates for the team.
- Connecting tools to company knowledge sources: drive, email and calendar.
- Testing new models and features on company tasks before the team starts using them.
- A library of proven prompts and before-and-after examples the team can come back to.
- Rules for using AI safely and privacy settings in the chosen tools.
BEFORE AND AFTER
Write a reply to a complaint.
A generic, polite text you end up rewriting anyway.
You're a customer service specialist at a furniture company. The customer says the table arrived with a scratched top (email below). Write a reply: apologise, offer a replacement top within 7 days or a 15% discount, and ask for photos. Keep the tone matter-of-fact, no excessive apologies, 120 words max. [customer's email]
A ready reply you only need to read and send.
Summarise this contract.
Half a page of generalities without what matters for the decision.
The attached file is a supplier contract. List the payment terms, contractual penalties, termination terms and anything that differs from standard clauses. Give the section number for each point. If something isn't in the contract, say so explicitly.
A list of points with section numbers, easy to check against the original.
AI tools change every few weeks: new models, working with files, deep research modes, assistants built into office suites and agents that carry out tasks step by step. Most teams use a fraction of these capabilities, because nobody has time to follow them and try them out. That's what we do: we show what today's AI tools can really do, match them to your team's work and make sure the team uses the best of them.
01What AI tools can do today
Chatting with a language model is only the start. The same tools can work with company documents, analyse spreadsheets, research the web with sources, take meeting notes and help write code. Many of these features are already included in the plans the company pays for; nobody has switched them on or knows how to use them.
We start with a review: what tasks the team does, which tools it uses and what it doesn't use in them. Based on that, we show specific features on examples from the company's work, not on generic demos from the internet.
02We test new models and features for you
A new model or feature doesn't always mean better work. So before we recommend anything, we test it on your company's tasks: we compare the results with what the team uses today and judge whether the change is worth the time and money. The team gets a short summary: what changed, what's worth trying and how, and what can safely be skipped.
We take the same approach to plans and licences. We help you judge whether a more expensive plan or an extra tool gives the team something it doesn't have today. We write more about this in Chat GPT Pro – is it worth paying for the Plus and Pro versions.
03How we work together
- Review. We talk to the team, collect typical tasks and check how it uses AI today.
- Selection and setup. We choose tools and plans, set up projects with company files, custom assistants and shared instructions, so everyone starts from good settings.
- Learning by example. We show features on the team's own tasks and build a library of proven prompts. Before-and-after examples show it best: the same task described vaguely, and with context, a goal and the expected format.
- Ongoing updates. We follow new models and features, test them on company tasks and let you know what's worth adopting.
To use models well, it also helps to understand how they work and where their limits and costs come from. We cover the basics in what is a token in AI.
04Security and privacy
Together with choosing the tools, we agree which data can go into them and which can't. We check privacy and data retention settings in the chosen plans and help write simple rules for using AI in the company. If the company already has a policy on this, we build on it.
05AI tools and training
This service is ongoing help with choosing, setting up and updating AI tools. If what you need most is workshops for your team, see our AI training for teams. The two work well together: training gives the team the basics, and the updates keep those basics from going stale.
Want your team to get more out of AI tools? Tell us what you do and which tools you use, and we'll show you what you can gain.
HOW WE WORK
- 01DISCOVERYBusiness goals, constraints, technical audit.
- 02STRATEGYScope, architecture direction, delivery plan.
- 03DESIGNInterface system, flows, validated prototypes.
- 04DEVELOPMENTIterative builds against a working environment.
- 05TESTINGAutomated coverage, load and security passes.
- 06DEPLOYMENTPipelines, monitoring, controlled rollout.
- 07SCALEPerformance, cost and capability expansion.
TECHNOLOGIES
FREQUENTLY ASKED QUESTIONS
How is this different from AI training?
Training is a workshop after which the team knows the basics. Here we help choose and set up the tools, then keep testing new models and features and tell the team what's worth adopting.
Which AI tool should a company choose?
It depends on the team's tasks and the tools the company already uses. A company working in Microsoft 365 may be well served by Copilot, while a team that writes a lot and analyses documents may do better with ChatGPT or Claude. We base the choice on tests with the company's own tasks.
Is it worth paying for paid AI plans?
Often yes, because paid plans offer better models, work with files and business privacy settings. But a more expensive plan doesn't always mean better work, so we check what the team actually needs.
How will we hear about new models and features?
We follow what's new, test it on your company's tasks and send a short summary: what changed, what's worth trying and how to use it.
Is company data safe?
We agree what data can go into AI tools, check privacy and data retention settings in the chosen plans and help write rules for using AI in the company.