Anthropic’s Fable 5 model is back online today after a government suspension, and new data from PwC shows AI’s payoff still depends entirely on where you point it. This issue covers how ChatGPT and Google Gemini can strengthen your next business plan, plus an honest verdict on whether Notion AI is worth its Business plan price tag.
Prompt Engineer
Before you send your business plan to a lender or investor, get an AI tool to find the holes first. This prompt turns it into a sceptical reviewer instead of a cheerleader, which is a much cheaper place to get hard feedback than after a rejection.
I’m writing a business plan for [type of business]. I’ll upload a document containing all my notes, research and draft sections.
Read it as a sceptical UK lender or investor would. Then:
- List the three weakest assumptions in my forecast and explain why they’re weak.
- List the questions you’d ask before agreeing to fund this.
- Point out anything unclear or missing that a first-time reader would need explained.
Be direct. Don’t soften your feedback.
AI Updates Roundup
This week brings a study in contrasts: one AI model getting its safety net rebuilt in public and one report reminding us that AI’s payoff still depends entirely on where you point it.
Overview
- Anthropic’s Fable 5 model is back online globally after a government-ordered suspension over a jailbreak concern.
- New McKinsey data shows AI is paying off strongly in some business functions and barely at all in others.
- So what? Fable 5 users regain full access today, though some routine coding tasks still route to Opus for now.
- So what? The gap between AI leaders and laggards is widening fast, so waiting no longer looks like the safe option.
*An AI jailbreak is a technique that gets around an AI model’s built-in safety guardrails. By using carefully crafted prompts or psychological manipulation, users trick the AI into ignoring its own ethical constraints and generating responses it’s explicitly programmed to refuse.
How can AI help with my business plan?
A business plan asks you to be your own researcher, writer and financial forecaster all at once. Most business owners are none of those things by trade. That’s where AI earns its place: not as a shortcut, but as a way to get past the parts that eat the most time.
Overview
- AI tools can pull scattered research into a clear summary and draft the sections built on your own knowledge.
- Use AI to stress test your assumptions and list the questions a sceptical lender would ask.
- Financial forecasts and final judgement calls are still yours to make.
Getting your research in order
Every business plan starts with proving you understand your market: competitors, trends, customer behaviour. General-purpose AI assistants such as ChatGPT, Claude and Google Gemini are all capable of pulling scattered research into a clear summary and flagging gaps in your thinking. They’re a starting point for structuring what you already know.
Turning notes into a first draft
A blank page is where most business plans stall. Feed an AI tool your notes on your business idea, your customer and how you’ll make money, and it can turn that into a working first draft. Workspace tools like Notion and document assistants built into Google Docs or Microsoft Copilot can do this directly in the document youβre already writing.
Stress testing your own thinking
One underused trick: ask an AI tool to argue against your plan. Get it to list the weakest assumptions in your forecast or the questions a sceptical lender would ask. You’ll get sharper answers back than from a generic summary request.
Where it can’t help
Financial forecasts are the part of a business plan lenders and investors scrutinise hardest. AI can help you build a spreadsheet or explain what a cash flow forecast should include, but it can’t verify your numbers or judge whether your assumptions hold up. GOV.UK’s guidance on writing a business plan sets out what a lender actually expects to see, AI generated or not.
It’s also worth remembering what a business plan is for. It needs to prove you understand your own business. A plan that reads like it was written entirely by AI, with none of your own thinking in it, works against you.
Quick Review: Notion AIπ§
Notion is a workspace tool that pulls scattered notes, research and documents into one place, then uses AI to draft, summarise and organise them. For UK SMEs juggling emails, spreadsheets and half finished Google Docs, it offers a single home for the messy first stages of any business document.
Key Takeaways
What’s Good β :
- Drafts sections from bullet points and turns long research into a short summary
- Pulls structured data out of messy notes and into a table
- Works well as a single workspace for market research, competitor notes and financial projections
What’s Less Good β:
- Full AI access now sits behind the Business plan rather than the cheaper Plus tier
- Free and Plus plans only offer a limited AI trial before prompting an upgrade
- Notion has changed its AI pricing structure more than once in the past year
Impact on Your Business π: For a sole trader or small team writing a plan for a loan application or investor pitch, Notion solves a real problem: too many scattered notes and not enough time to turn them into a coherent document. It will not write your financial forecasts and any factual claims still need a human check, so it works best as a drafting assistant rather than a replacement for your own judgement.
Financial ROI π°: The value sits in synthesis rather than pure writing. Turning a few hours of scattered research into a structured first draft in twenty minutes is where the time saving happens. Also, having everything organised and easy to find also cuts down the hours lost searching through old notes and documents.
Price Point π³: Notion’s Free plan costs nothing and covers a solo user, though AI is limited to a short trial. Full AI drafting and summarising requires the Business plan, which runs to around Β£16 per user per month on an annual contract, plus VAT.
Upcoming Events & Conferences π
- From Copilot to Autopilot: Agentic AI’s Impact on Treasuryπ¦
- Date: 14 July 2026
- Location: Online
- Cost: Free
A Finextra webinar for bank and corporate treasurers on where agentic AI can safely take over cash forecasting and liquidity management, and where a human still needs to stay in the loop.
- Digital Sovereignty: Banking Transformationπ
- Date: 28 July 2026
- Location: Online
- Cost: Free
Useful for operations and IT leads at financial institutions weighing up how to modernise core systems and embed AI without losing control of sensitive data.
- AI, HR & Financial Services: Modernising Banking Systemsπ₯
- Date: 30 July 2026
- Location: Online
- Cost: Free
Aimed at HR and operations leaders in banking who want a straight answer on where AI actually helps with workforce and system modernisation, rather than just bolting AI onto old tech.
Training & Skills Development π
- ICAEW β Generative AI in Finance: Smarter Operations & Sharper Decisionsπ
- Date: 16β18 November 2026
- Location: Online (virtual classroom)
- Cost: From Β£625 (+VAT)
A CPD-verified course for finance professionals wanting hands-on GenAI skills for forecasting, budgeting and internal controls, with a 90-day adoption plan built in.