Knowledge Resource — 2025 Edition

Transforming AI &
Intelligent Tools

A structured reference module for understanding AI, its categories, tools, token economics, and practical use — prepared for finance and audit professionals.

25+ AI tools covered
7 AI categories
$0–$200 Monthly cost range

What is Artificial Intelligence?

01
AI refers to computer systems that perform tasks normally requiring human intelligence — reasoning, understanding language, recognising patterns, and making decisions. Modern AI is largely powered by large language models (LLMs) trained on vast datasets.

Narrow AI

Designed for one specific task — image recognition, spam filtering, recommendation engines. Most AI tools in use today are narrow AI.

Generative AI

Creates new content — text, images, code, audio. Powered by models like GPT-4, Claude, and Gemini. The fastest-growing category.

General AI (AGI)

Hypothetical AI with human-level reasoning across all domains. Does not yet exist but is the long-term research goal of labs like Anthropic and DeepMind.

How modern AI works: Large Language Models (LLMs) are trained by predicting the next token in a sequence across billions of documents. They learn grammar, facts, reasoning patterns, and context. When you prompt them, they generate a statistically likely continuation — which is why they are fluent but can sometimes hallucinate facts.

Categories of AI

02

Conversational AI

Understands and generates human language. Powers chatbots, writing assistants, and question-answering systems. Uses transformer-based LLMs.

ChatGPT Claude Gemini

Image Generation AI

Generates or edits images from text descriptions using diffusion models. Used in design, marketing, and visual prototyping.

DALL-E Midjourney Stable Diffusion

Code AI

Writes, reviews, debugs and explains code. Dramatically accelerates software development and is increasingly used in data analytics workflows.

GitHub Copilot Cursor Replit AI

Search & Research AI

Combines web search with AI reasoning to answer questions with cited sources. Replaces traditional keyword-only search for many queries.

Perplexity You.com Bing AI

Audio & Speech AI

Transcribes speech to text, converts text to voice, and generates music or audio. Key for accessibility, documentation, and media production.

Whisper ElevenLabs Otter.ai

Data & Analytics AI

Analyses structured and unstructured data to surface insights, forecast trends, detect anomalies, and automate reporting.

Julius AI Tableau AI Power BI Copilot

Major AI Tools & Their Purpose

03
Tool Category Primary Purpose Best For Model
ChatGPTOpenAI
Conversational Writing, analysis, coding, Q&A, document summarisation General professionals, students GPT-4o / o1
ClaudeAnthropic
Conversational Long-document analysis, audit reports, finance writing, reasoning Finance, legal, compliance Claude 3.5 / Sonnet
GeminiGoogle
Conversational Multimodal tasks — text + images + Google Workspace integration Google users, researchers Gemini 1.5 Pro
CopilotMicrosoft
Productivity Embedded AI in Word, Excel, Outlook, Teams — drafts, summaries Microsoft 365 users GPT-4 / Azure OpenAI
PerplexityPerplexity AI
Research Real-time web search with cited answers; fact-checking Researchers, due diligence Multiple models
GitHub CopilotMicrosoft / GitHub
Code Autocomplete and generate code in IDE; code review, test writing Developers, data analysts GPT-4 / Codex
MidjourneyMidjourney Inc.
Image High-quality image generation from text prompts Designers, creative teams MJ v6
Otter.aiOtter AI
Audio Meeting transcription, real-time captions, AI meeting notes Board meetings, interviews Whisper-based
Julius AIJulius
Data Upload CSV/Excel and ask questions; auto-charts, statistics Finance, audit analytics GPT-4 backbone
Canva AICanva
Design AI-generated presentations, social graphics, brand materials Business communication Multiple models
Notion AINotion
Productivity Summarise pages, draft content, extract action items within Notion Project management, notes Claude + GPT-4
GrokxAI (Elon Musk)
Conversational Real-time X/Twitter data + conversational AI; less filtered responses Social media intelligence Grok-2

Understanding Tokens

04
Why this matters for cost: Every interaction with a paid AI API is billed by token — both what you send in (input/prompt tokens) and what the AI generates back (output/completion tokens). Understanding tokens is essential for cost management.

Example: "Audit the financial statements for FY2024"

= 8 tokens (roughly 6–7 words per token on average)

Audit the financial state ments for FY 2024
Avg. words per token
0.75
~4 characters = 1 token
1,000 tokens ≈
750 words
~1.5 pages of A4 text
Context window
200K+
Claude's max input (≈500 pages)
Token cost range
$0.003–$15
Per 1,000 input tokens (varies by model)

Token types explained

Input tokens
Your prompt + system instructions + conversation history + any documents you paste in. Billed at input rate (usually cheaper).
Output tokens
The AI's response text. Billed at output rate (usually 3–5x more expensive than input). Shorter, precise responses save cost.
Cached tokens
Repeated context (like a system prompt) can be cached and re-used at a discount. Claude and GPT-4 both support prompt caching (≈90% cost reduction on cached portions).
Why do tokens matter for finance professionals?
When using AI for tasks like contract review, financial statement analysis, or drafting audit reports, you are often pasting in large documents. These become input tokens. If you paste a 50-page financial report (≈25,000 tokens) into every conversation, you are paying for those tokens every time. Good practice is to extract only the relevant section — Balance Sheet, Notes, etc. — rather than the whole document.
What is a context window and why does it matter?
The context window is the maximum total tokens (input + output) an AI can process in one conversation. Claude 3.5 Sonnet has a 200K token window — enough for roughly 500 pages. GPT-4o supports 128K. Once you exceed the window, the model forgets earlier parts of the conversation. For audit work involving large files, a larger context window is strongly preferable.
Do subscription plans (ChatGPT Plus, Claude Pro) also use tokens?
Yes, but they use a rate-limit model rather than per-token billing. You pay a flat monthly fee and get a generous message/usage allowance. If you hit the limit, you either wait or get downgraded to a weaker model. API access is per-token and is better value for high-volume automated tasks.

Efficient Use of Tokens

05
Each of the tips below can reduce your token usage by 20–70% with no loss in output quality. For a CA firm or bank processing high volumes of documents via API, this translates to significant monthly savings.

Trim your prompts

Remove pleasantries ("Could you please kindly...") and redundant instructions. A clean, directive prompt uses 30–40% fewer input tokens and often gets a better response.

Extract, don't paste whole documents

Instead of pasting a 60-page annual report, extract only the Income Statement and Notes. You get the same quality analysis at 10% of the token cost.

Request structured output

Tell the AI exactly what format you want — "Return a 5-row table with columns: Issue, Risk Level, Recommendation." Structured outputs are shorter and avoid verbose explanatory prose.

Use new conversations for new tasks

Each message in a conversation thread carries the full prior history as input tokens. Start a fresh conversation for unrelated tasks to avoid paying for irrelevant context.

Use system prompts & templates

For repetitive tasks (audit checklist review, VAT return analysis), write one precise system prompt once. Cache it if using the API. Reuse across all similar tasks.

Use the right model tier

Not every task needs GPT-4o or Claude Sonnet. Summarising a meeting transcript is a low-complexity task — use GPT-4o-mini or Claude Haiku (10–20x cheaper) for those.

Quick cost estimation formula

Estimated cost per task
= (Number of input tokens × input rate) + (Number of output tokens × output rate)

Example — reviewing a 10-page contract:
Input: ~5,000 tokens × $0.003/1K = $0.015
Output: ~800 tokens × $0.015/1K = $0.012
Total ≈ $0.027 per contract review

Paid AI Plans — Cost Comparison

06

All major AI providers offer a free tier with limitations and a paid subscription. Enterprise and API pricing is separate and usually more cost-effective for business use above ~50 hours/month. Prices are in USD and reflect 2025 rates.

Anthropic Pro
$20 / month

Claude 3.5 Sonnet (flagship)
200K context window
5x more usage than free
Priority access, Projects
No image generation
Google One AI Premium
$20 / month

Gemini 1.5 Pro access
2TB Google Drive storage
Gemini in Gmail, Docs, Sheets
1M token context window
Workspace add-on extra
Microsoft 365 Copilot
$30 / user / month

AI in Word, Excel, Outlook
Teams meeting summaries
PowerPoint generation
Requires M365 Business license
Best value at 10+ users
Perplexity Pro
$20 / month

Unlimited searches
300 Pro searches/day
GPT-4 / Claude / Gemini choice
File upload analysis
No coding / generation features
Individual
$10 / month

Code autocomplete in VSCode
Chat in IDE
Multi-model (GPT-4 / Claude)
PR summaries on GitHub
Requires coding workflow

API pricing — for automated / high-volume use

Model Input (per 1M tokens) Output (per 1M tokens) Context window Ideal use
GPT-4oOpenAI
$2.50 $10.00 128K Complex reasoning, vision, balanced cost
GPT-4o miniOpenAI
$0.15 $0.60 128K High-volume simple tasks, classification
Claude 3.5 SonnetAnthropic
$3.00 $15.00 200K Document analysis, audit, finance tasks
Claude 3 HaikuAnthropic
$0.25 $1.25 200K Fast, cheap summarisation, extraction
Gemini 1.5 ProGoogle
$3.50 $10.50 1M Very long documents, video analysis
Gemini 1.5 FlashGoogle
$0.075 $0.30 1M Cheapest per token for bulk tasks

AI Use Cases for Finance & Audit Professionals

07

Audit & Assurance

  • Summarise audit findings from working papers
  • Draft management letters and audit reports
  • Review disclosures against IFRS / BAS standards
  • Detect anomalies in trial balance data (Julius AI / Python)

Tax & VAT Advisory

  • Research NBR circulars and SRO interpretations
  • Draft objection letters and appeal documents
  • Summarise complex tax rulings for client memos
  • Compare year-on-year tax computation variances

Banking & Finance

  • Credit proposal drafting and financial spreading
  • Loan document review and covenant monitoring
  • Regulatory report drafting (BRPD, DOS circulars)
  • AML/CFT policy drafting and transaction narratives

Valuation & Advisory

  • Draft due diligence information request lists
  • Summarise target company annual reports
  • DCF model explanation and sensitivity narration
  • Comparable company research and benchmarking
Professional caution: AI models can hallucinate — they may produce plausible-sounding but incorrect figures, citations, or legal interpretations. Always verify AI-generated content against primary sources before signing off or submitting to a client, court, or regulator. AI is an accelerator, not an auditor. Professional judgement remains irreplaceable.
AI Intelligence Module — Compiled June 2025 • For professional reference and educational use • Prices and specifications subject to change