Is AI Creating a Finance Jobs Crisis?, Ademola Odewade, CEO, K-Dems Consulting Ltd.

Core claim: AI and jobs

• AI is framed as something that will not replace humans directly.
• Instead, people may be replaced by colleagues who can use AI effectively.
• AI is treated as a tool, not a standalone decision-maker.

Macro/job concern

The guest argues the economy is already moving toward a “jobs crisis,” citing:

• UK economic growth described as approximately 0–1% (approximate range).
• Millions of unemployed youth, including over 1 million unemployed young people in the UK (exact figure not provided).

Practical finance workflow automation

An operations/finance reporting example explains how AI can compress work:

• Create a briefing-style report by prompting Claude and ChatGPT to synthesize what major firms and analysts say (e.g., PwC, Deloitte, KPMG, EY, Forrester, Bain, McKinsey).
• Claimed outcome: an answer in ~15 minutes, instead of weeks of traditional analyst effort.

Risk and ethics emphasis (cognitive + governance)

Key cautions and themes:

• Removing “manual” discipline can cause “cognitive surrender”—people stop thinking or validating.
• AI can hallucinate, so outputs require:
  • Controls
  • Human verification
• Data quality is foundational:
  • “Garbage in, garbage out.”
  • If data quality is poor, the AI strategy fails.

How AI changes value creation (efficiency framing)

• Referenced concept (attributed to Dave Ulrich): an “equation” where
  • AI multiplied by AI/ingenuity → efficiency
  • Presented conceptually as: AI × human ingenuity = efficiency
• Productivity scoring analogy:
  • If tool and user are both around 9, then 9×9 ≈ 80-something (“A+”).
  • If 10×9 = 90 (“what does that mean your score will be?”).
  • If the human component is 2 or 1, results shrink accordingly (e.g., 2×10).
  • If someone is 0, the outcome becomes 0.

“ACCOUNT” framework for safe, effective AI in finance

A structured methodology for governance and implementation, especially for cash flow forecasting:

1. A — Goal

   • Start with the business finance objective (not “implement AI” for its own sake).
   • Examples: improved prognosis, time efficiency, accuracy, early identification of risks/opportunities.

2. C — Capabilities

   • Ensure the team understands the model and can apply it with commercial/financial judgment.
   • Provide training if capabilities are lacking.

3. C — Control

   • Use controls and maintain a “person in the loop.”
   • Validate AI-generated numbers so they can be explained to auditors/board.
   • AI can help flag anomalies across many transactions, but governance remains essential.

4. O — Operational workflows

   • Redesign workflows around AI rather than bolting AI onto spreadsheet-heavy processes.
   • Change management spans people, processes, systems.
   • Guiding actions: simplify, standardize, automate.

5. U — Basic data

   • Ensure data integrity/completeness and documented processes.
   • “If it wasn’t documented, it wasn’t done.”
   • Incomplete/low-quality data can lead to creating a bad forecast faster.

6. N — Numbers (KPIs/metrics)

   • Track whether AI improves outcomes:
     • time to forecast
     • variance reduction (actual vs predicted)
     • identification of predictive risk
   • If metrics don’t tell a compelling story, reconsider strategy.

7. T — Trust

   • Trust is earned through consistent reliability.
   • Leadership (e.g., CFO/management) must be able to:
     • explain and defend assumptions
     • operationalize how the system is used
   • Outputs must be presentable to the board and defensible to auditors.

Key examples and recommendations/cautions

• Speed is a double-edged sword:
  • AI can produce correct results quickly,
  • but can also produce wrong results quickly if the setup is flawed.

Cash flow forecasting example

• Move from manually built spreadsheets to an AI-assisted connected model using:
  • income statement, balance sheet, cash flow
• AI supports scenario changes (e.g., adjusting credit days like 30 or 45) without rewriting formulas from scratch.
• Emphasis remains on assumptions and the forecasting goal.

Tool selection (no single “best” model)

Different models are suggested depending on the task:

• Claude: research and collaboration / moving content between apps
• ChatGPT: writing reports
• Google Gemini: strong at image generation
• DeepSeek: strong at research

The takeaway: compare strengths, rather than declaring one universally superior.

Crypto assets mentioned (tickers)

• Bitcoin
• Bitcoin Gold
• Bitcoin Cash
• Ripple (XRP)

(No stocks/ETFs/bonds/commodities were mentioned.)

Key numbers and timelines mentioned

• Crypto story
  • Bitcoin position: 5 units held around 2016–2017
  • Sold everything at the end of 2017 after a hard fork
  • Later regret mentioned when Bitcoin reached $100,000 (exact date not provided)
• AI research example
  • Synthesis response time: ~15 minutes
• UK labor/macro context
  • Growth described as above zero / around ~1% (approximate)
  • Unemployed youth: over 1 million (approximate; year not specified)
• Book release
  • Book available first week of September (year not explicitly stated)
  • Discussion occurs “this year”
• LinkedIn newsletter
  • Launched in January

Disclosures / disclaimers

• No explicit “not financial advice” disclaimer was included in the provided subtitles.
• The guest’s content is primarily framed as professional/operational guidance on AI in finance, not investment advice.

Presenters and sources mentioned

People

• Kevin Appleby (host)
• Ademola Odewade / Ademola Odowada (guest; CEO, K-Dems Consulting Ltd.)

Referenced sources and companies

• Money 2020 (conference)
• Dave Ulrich
• Ron Schiavone
• Major firms: PwC, Deloitte, KPMG, EY
• Research/consulting: Forrester, Bain, McKinsey
• AI tools/models: Claude, ChatGPT, Google Gemini, DeepSeek
• Apps mentioned: Notion, HubSpot