The team is the hard part, not the tool
Buying the tool is easy. Getting your people to use it well - and to trust each other's AI-assisted work - is the real leadership task. The finance professionals who thrive with AI are not the ones who prompt fastest; they are the ones whose judgment and skepticism got sharper because the drudgery got cheaper. Your job as a leader is to build that: shift the roles deliberately, grow the four skills that matter, and manage the emotional change from fear to curiosity. Tonight you build a skills-map for your actual team and pick two change moves you can make this quarter.
The roles are shifting 6 min live
No finance job title disappears overnight, but what people spend their hours on changes fast. The best way to lead this is to name the shift out loud, so your team sees a better job on the other side rather than a threat.
LiveThe AI-literate analyst and the finance-data translator3 min▶
Two role shifts matter most, and both are promotions in disguise.
- The AI-literate analyst. Today many analysts spend the majority of their time gathering, cleaning, and formatting data. With AI doing more of that, the job tilts toward the part that was always the point: interpreting the numbers, spotting what matters, and advising the business. The skill mix shifts from spreadsheet mechanics toward judgment and communication.
- The finance-to-data translator. A newly valuable role - the person who speaks both finance and data fluently, can frame a business question so a tool can help, and can sanity-check what comes back. This is often an existing team member who leans in, not a new hire.
The analyst who got her Fridays back. A senior FP&A analyst spent most of Thursday and Friday assembling the weekly pack. With AI drafting the first cut of commentary from numbers she supplied, that dropped to an afternoon - and the freed time went into a segment-profitability question the business had been asking for a year. Same person, same title, far more valuable work. That is the story to tell your team, not "we are automating your job".
LiveCitizen developers and where a data/AI partner sits3 min▶
As tools get easier, finance people start building their own small workflows - a prompt template, a spreadsheet copilot flow, a document Q&A helper. This is powerful and needs guardrails.
- Citizen developers with guardrails. Encourage building, but inside the fence: approved tools only, no real data in ungoverned tools, and anything that touches a decision gets the same human sign-off as everything else (the a3 stack). Unguarded citizen development is how shadow AI spreads.
- Where a data/AI partner sits. Most finance teams benefit from at least one person or a small central group who owns tooling, policy, and training - a bridge to IT and data teams. Whether that sits inside finance or is a shared service is the operating-model question in Part 3.
The skills to build 5 min live
Four skills separate a finance professional who uses AI well from one who is quietly at risk with it. Notice that only one is about the tool - the rest are finance judgment, sharpened. That is reassuring news for a team worried they need to become engineers.
LiveThe four skills, and how to build each3 min▶
| Skill | How to build it in your team |
|---|---|
| Verification discipline - the #1 skill | Make checking AI output a normal, expected step. Teach the "ground it, show the work, verify the math" habit. Reward the person who catches the error, not just the one who ships fast. |
| Prompt literacy | Short internal training plus a shared library of prompts that work for your recurring tasks. This is the most learnable skill - a few hours moves most people a long way. |
| Data fluency | Comfort reading, questioning, and framing data - what a number means, where it came from, what could make it wrong. Build via real work, not abstract courses. |
| Judgment & skepticism | The finance instinct for "that does not look right". Protect and grow it - AI makes it more valuable, not less. Pair juniors with seniors who model it. |
Self-studyReskill or hire? Mostly reskill2 min read▶
Leaders often reach for hiring first. For most finance teams, reskilling wins:
- Reskill your existing team for prompt literacy, verification, and data fluency. These are learnable, and your people already have the scarce thing - deep finance context and business relationships an outside hire would take a year to build.
- Hire selectively for genuinely new capability you cannot grow fast enough - often a single data/AI translator or a lead who has done this before, to accelerate everyone else.
- The trap to avoid: hiring a data scientist to "do AI" while the finance team stays passive. Adoption sticks when your finance people own it, supported by a specialist - not when it is outsourced to one.
Change management: fear to curiosity 5 min live
The barrier to AI in finance is rarely technical. It is fear - of looking foolish, of being replaced, of making a costly mistake. Good change leadership converts that fear into curiosity, and it follows a pattern you can run deliberately.
LiveThe moves that turn fear into curiosity3 min▶
- Name the fear, then reframe the role. Say plainly that AI changes the work, and show the better job on the other side (Part 1). People fight a threat and lean into an upgrade.
- Champions and power users. Find the two or three people already curious, give them time and cover, and let them teach peers. Adoption spreads sideways far better than top-down.
- Safe sandboxes with synthetic data. Give people a place to experiment with fake or public data where nothing can break and no rail is crossed. Fear drops fast when the stakes are zero.
- Quick wins first. Start with the drudgery everyone hates - a first-draft commentary, a reconciliation helper. An early, visible time-saver buys you permission for the harder changes.
- Make verification a badge of honor. Celebrate the catch. The person who found the AI's error is the hero of the story, not the bottleneck.
The sandbox that thawed a frozen team. A controller whose team was split between fear and quiet rule-breaking gave everyone a synthetic dataset and one hour a week to "try to break it". Within a month the skeptics were the loudest advocates, because they had proven to themselves what was safe and what was not. No mandate could have done what curiosity did.
LiveThe operating model: center-of-excellence vs embedded2 min▶
You also have a structural choice about who owns tools, policy, and training. Two common models, and most teams end up blending them.
- Center-of-excellence (CoE). A small central group owns standards, tooling, and training for the whole function. Strength: consistency, governance, no duplicated effort. Risk: it becomes a distant gatekeeper the front line routes around.
- Embedded. AI capability lives inside each team - FP&A, controllership, treasury each own their own use. Strength: close to the real work, fast. Risk: inconsistency and gaps in governance.
- The common answer: a light CoE that owns policy, tool tiers, and training (the a3 stack), with embedded champions who own use within each team. Central sets the rails; the edge does the driving.
Build your team skills-map + a change plan ◐ 20-30 min
No tools - a grid and honest reflection. Map your team against the four skills, then commit to two change moves for the quarter. This is the artifact you take into your next team or one-to-one meeting.
- Score the four skills. For your team as a whole, rate prompt literacy, verification discipline, data fluency, and judgment/skepticism as Strong, Developing, or Gap. Verification is the one to be hardest on - it is the #1 skill and the easiest to skip.
- Name your people. Who are your two or three natural champions? Who is most anxious and needs a sandbox and a quick win first? Write real names.
- Pick two change moves for this quarter. Choose from: launch a synthetic-data sandbox, appoint champions, ship one quick win on hated drudgery, run a prompt-literacy session, or make verification a visible badge. Two, with owners and dates - not five aspirations.
- Choose your operating model. Decide, at least provisionally: light CoE, embedded, or the blend. Name who would own policy/tools/training and who champions each team.
Official sources covered
This session teaches the people-and-adoption content from the finance-AI programs that treat team readiness as a first-class topic. Certificates, graded assessments, and paid tooling stay with the official programs - said honestly. This session covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · What is the single most important individual skill for a finance professional using AI?
AI makes drafting cheap and judgment scarce. The professional who verifies output and trusts their "that does not look right" instinct is the one who keeps the firm out of trouble. It is the #1 skill.
2 · Your team is split between fear of AI and quietly using free chatbots. What is the best change move?
Fear turns to curiosity when the stakes are zero (sandbox), peers lead (champions), and the value is obvious fast (quick win). Mandates breed resentment; waiting lets shadow AI spread.
3 · What is the trade-off between a center-of-excellence and an embedded operating model?
CoE centralises standards and governance but can drift from the front line; embedded is close and fast but uneven. The common answer is a light CoE for policy and training plus embedded champions per team.