# Wall Street Ties Pay to AI Skills. How to Measure It Fairly > Selby Jennings: nearly 25% of finance firms tie pay to AI proficiency, 30% make staff disclose AI use. Why usage metrics get gamed, and a fair 4-skill rubric. Source: https://wecallshotgun.com/blog/ai-skills-bonuses-performance-reviews-wall-street-2026 Language: en Author: Toni Dos Santos Published: 2026-10-15 Updated: 2026-10-15 --- AI Tools # Wall Street Ties Bonuses to AI Skills. How to Measure Them Without Paying for Clicks By [Toni Dos Santos](https://wecallshotgun.com/toni-dos-santos) • Oct 15, 2026 • 18 min read • **Summary:** A Selby Jennings survey of 500 finance respondents, reported by Reuters (Nupur Anand) on 14 October 2026, found that nearly a quarter of financial firms tie raises, bonuses or promotions to employees' AI proficiency, another 30% require staff to disclose AI tool use in formal performance reviews, 42% have formal criteria to evaluate AI skills and 30% have teams or roles tracking whether AI investments pay off; nearly a third lack internal experts who can assess AI skills. Selby Jennings' Dylan Gomez said there's no established formula yet for how much pay is tied to AI, and Jacob Summers said the practice is more common at fintechs and smaller firms. The risk: usage is the easiest thing to count and the easiest to game. Meta's employee-built "Claudeonomics" token leaderboard came down in April 2026, Amazon retired KiroRank in May 2026 after staff ran pointless tasks through AI agents (tokenmaxxing) and now tracks useful code shipped, Accenture tracks senior staff's weekly AI log-ins for promotion, and METR's 2025 trial showed developers believed AI made them 20% faster when it made them 19% slower, so self-disclosure is a weak skill measure. Our method: assess AI proficiency on one recurring responsibility per role, take a baseline before anyone knows it counts, review three real work samples against Teach Them to Drive's four skills (Frame, Prompt, Evaluate, Iterate), check how people review AI output (Check-Challenge-Confirm), measure at 30, 60 and 90 days, and credit workflows that spread. Usage data belongs in support and cost control, disclosure in governance, and workflow change in the bonus formula. Nearly a quarter of financial firms now tie raises, bonuses or promotions to employees' AI proficiency, according to a Selby Jennings survey of 500 finance respondents reported by Reuters on 14 October 2026. Another 30% make staff disclose their AI use in formal performance reviews, and 42% have written criteria for judging AI skills. The easiest thing to count is usage. Pay for usage and you get usage. A fair test of AI proficiency looks at one recurring piece of someone's work and asks whether it got faster or better against a baseline. I started in corporate banking at Crédit du Nord, at the bottom of the ladder, and later ran product marketing inside a regulated bank. Nobody who has worked in a bank needs to be told what happens to a number once it sits in the bonus formula. It gets hit. Today I run AI adoption programmes for teams in the UK and France at We Call Shotgun. The Reuters story asks the question every manager with an AI licence budget will meet at the next review cycle: how do you judge someone's AI skill fairly, when the only data on the dashboard is how often they opened the tool? ## What did the Selby Jennings survey find? That AI skill has entered the pay conversation in finance before anyone agreed how to measure it. Selby Jennings is a financial-services recruitment firm. Reuters' Nupur Anand reported its findings on 14 October 2026 under the headline ["Wall Street firms tie bonuses, promotions to AI skills, survey finds"](https://www.reuters.com/business/finance/wall-street-firms-tie-bonuses-promotions-ai-skills-survey-finds-2026-10-14/). Every figure in the coverage, with the people each one counts: | Finding | Figure | Who it counts | | --- | --- | --- | | Tie raises, bonuses or promotions to AI proficiency | **Nearly 25%** | Surveyed firms. Reported as "nearly a quarter" | | Require employees to disclose AI tool use in formal performance reviews | **30%** | Surveyed firms. Reported as "another 30%", which suggests a separate group from the quarter above | | Have introduced formal criteria to evaluate employees' AI skills | **42%** | Respondents | | Created dedicated teams or roles to track whether AI investments deliver results | **30%** | Respondents. A separate question from the disclosure 30% | | Lack internal experts who can reliably assess AI skills | **Nearly a third** | Respondents, from coverage of the full report | ![Bar chart of how US financial firms handle employee AI skills, from the Selby Jennings survey of 500 finance respondents reported by Reuters on 14 October 2026. 42 percent have formal criteria to evaluate AI skills. 30 percent require employees to disclose AI tool use in formal performance reviews. 30 percent have dedicated teams or roles tracking whether AI investments deliver results. Nearly 25 percent tie raises, bonuses or promotions to AI proficiency.](https://wecallshotgun.com/images/blog/finance-ai-skills-pay-survey-2026.svg) The sample is described three ways. Reuters says 500 executives across the finance industry. Selby Jennings' own [AI Talent Pulse Report](https://www.selbyjennings.com/insights/ai-talent-pulse-report/) page describes 500 HR professionals in finance, polled through Censuswide. [Semafor](https://www.semafor.com/article/10/14/2026/a-quarter-of-financial-firms-tie-pay-and-promotions-to-ai-proficiency-survey) calls it a survey of 500 companies. Either way, the respondents are reporting their firm's policy, which is what the figures measure. Dylan Gomez, an executive director at Selby Jennings, told Reuters that firms have moved past checking whether employees use AI tools and now track its effect on output. He also said there's no established formula yet for how much compensation is tied to AI usage and performance, and he expects one to take shape by next year. Jacob Summers, who heads technology recruiting at the firm, told Semafor the practice is more common at fintechs and smaller firms than at the big banks and trading houses. So the formula is being written now, mostly at firms small enough to change it quickly. That's the window to get it right. ## Why does tying pay to AI usage backfire? Because usage is the easiest number to move without doing better work. Goodhart's law, in the anthropologist Marilyn Strathern's phrasing: "When a measure becomes a target, it ceases to be a good measure." Some of the biggest employers in tech and finance ran the experiment this year, in public. - **Meta.** [Fortune reported on 9 April 2026](https://fortune.com/2026/04/09/meta-killed-employee-ai-token-dashboard/) that a Meta employee had built an internal leaderboard, "Claudeonomics", ranking the top 250 token users among more than 85,000 staff and handing out titles such as "Token Legend" and "Cache Wizard". It came down two days after the story broke. - **Amazon.** In late May 2026 the Financial Times reported that Amazon had retired KiroRank, a staff-built leaderboard of AI usage on its Kiro developer platform, after employees ran pointless tasks through AI agents to climb it. The practice has a name now: tokenmaxxing. Senior vice president Dave Treadwell reportedly told staff not to use AI for the sake of using AI. Amazon now tracks what it calls normalised deployments, evidence that engineers use AI to ship useful code. - **Accenture.** In February 2026 the Financial Times reported that Accenture [tracks senior staff's weekly log-ins to its internal AI tools](https://www.computing.co.uk/news/2026/ai/accenture-ties-progression-to-ai) and requires "regular adoption" for promotion to leadership roles. Some staff told the FT the tools were "broken slop generators". - **JPMorgan.** Business Insider reported in March 2026 that the bank [tracks how often its engineers and technologists use AI tools](https://www.artificialintelligence-news.com/news/jpmorgan-begins-tracking-how-employees-use-ai-at-work/), sorting them into light, heavy and non-users, and that the data may influence performance reviews. Neither leaderboard was an official pay metric. Amazon said its token statistics wouldn't inform performance reviews. People gamed them anyway. My reading, and it's an opinion: if an unofficial ranking gets gamed within weeks, a line in the bonus formula will get gamed harder. Usage metrics also miss where the cost lands. Templafy's 2026 survey of 2,000 UK and US knowledge workers found people spend almost four hours a week reviewing and correcting AI-generated documents, and 48% say that leaves little time saved ([our breakdown of the Templafy figures](https://wecallshotgun.com/blog/uk-workers-edit-ai-documents-templafy-2026)). Researchers at BetterUp Labs and the Stanford Social Media Lab, writing in [Harvard Business Review in September 2025](https://hbr.org/2025/09/ai-generated-workslop-is-destroying-productivity), named the problem: workslop, "AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task". Their respondents put the clean-up at nearly two hours per incident. A usage bonus pays twice. Once in tokens, once in the hours someone downstream spends fixing the output. ### Why self-disclosure is a weak measure of AI skill Disclosure asks people how much AI helped. People are poor judges of that, experts included. In a randomised trial [published by METR in July 2025](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/), 16 experienced open-source developers expected AI tools to speed them up by 24%. On 246 real tasks they were 19% slower with AI, and afterwards they still believed they'd been about 20% faster. Disclosure has a real job in a regulated firm: knowing which work AI touched. That's a governance control. Scoring it as skill turns a compliance form into a self-assessment, and the incentive runs one way. ## What's the difference between AI usage and AI proficiency? Usage counts how often someone opens the tool. Proficiency is whether a recurring piece of their work comes out faster, better or cheaper because of it, and whether they can tell when the output is wrong. Our glossary draws the same line at team level as [activation versus adoption](https://wecallshotgun.com/ai-adoption-glossary#ai-activation-vs-adoption): activation is a licence metric, adoption is a workflow that runs differently because of AI. A performance review needs the individual version of the same distinction. ![Diagram of four levels of AI measurement and what each one proves. Level 1, logged in: proves access, gamed by opening the tool once a week. Level 2, weekly use and token counts: proves volume, gamed by trivial prompts as at Amazon's KiroRank. Level 3, runs Frame, Prompt, Evaluate and Iterate on a recurring task: proves proficiency, evidenced by work samples. Level 4, the workflow runs differently against a baseline: proves value, evidenced by cycle time and error rate. Only levels 3 and 4 belong in pay decisions. We Call Shotgun framework from Teach Them to Drive.](https://wecallshotgun.com/images/blog/ai-usage-vs-proficiency-ladder.svg) Every AI metric a firm can collect, with what it proves and how it gets gamed: | What firms count | What it proves | How it gets gamed | What to use it for | | --- | --- | --- | --- | | Log-ins, weekly active use | Access | Open the tool every Monday | Licence management | | Prompt or token counts | Volume | Trivial tasks, agents doing busywork | Cost control | | Self-disclosure in the review | What people believe | Everyone reports "a lot" | Governance: which work AI touched | | Course completions, certificates | Attendance | Click through the modules | Compliance records, such as [AI literacy under Article 4 of the EU AI Act](https://wecallshotgun.com/blog/eu-ai-act-ai-literacy-article-4-risks-action-plan) | | Work samples on a recurring task | The four skills, on real work | Hard to fake: the reviewer sees the prompt, the rejected output and the fix | Proficiency rating | | Before-and-after on the workflow | Value | Sandbagging the baseline, so take it before anyone knows it counts | Bonus and promotion | The first four rows are what most AI dashboards show. [Claude, ChatGPT, Copilot and Gemini all report usage by user](https://wecallshotgun.com/blog/ai-spend-monitoring-dashboards-claude-chatgpt-copilot-gemini), which makes them good for spend and support. The last two rows need a manager and a baseline. No admin console produces them. ## What does Teach Them to Drive say about paying for AI skills? That pay has to follow the value, and the value moved. Three ideas from the book apply directly. **Value Migration.** After AI lands, the value a team creates moves away from execution speed toward judgment, framing, evaluation and taste. The book's glossary entry for [Value Migration](https://wecallshotgun.com/ai-adoption-glossary#value-migration) spells out the pay consequence: > "Roles, comp bands, and promotion criteria all need to follow the migration. If a senior IC is still being paid for first-draft speed two years into the AI era, the org is mispricing the work." Toni Dos Santos, Teach Them to Drive: The AI Adoption Playbook for Teams **Skill Inversion.** [Skill Inversion](https://wecallshotgun.com/ai-adoption-glossary#skill-inversion) is the mechanism behind the migration: AI drops the cost of a credible first draft to near zero, so judgment and evaluation become the scarce skills. A usage metric rewards drafting volume. It pays for the part of the job that just got cheap. **The four core skills.** The book names four skills that survive every tool change: Frame (define the real job before reaching for the prompt), Prompt (turn it into instructions a model can execute), Evaluate (read the output critically, catch hallucinations, omissions and tone drift) and Iterate (refine, re-prompt, ship). Usage logs see one of the four. Prompt leaves a trace. Frame happens before the tool opens, Evaluate after it answers, and Iterate shows up in what ships. A token count can't separate the analyst who sent 200 prompts and shipped nothing from the one who sent 12 and rebuilt the month-end pack. > "The tool was never the problem. Nobody taught them to drive." Toni Dos Santos, Teach Them to Drive: The AI Adoption Playbook for Teams Paying by the mile rewards whoever sits in the car longest. ## How do you assess AI proficiency fairly in a performance review? On one recurring responsibility per person, with a baseline taken before anyone knew it would count, and with evidence from real work. This is the sequence we run with teams, adapted to a review cycle. - **Pick one recurring responsibility per role.** The [Painful Tuesday](https://wecallshotgun.com/ai-adoption-glossary#painful-tuesday): the task everyone complains about every week. In finance that's often the credit memo, the client review pack, the reconciliation exceptions, the month-end commentary or the alert write-ups in compliance. - **Take the baseline now.** Cycle time, rounds of correction, errors found downstream. Two weeks of honest timing, before anyone knows it feeds a review. - **Review three work samples against the four skills.** A 30-minute conversation over the last three pieces of AI-assisted work: what the job was, what the person gave the model, what they threw away, what they changed. - **Check the checks.** We use [Check-Challenge-Confirm](https://wecallshotgun.com/ai-adoption-glossary#check-challenge-confirm) for anything going to a client, a regulator or an executive: check the facts, challenge the reasoning, confirm the tone and policy fit. In a regulated firm, the person who catches the invented figure before it reaches a client is worth more than the one who generated it fastest. - **Measure at 30, 60 and 90 days against the baseline.** Same units. If the number didn't move, look at the process and the context the tool was given before you look at the person. - **Credit what spreads.** The 90-Day Playbook ends with codifying the workflow and handing it to the team. The analyst whose prompt and checklist the whole desk now uses has multiplied the gain. Score that. The rubric we hand managers, one row per skill: | Skill | What the reviewer asks | Weak evidence | Strong evidence | | --- | --- | --- | --- | | **Frame** | "What job were you trying to get done, and for whom?" | "I use it for emails" | Names the task, the reader, the constraint and what done looks like | | **Prompt** | "Show me what you gave the model." | One-line requests with no context | Reusable instructions with sources, examples and the output format | | **Evaluate** | "What did you reject or correct last time?" | Nothing, ever | Names the error caught and the check that caught it | | **Iterate** | "What's different in this workflow since last quarter?" | Same output, more prompts | A before-and-after number and a written process someone else now uses | ### What goes in the bonus formula, and what stays out Our position, and it's an opinion: - **In:** the workflow's change against its baseline, the four-skill rating from work samples, and whether the improved workflow spread to the team. - **Out:** log-ins, prompt counts, tokens and hours in the tool. They belong in the manager's support conversation, where a low number means someone needs help with a workflow. - **On its own line:** disclosure. Keep it as a governance record with its own purpose, and keep it out of the skill score. The assessor is the line manager. Nearly a third of the firms in the Selby Jennings survey say they lack internal experts who can assess AI skills, and the rubric above is built so a manager who can read a credit memo can run it. [Our manager playbook](https://wecallshotgun.com/blog/ai-skeptic-to-champion-manager-playbook) covers the rest of that job, and [Microsoft's account of its own Copilot rollout](https://wecallshotgun.com/blog/why-ai-training-fails-microsoft-lessons) shows the measurement in practice: it judged 687 sellers on revenue per account manager and close rates, compared with low-usage sellers. That sequence is the method behind our [enterprise AI adoption programmes](https://wecallshotgun.com/enterprise) and our [work with financial services teams](https://wecallshotgun.com/ai-training-financial-services), tool-agnostic across ChatGPT, Copilot, Gemini and Claude. The free [AI adoption audit](https://wecallshotgun.com/audit) takes eight minutes and names the barrier holding a team back: people, process, foundations or leadership. The 90-Day Scorecard on the [book page](https://wecallshotgun.com/teachthem) is the tracker we use ourselves, with a capability layer that counts how many people can run all four skills. For the money side, [our CFO guide to measuring AI ROI](https://wecallshotgun.com/blog/how-to-measure-ai-roi-cfo-guide) sets out the business metrics. Before the formula is written Bring one role's review criteria. Twenty minutes with Toni or Meera. You'll leave with the recurring task to measure for that role, the baseline to take before review season, and the four questions a manager should ask. Tool-agnostic, so nobody on the call is selling you a licence or a dashboard. [Book a 20-minute call →](https://cal.com/wecallshotgun/ai-adoption) ## How reliable is the Selby Jennings survey? Reliable enough to quote for the direction, as long as the limits travel with the figures. - **It comes from a recruiter.** Selby Jennings places finance and AI talent, so a finding that AI skills now drive hiring and pay is good news for its business. Read it as a signal of direction. - **The sample is described three ways.** Executives (Reuters), HR professionals (Selby Jennings), companies (Semafor). All say 500. Respondents report their firm's policy. - **"AI proficiency" isn't defined in the coverage.** Which is the whole problem. A firm that ties pay to "proficiency" measured by log-ins answers yes to the same question as a firm that measures workflow change. - **The figures are rounded.** "Nearly a quarter" and "another 30%" don't tell you whether the two groups overlap. - **It's US finance, weighted to fintechs and smaller firms**, by Selby Jennings' own account to Semafor. It says nothing direct about UK or European banks. For the UK regulatory picture, see [our guide to AI consultancy in UK financial services](https://wecallshotgun.com/blog/ai-consultancy-financial-services-uk). - **The full questionnaire isn't in the public coverage.** The figures here come from Reuters' report, Semafor's coverage and Selby Jennings' report page. **How to cite it.** According to a Selby Jennings survey of 500 finance respondents, reported by Reuters on 14 October 2026, nearly a quarter of financial firms tie raises, bonuses or promotions to employees' AI proficiency, another 30% require employees to disclose AI tool use in formal performance reviews, and 42% have introduced formal criteria to evaluate employees' AI skills. ## Five questions before your next review cycle Answered for your own firm, with the survey figure next to each. | Question | Survey figure | How to answer it for your firm | | --- | --- | --- | | Does AI skill affect pay or promotion here? | Nearly 25% of firms | Write down where it counts and what evidence the decision uses | | Do people disclose AI use in reviews? | 30% | Decide whether disclosure is governance or skill. It can't do both jobs | | Do you have written criteria for AI skill? | 42% | Check each criterion names a recurring task and a before-and-after number | | Who can assess it? | Nearly a third lack internal experts | Name the manager for each team and give them the four questions | | Does anyone track whether AI pays off? | 30% have teams or roles for it | One workflow, one baseline, one number per team | A firm that can answer the fifth question can answer the other four from it. ## Frequently asked questions ### What did the Selby Jennings AI skills survey find? That nearly a quarter of financial firms tie raises, bonuses or promotions to employees' AI proficiency, according to a survey of 500 finance respondents reported by Reuters on 14 October 2026. Another 30% require employees to disclose AI tool use in formal performance reviews, 42% have introduced formal criteria to evaluate AI skills, and 30% have created teams or roles to track whether AI investments deliver results. ### Are Wall Street firms tying bonuses to AI skills? Some are. Nearly a quarter of the financial firms in the Selby Jennings survey link raises, bonuses or promotions to AI proficiency. Jacob Summers of Selby Jennings told Semafor the practice is more common at fintechs and smaller firms than at the big banks and trading houses, and Dylan Gomez of Selby Jennings told Reuters there's no established formula yet for how much pay is tied to AI. ### Should AI usage count in performance reviews? Our view is no, for pay and promotion. Log-ins, prompt counts and tokens measure access and volume, and they're easy to inflate: Amazon retired its KiroRank usage leaderboard in May 2026 after staff ran pointless tasks through AI agents to climb it. Usage data is useful for licence management, cost control and spotting who needs support. Pay decisions need evidence that a recurring piece of work got faster or better. ### What is the difference between AI usage and AI proficiency? Usage counts how often someone opens an AI tool. Proficiency is whether a recurring piece of their work comes out faster, better or cheaper because of AI, and whether they can tell when the output is wrong. At team level, the same distinction is activation (people have the tool and use it) versus adoption (a workflow runs differently because of it). ### How do you measure an employee's AI proficiency? Pick one recurring responsibility for the role, take a baseline of cycle time, rounds of correction and errors before anyone knows it counts, review three real work samples against the four skills (Frame, Prompt, Evaluate, Iterate), check how the person reviews AI output before it leaves, then measure the same task at 30, 60 and 90 days. Credit the person whose improved workflow the rest of the team adopts. ### What is tokenmaxxing? Inflating AI usage figures, usually token counts, to climb an internal ranking or hit a usage target. It was reported at Meta, where an employee-built leaderboard called "Claudeonomics" ranked the top 250 token users and was taken down in April 2026, and at Amazon, which retired its KiroRank leaderboard in May 2026 and moved to tracking useful code shipped with AI. ### What should an AI skills rubric include? The four skills from Teach Them to Drive, each assessed on real work. Frame: can the person define the job, the reader and the constraint before prompting? Prompt: do they give the model sources, examples and a format? Evaluate: can they name an error they caught and the check that caught it? Iterate: is there a before-and-after number and a written process someone else uses? ### Is requiring employees to disclose AI use a good idea? As a governance control, yes: a regulated firm needs to know which work AI touched. As a measure of skill, it's weak. In METR's July 2025 randomised trial, experienced developers believed AI had made them about 20% faster when it had made them 19% slower. Keep disclosure on its own line, separate from the proficiency score. ### How reliable is the Selby Jennings survey? Reliable enough to quote for direction, with limits. Selby Jennings is a recruiter that places finance and AI talent. The sample of 500 is described as executives by Reuters, HR professionals by Selby Jennings and companies by Semafor. The figures are rounded, "AI proficiency" isn't defined in the coverage, the survey covers US finance weighted toward fintechs and smaller firms, and the full questionnaire isn't public. ## Sources and further reading - Nupur Anand, [Wall Street firms tie bonuses, promotions to AI skills, survey finds](https://www.reuters.com/business/finance/wall-street-firms-tie-bonuses-promotions-ai-skills-survey-finds-2026-10-14/), Reuters, 14 October 2026 - [A quarter of financial firms tie pay and promotions to AI proficiency: survey](https://www.semafor.com/article/10/14/2026/a-quarter-of-financial-firms-tie-pay-and-promotions-to-ai-proficiency-survey), Semafor, 14 October 2026 - [AI Talent Pulse Report: Financial Services](https://www.selbyjennings.com/insights/ai-talent-pulse-report/), Selby Jennings - [Meta's employee AI token dashboard](https://fortune.com/2026/04/09/meta-killed-employee-ai-token-dashboard/), Fortune, 9 April 2026 - [Accenture ties career progression to AI use](https://www.computing.co.uk/news/2026/ai/accenture-ties-progression-to-ai), Computing, February 2026, after the Financial Times - [JPMorgan begins tracking how employees use AI at work](https://www.artificialintelligence-news.com/news/jpmorgan-begins-tracking-how-employees-use-ai-at-work/), AI News, March 2026, after Business Insider - Kate Niederhoffer, Gabriella Rosen Kellerman, Angela Lee, Alex Liebscher, Kristina Rapuano and Jeffrey T. Hancock, [AI-Generated "Workslop" Is Destroying Productivity](https://hbr.org/2025/09/ai-generated-workslop-is-destroying-productivity), Harvard Business Review, September 2025 - [Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity](https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/), METR, 10 July 2025 - [96% of UK Workers Still Edit AI-Generated Documents](https://wecallshotgun.com/blog/uk-workers-edit-ai-documents-templafy-2026), We Call Shotgun, 4 October 2026 - [Why AI Training Fails After the Workshop: What Microsoft Learned From 200,000 Employees](https://wecallshotgun.com/blog/why-ai-training-fails-microsoft-lessons), We Call Shotgun, 24 September 2026 - [AI Spend Visibility: How to Monitor and Control AI Usage Across Claude, ChatGPT, Copilot and Gemini](https://wecallshotgun.com/blog/ai-spend-monitoring-dashboards-claude-chatgpt-copilot-gemini), We Call Shotgun, 3 July 2026 - Toni Dos Santos, [Teach Them to Drive: The AI Adoption Playbook for Teams](https://www.amazon.com/dp/B0GYSSFNPK/ref=tmm_pap_swatch_0) **About the author.** [Toni Dos Santos](https://wecallshotgun.com/toni-dos-santos) is co-founder of We Call Shotgun, an AI adoption practice for teams in the UK and France, with Meera Sanghvi. He wrote [Teach Them to Drive: The AI Adoption Playbook for Teams](https://wecallshotgun.com/teachthem). Before that he worked in corporate banking at Crédit du Nord (Société Générale), ran B2B product marketing at Banque BCP, and was a B2B product manager at lemlist. ## Keep Reading [AI Tools Oct 30, 2026 Microsoft Copilot Notebooks for teams: how to replace scattered chats with one shared workspace (2026 guide) Copilot Chat starts every conversation from zero, and a team's best prompt lives in one person's notes file. A Copilot Notebook, the shared workspace Microsoft also calls a project workspace, keeps the references, the instructions and the outputs for one piece of work in one place. Where chat fails teams, what a notebook changes, the October 2026 releases, a six-step setup, a worked monthly-report example and how to create Word, PowerPoint, Excel and audio artifacts. 28 min read Read →](https://wecallshotgun.com/blog/microsoft-copilot-notebooks-for-teams-shared-workspace) [AI Tools Oct 26, 2026 The Average UK Business Spends £20.6m a Year on AI. 11% Invest in It Strategically The average UK business spends £20.6 million a year on AI, up 27% on 2025, and 11% of UK businesses invest in it strategically, according to SAP and Oxford Economics' Value of AI in the UK 2026. 70% are piloting agents and 49% say agents are being deployed faster than they can be standardised and governed. Every figure with what it counts, what the ONS and DSIT data add, what Teach Them to Drive says about budget without strategy, and the six-step sequence that turns a budget line into a measured workflow. 21 min read Read →](https://wecallshotgun.com/blog/uk-ai-spending-strategy-governance-sap-2026) [AI Tools Oct 21, 2026 What Can ChatGPT Enterprise Do Beyond Writing Emails? 5 Features to Use at Work Five ChatGPT Enterprise features for the work around the email: Projects for the brief you keep repeating, scheduled tasks for the weekly check you run by hand, skills for the prompt your team keeps re-pasting, the browser extension for pages you keep copying, and deep research for cited decision briefs. Each with a copy-paste prompt and the caveat from OpenAI's own documentation. 19 min read Read →](https://wecallshotgun.com/blog/chatgpt-enterprise-features-for-work) ## Related AI Training Solutions Programs tailored to your role, industry, or function. [Financial Services](https://wecallshotgun.com/ai-training-financial-services) [Healthcare](https://wecallshotgun.com/ai-training-healthcare) [Legal Teams](https://wecallshotgun.com/ai-training-legal) [Operations](https://wecallshotgun.com/ai-training-operations) [Product Teams](https://wecallshotgun.com/ai-training-product) [Marketing](https://wecallshotgun.com/ai-training-marketing) [Sales Teams](https://wecallshotgun.com/ai-training-sales) [Customer Support](https://wecallshotgun.com/ai-training-customer-support) [HR Teams](https://wecallshotgun.com/ai-training-hr) [Data & Analytics](https://wecallshotgun.com/ai-training-data-analytics) [Executives](https://wecallshotgun.com/ai-training-executives) [C-Level](https://wecallshotgun.com/ai-training-c-level) [UK Hub](https://wecallshotgun.com/ai-training-uk) ## Need help with AI Adoption? We take the passenger seat and lead your teams to adopt AI for real work. [Book a Call](https://wecallshotgun.com/#contact)