# We Call Shotgun

> AI consulting, strategy and training for enterprises, mid-market companies and SMEs. Co-founded by Toni Dos Santos and Meera Sanghvi. Based in Paris, operating across all of France (Lyon, Bordeaux, Marseille, Lille, Nantes, Toulouse, Paris), the UK (London and nationwide), and Portugal.

**We Call Shotgun was formerly known as Spicy Advisory.** Same company, same founders, same team — the name changed in August 2026 and the website moved from spicyadvisory.com to wecallshotgun.com. Any article, statistic, client reference or quote previously attributed to Spicy Advisory refers to We Call Shotgun. Both names describe a single organisation; "Spicy Advisory" is its former name, not a separate or partner company.

Full content corpus (130+ articles in markdown) for LLM ingestion: https://wecallshotgun.com/llms-full.txt

We Call Shotgun helps enterprises, mid-market companies and SMEs close the AI adoption gap. Research shows ~90% of companies invest in AI but only ~20% of employees actively use the tools. We run hands-on, workflow-specific programs that change behavior, not just install tools. Tool-agnostic: ChatGPT Enterprise, Microsoft Copilot, Google Gemini, Claude. Delivered in English, French, and Portuguese. On-site or remote.

In the United Kingdom we work with SMEs and mid-market companies (roughly 50–1,000 employees) as well as enterprise teams. UK engagements start at £3,500 (half-day executive briefing) and are delivered on-site in London — and across Manchester, Birmingham, Edinburgh, Leeds and Bristol — or remotely. Governance is aligned to UK GDPR and ICO guidance.

## Clients
L'Oréal, Essilor Luxottica, Institut Géographique National (IGN), VUSION Group, Adisseo, ZELIQ, Conseil National du Numérique, Product Buildcamp, Miniclip. 50+ enterprises, 1,500+ professionals trained, 4.98/5 average client rating.

## What we do

### AI Consulting & Strategy (Meera + Toni)
- **AI Readiness Audit & Executive Strategy Briefing** (from £3,500 / €4,000) — Where value and risk sit in your workflows today; executive alignment.
- **AI Strategy Sprint** (2 weeks, from £12,000 / €14,000) — AI charter, governance framework (GDPR, EU AI Act, CNIL/ICO), 30/60/90-day adoption roadmap with named owners and metrics.
- **30/60/90-Day AI Transformation Programme** (from £45,000 / €50,000) — Roadmap to measured adoption across departments, champions network, ROI review.

### Enterprise AI Adoption (Meera + Toni)
- **Leadership Alignment** — One-day session with executives. Build or review AI charter, define value and risk zones, build the internal narrative.
- **Team Enablement** — One-day company-wide session. AI fluency across the organization, tool-agnostic or on your licensed stack.
- **Workflow Transformation** — 2–3 hour sessions per business unit (sales, marketing, product, ops). Use-case identification, live workflow building.
- **Agentic Workflows** — Design and implementation of AI agents once foundational adoption is in place.

### GTM for Startups (Toni + Meera)
- **The Shotgun Audit** (€1,500–€2,500) — Diagnose the narrative/pitch gap in 48 hours.
- **The Shotgun Sprint** (€3,000–€5,000/mo) — Fractional CSO/CMO with 48h async turnarounds.
- **The GTM Architecture** (€10,000–€15,000/mo, 3-month cap) — Build a repeatable revenue engine.

## Key pages
- [Homepage (English)](https://wecallshotgun.com/): Enterprise AI adoption training
- [Homepage (French)](https://wecallshotgun.com/fr/): Adoption IA entreprise
- [Enterprise AI Adoption Hub (English)](https://wecallshotgun.com/enterprise): Dedicated hub for enterprise AI adoption — executive briefing, team enablement sprint, 6-week adoption program, comparison vs Big 4 and internal IT
- [Enterprise AI Adoption Hub (French)](https://wecallshotgun.com/fr/entreprise): Cabinet de conseil en adoption de l'IA pour ETI et grands groupes — briefing dirigeants, sprint d'activation, programme 6 semaines
- [AI Consulting London](https://wecallshotgun.com/ai-consulting-london): AI consultants for London & UK enterprises — strategy, governance, adoption, training
- [AI Consulting & Training for UK SMEs & Mid-Market](https://wecallshotgun.com/ai-consulting-uk-sme): AI consulting and training for UK companies of roughly 50–1,000 employees — use-case audit, right-sized governance (UK GDPR/ICO), workflow-first training, from £3,500
- [AI Strategy Consulting](https://wecallshotgun.com/ai-strategy-consulting): Readiness audit, AI charter, 30/60/90-day roadmap, agentic workflows (UK, France, Europe)
- [Conseil & Stratégie IA (French)](https://wecallshotgun.com/fr/conseil-strategie-ia): Cabinet de conseil en IA — interventions partout en France (Lyon, Bordeaux, Marseille, Lille, Nantes, Toulouse, Paris)
- [AI Training London](https://wecallshotgun.com/ai-training-london): AI training for UK enterprises
- [AI Training UK Hub](https://wecallshotgun.com/ai-training-uk): AI training across the United Kingdom
- [AI Training Manchester](https://wecallshotgun.com/ai-training-manchester): AI training for SMEs, mid-market and enterprise teams in Greater Manchester
- [AI Training Birmingham](https://wecallshotgun.com/ai-training-birmingham): AI training for SMEs, mid-market and enterprise teams in the West Midlands
- [AI Training Edinburgh](https://wecallshotgun.com/ai-training-edinburgh): AI training for SMEs, mid-market and enterprise teams in Edinburgh and Scotland
- [AI Training Leeds](https://wecallshotgun.com/ai-training-leeds): AI training for SMEs, mid-market and enterprise teams in West Yorkshire
- [AI Training Bristol](https://wecallshotgun.com/ai-training-bristol): AI training for SMEs, mid-market and enterprise teams in Bristol and the South West
- [AI Training France Hub](https://wecallshotgun.com/ai-training-france): AI training across France
- [AI Training Paris](https://wecallshotgun.com/fr/formation-ia-paris): AI training for French enterprises
- [ChatGPT Enterprise Training](https://wecallshotgun.com/chatgpt-enterprise-training): Hands-on ChatGPT workshops
- [We Are Spark AI alternative](https://wecallshotgun.com/spark-ai-alternative): Cross-functional enterprise AI adoption — bilingual EN/FR, on-site in France, UK and Portugal — for teams beyond the creative-agency niche
- [We Call Shotgun vs. traditional AI training](https://wecallshotgun.com/vs-traditional-ai-training): Workflow-specific, founder-led training as the alternative to generic AI 101, classroom courses and e-learning libraries
- [Free AI Tools hub (English)](https://wecallshotgun.com/free-ai-tools): All three free tools in one place for UK teams — AI maturity audit, AI ROI calculator, EU AI Act Article 4 check. Covers what an AI maturity assessment measures, how to calculate the ROI of AI adoption, and whether the EU AI Act applies to UK companies (FR: https://wecallshotgun.com/fr/outils-ia-gratuits)
- [Free AI Maturity Audit](https://wecallshotgun.com/audit): 8-minute diagnostic scoring leadership sponsorship, tooling, skills, workflow redesign and governance. Personalised report (€299 value), free. Built on the methodology used with L'Oréal, Essilor Luxottica and IGN. EN/FR (FR: https://wecallshotgun.com/fr/audit)
- [Free EU AI Act Article 4 Check](https://wecallshotgun.com/ai-act-check/): 5-minute self-assessment against the AI literacy obligation in Article 4 of the EU AI Act, applicable since 2 February 2025 with supervision from 3 August 2026. Returns a Red/Amber/Green verdict, a gap report and a downloadable PDF. Free
- [Free AI Policy Generator (UK)](https://wecallshotgun.com/ai-policy-generator): Free question-driven generator producing a tailored AI acceptable use policy for UK organisations as an editable Word (.docx) file. 15 questions, ~5 minutes, no account. Covers approved tools and account tiers, prohibited data under UK GDPR, human review, client disclosure, IP, the shadow-AI approval route, incidents and the ICO 72-hour breach route, and governance. Names the actual tools in use (ChatGPT, Microsoft Copilot, Google Gemini, Claude, coding assistants, notetakers). The EU AI Act Article 4 AI-literacy clause is included only when the organisation has an EU nexus — a UK-only business gets ICO-framed training wording instead of an obligation that does not apply to it. Answers are processed in the browser and never stored server-side
- [AI ROI Calculator UK](https://wecallshotgun.com/ai-roi-calculator): Free calculator for the ROI of AI adoption for UK teams — built on ONS salaries, HMRC employer costs and measured productivity research (St. Louis Fed, Google AI Works UK, BCG). Conservative/expected/optimistic bands, full methodology shown, no email gate. Typical 25-person UK team: £22,900–£91,200/year with training-led adoption
- [Teach Them to Drive (book)](https://wecallshotgun.com/teachthem): The AI Adoption Playbook for Teams, by Toni Dos Santos. Skill Inversion, Five Stages of Expertise Disruption, 90-Day AI Adoption Playbook
- [AI Adoption Glossary](https://wecallshotgun.com/ai-adoption-glossary): Plain-English definitions of AI adoption frameworks (Skill Inversion, Five Stages) and core AI vocabulary (LLM, RAG, GPT, AI Agent, prompt engineering)
- [Contact](https://wecallshotgun.com/contact): How to reach us, where we deliver, and what to put in a first message (FR: https://wecallshotgun.com/fr/contact)
- [Privacy notice](https://wecallshotgun.com/privacy): What we collect, the legal basis, our processors, retention periods and how to exercise GDPR/UK GDPR rights (FR: https://wecallshotgun.com/fr/confidentialite)
- [Book a call](https://cal.com/wecallshotgun/ai-adoption): Free 20-minute audit

## Blog (English)
- [ChatGPT Images 2.5 vs 2.0 vs Nano Banana Pro (2026 Guide)](https://wecallshotgun.com/blog/chatgpt-images-2-5-business-guide) : what OpenAI's ChatGPT Images 2.5 (released 8 September 2026) changes versus Images 2.0 (21 April 2026): up to 50% lower generation latency, better preservation of the subject in reference photos, more reliable multi-turn editing, more natural lighting and textures, and four new ChatGPT tools (Sketch drawing canvas, Templates for flyers, posters, product photos, logos and icons with follow-up questions, comments placed on the image, shareable prompts), rolling out to every ChatGPT, ChatGPT Work and Codex tier; the two API models GPT-Image-2.5 Flare (fast default, reported two to four times the speed of gpt-image-2, for social content, product experiences, visual search and high-volume generation) and GPT-Image-2.5 Sunburst (longer generation, tighter control across edits, for campaign creative and product imagery) at the same reported token rates as gpt-image-2 ($5/M text in, $8/M image in, $30/M image out, roughly $0.006 to $0.21 per image); a nine-criterion comparison with Google's Nano Banana Pro (Gemini 3 Pro Image, 20 November 2025: 1K/2K/4K output, up to 14 reference images, five consistent people, Google Search grounding, SynthID, $0.139 per image up to 2K and $0.24 at 4K, in Gemini app, Slides, Vids, NotebookLM, Google Ads and Vertex AI; Nano Banana 2 the default since 26 February 2026) where ChatGPT wins on text, layout, iterative edits and speed and Google wins on 4K, photorealism and multi-reference consistency; the decision rule (use the licence you already pay for, route the two or three weak jobs, stop re-evaluating per release); six hands-on business workflows from Images 2.0 client work (campaign variants with real headlines and promo codes, product in the customer's context for proposals, wireframes from a sketch, internal flyers and onboarding visuals from Templates, packshots and multilingual label localisation, infographics from a data table) each with the human check on digits, names and product details; the three prompt rules (quote text exactly, give every element an address, name the junk you don't want); a six-step team rollout (ChatGPT Work or Enterprise seats, likeness and logo consent, shared prompt library, named reviewer, C2PA provenance kept, a cycle-time baseline on ten real requests); and a stated limit: We Call Shotgun's own Images 2.5 test set hasn't run yet, so 2.5 rows are OpenAI's claims (FR: https://wecallshotgun.com/fr/blog/chatgpt-images-2-5-guide-entreprise)
- [Switch from Claude to ChatGPT for GPT-6 Astra: Migration Guide (2026)](https://wecallshotgun.com/blog/claude-to-chatgpt-migration-gpt-6-astra) : the seat-level price comparison (ChatGPT Business and Claude Team both at $25/$20 standard and $125/$100 premium; ChatGPT includes GPT-6 Astra as GPT-6 Pro with 15 messages a month on standard and 50 a week on premium, plus image generation, agent mode, deep research and Codex; Anthropic bills Fable 5.1 as pay-as-you-go usage on standard seats and inside limits on premium; Astra off by default until an admin enables it), the API price list where both flagships cost $10/$50 per million tokens with cache reads at $1 vs $0.25 and a whole-request surcharge above 272,000 input tokens, Artificial Analysis Intelligence Index 57 vs 55 and Coding Index 70 vs 67 in Fable's favour on 6 September 2026, a pros and cons table between OpenAI and Anthropic for a switching decision, and the six-step migration if you switch anyway: inventory, native import of Claude Code, Claude Cowork and Cursor via the ChatGPT desktop app (Settings, Import) or Codex CLI (/import), manual rebuild of claude.ai Projects, memory, custom instructions and connectors, the API swap and blind retest of twenty real prompts, UK GDPR paperwork (DPA, ROPA, DPIA), and a two-to-three-week parallel run with a hard cutover date and the option of keeping a few Claude seats (FR: https://wecallshotgun.com/fr/blog/migrer-de-claude-vers-chatgpt-gpt-6-astra)
- [GPT-6 Astra at Work: A Guide for Non-Technical Teams (2026)](https://wecallshotgun.com/blog/gpt-6-astra-at-work-guide) : what OpenAI's GPT-6 Astra is (its most capable model as of September 2026, positioned above Terra and Luna, successor to GPT-5.6 Sol at the top of the range), what OpenAI's own prompting guidance says changes (more clarifying questions, stronger instruction following including instructions inside files, longer and more formatted output, better continuity on long tasks) and how to translate each into a brief; the BRIEF framework (Business result, Reference material, Independence, Expected output, Final checks) with a reusable template; a meeting follow-up prompt that separates agreed actions from proposals; two copy-paste workflows for marketing (campaign keep/investigate/pause with cost per qualified lead) and finance (budget variance commentary without invented causes) with fictional worked checks; three repair prompts; why self-awarded quality scores are not a check; and the measurement rule (baseline one recurring task, record preparation, generation and checking time, compare cycle time) that decides whether an AI licence earns its place (FR: https://wecallshotgun.com/fr/blog/gpt-6-astra-au-travail-guide)
- [How to Write an AI Policy for Employees: What to Include and a Free Generator](https://wecallshotgun.com/blog/how-to-write-an-ai-policy-for-employees) : the 12 sections of an AI acceptable use policy for employees (approved tools by name and account tier, personal accounts, prohibited data, human review, client disclosure, IP, the new-tool approval route, training, sector obligations, incidents and the ICO 72-hour route, ownership and review), the three clauses that matter most (account tier, approval route, human review), the free We Call Shotgun AI policy generator (15 questions, editable .docx, UK GDPR and ICO framing, EU AI Act Article 4 clause only with an EU nexus, answers never leave the browser), the evidence that a policy alone changes behaviour very little (Microsoft Work Trend Index 2024: 75% of knowledge workers use generative AI, 78% bring their own tools, 52% hide use for important tasks, 39% trained by their company; Cyberhaven Q2 2024: 73.8% of workplace ChatGPT use via non-corporate accounts, 27.4% of data sensitive; KPMG/University of Melbourne 2025: 48% contravene AI policy, 57% conceal AI use, 47% trained; ISACA 2026: 38% of organisations with a formal AI policy, up from 28%; CBI Adoption Decade 2026: 49% vs 15% ROI execution divide), and what the do's look like: a leadership narrative, a green list of tasks per team, the tool for each task, a review rule per output, a baseline and cycle-time measure over 90 days, and everyone shown how on their own work (FR: https://wecallshotgun.com/fr/blog/rediger-charte-ia-entreprise)
- [Vibe Coding vs Agentic Workflows: The 2026 Guide for UK Business Leaders](https://wecallshotgun.com/blog/vibe-coding-vs-agentic-workflows) : the research-backed distinction — vibe coding is a discovery method (Karpathy's original definition, no code review, fine when failure is cheap), agentic workflows are a delivery method (AI agents implement inside specs, tests, reviews, security gates, with a named accountable human); the dividing line is verification, not the tool; Anthropic's ~400k-session finding (humans ~70% of planning, AI ~80% of execution), the AppDirect, Leatherman, Lucid and Thoughtworks cases, the 2026 security evidence (WIRED ~2,000 exposed vibe-coded apps, Veracode 45% insecure code samples, OWASP Secure Coding with AI), the speed evidence (Asana, Virgin Atlantic, DORA amplifier, METR 19% slowdown), a 13-question engineering-gate framework, the vibe-first-engineer-second operating model, and role-by-role actions for CEO, COO, CMO, CFO and L&D (FR: https://wecallshotgun.com/fr/blog/vibe-coding-vs-workflows-agentiques)
- [Enterprise AI Agents in 2026: The Vendor-Neutral Benchmark for Companies](https://wecallshotgun.com/blog/enterprise-ai-agents-benchmark-2026) : the enterprise AI agent landscape as of August 2026 — four buying routes (frontier labs OpenAI/Anthropic, hyperscalers Microsoft/Google, app platforms Salesforce Agentforce/ServiceNow, specialists Glean/UiPath/CrewAI), a comparison and pricing table, how to read agent benchmarks (GDPval, tau2-bench, OSWorld, Terminal-Bench, SWE-bench) and why your own two-week workflow eval beats them, the four risks (prompt injection, over-permissioned access, silent errors, shadow agents), five hard limits, and the adoption evidence (Gartner 40%+ cancellations by 2027, MIT NANDA 95% pilots without P&L impact, CBI 49% vs 15% ROI divide) (FR: https://wecallshotgun.com/fr/blog/agents-ia-entreprise-comparatif-2026)
- [Zero Data Retention AI Models: OpenAI vs Claude vs Gemini vs Copilot](https://wecallshotgun.com/blog/zero-data-retention-ai-models-comparison) : what zero data retention (ZDR) means and how it differs from a no-training policy, data residency and confidential computing; the three mechanisms vendors use (Anthropic sells a contract, Microsoft takes an application per Azure subscription, Google gives you a configuration checklist that fails open because context caching runs by default on a 24-hour window); OpenAI's 19 Aug 2026 frontier-model ZDR announcement and Private Safety Processing preview; Anthropic's Covered Models policy of 9 June 2026 that forces 30-day retention and blocks ZDR entirely on its most capable models across Claude Enterprise, Bedrock, Google Cloud and Microsoft Foundry; per-vendor default retention windows (OpenAI 30 days, Vertex 30 or 90 days, Gemini paid API 55 days, Azure OpenAI 30 days); what breaks under ZDR (batch, file storage, code execution, MCP connectors) and the audit-trail conflict; the NYT preservation order that exempted ZDR customers; and 10 procurement questions to put in an RFP (FR: https://wecallshotgun.com/fr/blog/retention-zero-donnees-modeles-ia-comparatif)
- [AI Sycophancy at Work: The Hidden Cost of an AI Yes-Man](https://wecallshotgun.com/blog/ai-sycophancy-business-teams) : what AI sycophancy is and why it happens (RLHF), the April 2025 GPT-4o rollback, the Stanford findings (models affirm users 50% more than humans, tell both sides of a conflict they are right in 48% of cases), the organisational psychology behind it (MUM effect, organizational silence, CEO flattery research), concrete costs for marketing, sales, HR, finance and ops teams, and the five prompt moves that get honest answers (FR: https://wecallshotgun.com/fr/blog/sycophancie-ia-entreprise)
- [ChatGPT Event-Triggered Tasks: Automation Without an Automation Tool](https://wecallshotgun.com/blog/chatgpt-event-triggered-tasks-business-automation) : guide to ChatGPT's event-triggered (webhook) scheduled tasks released 25 Aug 2026 — Gmail, Slack and GitHub triggers, plan availability (Plus/Pro/Business/Enterprise/Edu, not Free/Go/FedRAMP), the 30-per-hour and 720-per-day run limits, active task caps by plan, shared tasks, use cases for marketing, sales, support, ops and C-level, why it does not replace Zapier or Make, and the prompt-injection risk of a standing Gmail listener
- [CBI Adoption Decade Report: The AI Execution Divide Is a People Problem](https://wecallshotgun.com/blog/cbi-adoption-decade-ai-execution-divide) : analysis of the CBI / Oliver Wyman report "The Adoption Decade" (18 Aug 2026) — the 49% vs 15% ROI execution divide between AI deployment leaders and laggards, the CBI's four goals, why the gap is behavioural rather than technological, and the 90-day sequence to close it
- [Executive's Guide to Leading AI Transformation](https://wecallshotgun.com/blog/executive-guide-ai-transformation)
- [CMO's Playbook for AI-Driven Marketing Operations](https://wecallshotgun.com/blog/cmo-playbook-ai-marketing-operations)
- [AI for Finance Teams: Workflows That Save 10+ Hours/Week](https://wecallshotgun.com/blog/ai-workflows-finance-teams)
- [AI for Marketing Teams: 20 Workflows](https://wecallshotgun.com/blog/ai-marketing-workflows-save-10-hours-week)
- [ChatGPT Enterprise vs Copilot vs Gemini](https://wecallshotgun.com/blog/chatgpt-enterprise-vs-copilot-vs-gemini)
- [Claude AI for Teams: Getting Started Guide](https://wecallshotgun.com/blog/claude-ai-getting-started-guide-teams-2026)
- [Protect Claude Usage Limits: 18 Tactics](https://wecallshotgun.com/blog/protect-claude-usage-limits-stop-burning-credits-work)
- [How to Choose an AI Training Provider in the UK](https://wecallshotgun.com/blog/choose-ai-training-provider-uk)
- [Best AI Training Providers in London (2026): An Honest Shortlist](https://wecallshotgun.com/blog/best-ai-training-providers-london-2026)
- [UK SME AI Adoption Statistics 2026: Every Number Worth Citing](https://wecallshotgun.com/blog/uk-sme-ai-adoption-statistics-2026)
- [How Much Does AI Training Cost in the UK? The 2026 Price Guide](https://wecallshotgun.com/blog/ai-training-cost-uk-2026)
- [Claude Opus 5 for Business: Costs, Capabilities and UK Use Cases (2026)](https://wecallshotgun.com/blog/claude-opus-5-business-guide)
- [The Risks of Shared AI Conversations: Guardrails Every Company Needs in 2026](https://wecallshotgun.com/blog/shared-ai-conversations-data-leak-risks)
- [The Digital Omnibus Is Law: What Regulation (EU) 2026/1744 Changes in the EU AI Act](https://wecallshotgun.com/blog/digital-omnibus-ai-act-2026-what-changed)
- [Microsoft Copilot in Banking: 5 Business Use Cases UK Banks Are Actually Deploying in 2026](https://wecallshotgun.com/blog/microsoft-copilot-banking-use-cases-uk)
- [Best AI Consulting Firms in London & the UK (2026): An Honest Comparison](https://wecallshotgun.com/blog/best-ai-consulting-firms-london-uk-2026)
- [Best AI Consulting Firms for SMBs & Mid-Market Companies in France (2026)](https://wecallshotgun.com/blog/meilleurs-cabinets-conseil-ia-pme-france)
- [From Prompting to Task Delegation: The 2026 AI Adoption Shift for UK Businesses](https://wecallshotgun.com/blog/from-prompting-to-task-delegation-ai-uk-2026) : why the unit of AI work moved from the prompt to the task with ChatGPT Work, Claude Cowork, Gemini Spark and Microsoft Scout; the six-field Delegation Brief that replaces prompt engineering, mapped to Anthropic's 4D AI fluency framework; ten ChatGPT Work delegations UK SMBs and mid-market firms are running
- [How to Import Your Claude Agents into ChatGPT: The 2026 Migration Guide for UK Teams](https://wecallshotgun.com/blog/import-claude-agents-into-chatgpt-uk-teams) : the exact click path for ChatGPT's native import of Claude Code, Claude Cowork and Cursor setups (Settings > Import, or General > Import other agent setup) plus the Codex CLI /import command; a table of what transfers automatically and what has to be rebuilt by hand (claude.ai Projects, custom instructions, memory, connectors, MCP servers); Claude Team vs ChatGPT Business vs Claude Enterprise vs ChatGPT Enterprise 2026 pricing; a five-day migration plan and the UK GDPR work (DPA, ROPA, DPIA) a change of processor triggers
- [AI Adoption Training: Why Most AI Projects Fail After Go-Live (UK Guide 2026)](https://wecallshotgun.com/blog/ai-adoption-training-uk) : what AI adoption training is (workflow-first enablement with usage measured at 30/60/90 days), the handover cliff that kills projects after go-live, the six components of a real programme, the day-90 test to run on any vendor, and UK price bands
- [How Much Does AI Consulting Cost in the UK? Day Rates and Project Fees (2026)](https://wecallshotgun.com/blog/ai-consulting-costs-uk-2026) : UK AI consultant day rates by tier (independents £500–£2,000, mid-tier £900–£1,600, large firms £2,000–£5,000 with £100k+ minimums), project fee benchmarks from £3,500 audits to £250k programmes, We Call Shotgun's published prices, and a five-step method to compare quotes like-for-like
- [How to Choose an AI Consultancy: 10 Questions Vendors Dodge](https://wecallshotgun.com/blog/how-to-choose-ai-consultancy-uk) : the ten questions that expose delivery reality (named team, day-90 outcomes, measured references, tool conflicts, IP, governance, wrong-fit work), the 2026 consolidation checks after Accenture–Faculty, CGI–BJSS and Indicium AI, and a four-hour shortlist method
- [AI Advisory vs Implementation vs Training: Which AI Partner Do You Actually Need?](https://wecallshotgun.com/blog/ai-advisory-vs-implementation-vs-training) : the three AI purchases defined (advisory decides, implementation builds, training changes behaviour), a symptom-to-purchase table, why the 2026 sequence inverts to advisory-then-training, and what each lane costs in the UK
- [AI Readiness Audit: What It Should Include and What It Costs (UK 2026)](https://wecallshotgun.com/blog/ai-readiness-audit-uk) : the five components of a real audit (workflow value map, data and stack, skills baseline, governance posture, sequenced quick wins), UK price bands from free scorecards to £50k+ Big 4 assessments, and the red flags that expose an audit-shaped sales document
- [Do You Need an AI Consultant or an In-House Hire? The UK Decision Guide (2026)](https://wecallshotgun.com/blog/ai-consultant-vs-in-house-hire) : real cost comparison (London AI lead £90k–£150k base and 3–6 months to hire vs £3,500–£45,000 consultancy engagements), when each wins, why first AI hires fail structurally, and the audit-champions-then-hire hybrid sequence
- [AI Consultancy for Financial Services in the UK: What to Look For (2026)](https://wecallshotgun.com/blog/ai-consultancy-financial-services-uk) : the sector filter for banks, insurers and fintechs — PRA SS1/23 model risk, Consumer Duty on AI-drafted outputs, SM&CR accountability, ICO guidance — the three artefacts to demand from any candidate firm, and who to shortlist by problem type

## Blog (Français)
- [Passer de Claude à ChatGPT pour GPT-6 Astra : guide de migration 2026](https://wecallshotgun.com/fr/blog/migrer-de-claude-vers-chatgpt-gpt-6-astra) : la comparaison des prix par siège (ChatGPT Business et Claude Team à 25 $/20 $ en standard et 125 $/100 $ en premium ; ChatGPT inclut GPT-6 Astra sous le nom GPT-6 Pro avec 15 messages par mois en standard et 50 par semaine en premium, plus génération d'images, mode agent, recherche approfondie et Codex ; Anthropic facture Fable 5.1 à l'usage sur les sièges standard et l'inclut dans les limites en premium ; Astra désactivé par défaut jusqu'à activation par un administrateur), la grille API où les deux modèles phares coûtent 10 $/50 $ par million de tokens avec lectures en cache à 1 $ contre 0,25 $ et majoration au-delà de 272 000 tokens en entrée, Intelligence Index d'Artificial Analysis 57 contre 55 et indice de code 70 contre 67 en faveur de Fable au 6 septembre 2026, un tableau avantages et inconvénients entre OpenAI et Anthropic pour décider de changer, et la migration en six étapes si vous changez quand même : inventaire, import natif de Claude Code, Claude Cowork et Cursor via l'application ChatGPT de bureau (Réglages, Importer) ou Codex CLI (/import), reconstruction manuelle des Projets claude.ai, de la mémoire, des instructions personnalisées et des connecteurs, changement d'API et retest à l'aveugle de vingt vrais prompts, documents RGPD, et période en parallèle de deux à trois semaines avec date de bascule ferme et l'option de garder quelques sièges Claude (EN: https://wecallshotgun.com/blog/claude-to-chatgpt-migration-gpt-6-astra)
- [GPT-6 Astra au travail : le guide 2026 pour les équipes non techniques](https://wecallshotgun.com/fr/blog/gpt-6-astra-au-travail-guide) : ce qu'est GPT-6 Astra d'OpenAI (modèle le plus capable en septembre 2026, au-dessus de Terra et Luna, successeur de GPT-5.6 Sol), ce que la documentation de prompting d'OpenAI dit qui change (plus de questions de clarification, suivi des instructions plus fort y compris dans les fichiers, réponses plus longues et formatées, meilleure continuité) et comment le traduire en brief ; le cadre BRIEF (résultat métier, sources, autonomie, livrable attendu, contrôles finaux) avec gabarit réutilisable ; un prompt de compte rendu de réunion qui sépare décisions et propositions ; deux workflows à copier pour le marketing (coût par lead qualifié) et la finance (commentaire des écarts sans causes inventées) avec contrôles fictifs ; trois prompts de correction ; et la règle de mesure (baseline sur une tâche récurrente, temps de préparation, génération et vérification, comparaison du temps de cycle) (EN: https://wecallshotgun.com/blog/gpt-6-astra-at-work-guide)
- [Tâches déclenchées par événement dans ChatGPT : automatiser sans outil d'automatisation](https://wecallshotgun.com/fr/blog/taches-declenchees-par-evenement-chatgpt) : guide des tâches planifiées déclenchées par webhook publiées le 25 août 2026 — déclencheurs Gmail, Slack et GitHub, éligibilité (Plus/Pro/Business/Enterprise/Edu, ni Free, ni Go, ni FedRAMP), limites de 30 exécutions par heure et 720 par jour, plafonds de tâches actives par offre, tâches partagées, cas d'usage marketing, ventes, support, opérations et direction, pourquoi cela ne remplace pas Zapier ou Make, et le risque d'injection de prompt lié à une écoute Gmail permanente
- [Rapport « Adoption Decade » de la CBI : l'écart d'exécution IA est un problème humain](https://wecallshotgun.com/fr/blog/cbi-adoption-decade-ai-execution-divide) : analyse du rapport CBI / Oliver Wyman « The Adoption Decade » (18 août 2026) — l'écart d'exécution 49% contre 15% sur le ROI, les quatre objectifs de la CBI, pourquoi l'écart est comportemental et non technologique, et la séquence de 90 jours pour le combler
- [Formation à l'adoption de l'IA : pourquoi les projets IA échouent après le déploiement](https://wecallshotgun.com/fr/blog/formation-adoption-ia-entreprise) : ce qu'est la formation à l'adoption (accompagnement par workflow avec usage mesuré à 30/60/90 jours), la falaise de passation, les six composants d'un vrai programme, le test des 90 jours et les prix
- [Combien coûte le conseil en IA au Royaume-Uni ? Taux journaliers et forfaits (2026)](https://wecallshotgun.com/fr/blog/cout-conseil-ia-royaume-uni-2026) : taux journaliers par niveau (indépendants 500–2 000 £, intermédiaires 900–1 600 £, grands cabinets 2 000–5 000 £), repères au forfait de l'audit à 3 500 £ au programme à 250 000 £, et la méthode pour comparer les devis
- [Comment choisir un cabinet de conseil en IA : 10 questions que les prestataires esquivent](https://wecallshotgun.com/fr/blog/choisir-cabinet-conseil-ia-10-questions) : les dix questions qui exposent la réalité de la livraison, les vérifications de consolidation 2026 (Accenture–Faculty, CGI–BJSS, Indicium AI), et une méthode de shortlist en quatre heures
- [Conseil, intégration ou formation IA : de quel partenaire avez-vous vraiment besoin ?](https://wecallshotgun.com/fr/blog/conseil-ia-vs-integration-vs-formation) : les trois achats IA définis, la table des symptômes, la séquence qui s'inverse en 2026, et le coût de chaque voie
- [Audit de maturité IA : ce qu'il doit contenir et ce qu'il coûte (2026)](https://wecallshotgun.com/fr/blog/audit-maturite-ia-entreprise) : les cinq composants d'un vrai audit, les bandes de prix, et les signaux d'alarme qui révèlent un document commercial déguisé
- [Consultant IA ou recrutement interne ? Le guide de décision (2026)](https://wecallshotgun.com/fr/blog/consultant-ia-ou-recrutement-interne) : la comparaison des coûts réels, quand chaque option gagne, pourquoi la première embauche IA échoue, et la séquence hybride audit-champions-recrutement
- [Cabinet de conseil en IA pour les services financiers au Royaume-Uni : les critères (2026)](https://wecallshotgun.com/fr/blog/conseil-ia-services-financiers-uk) : le filtre sectoriel banque-assurance (SS1/23, Consumer Duty, SM&CR, ICO), les trois artefacts à exiger, et qui shortlister par problème
- [Comment importer vos agents Claude dans ChatGPT : le guide de migration 2026](https://wecallshotgun.com/fr/blog/importer-agents-claude-vers-chatgpt) : le chemin exact de l'import natif ChatGPT des configurations Claude Code, Claude Cowork et Cursor (Réglages > Import) et la commande /import de Codex CLI ; ce qui se transfère automatiquement et ce qu'il faut reconstruire (Projets claude.ai, instructions, mémoire, connecteurs, serveurs MCP) ; tarifs Claude Team, ChatGPT Business, Claude Enterprise et ChatGPT Enterprise 2026 ; plan de migration en cinq jours et obligations RGPD (contrat de sous-traitance, registre, AIPD)
- [Du prompt à la délégation de tâches : le virage IA 2026 des PME et ETI françaises](https://wecallshotgun.com/fr/blog/du-prompt-a-la-delegation-de-taches-ia) : pourquoi l'unité de travail de l'IA est passée du prompt à la tâche avec ChatGPT Work, Claude Cowork, Gemini Spark et Microsoft Scout ; le brief de délégation à six champs qui remplace le prompt engineering, aligné sur le cadre 4D d'Anthropic ; dix délégations que des PME et ETI font tourner
- [Quels sont les meilleurs cabinets de conseil en IA pour les PME et ETI en France ? (2026)](https://wecallshotgun.com/fr/blog/meilleurs-cabinets-conseil-ia-pme-france) : comparatif des quatre familles de cabinets IA en France (grands cabinets, implémentation technique, boutiques de niche, formation au catalogue) et de leur adéquation aux entreprises de 50 à 1 000 salariés, avec les prix
- [Les meilleurs cabinets de conseil en IA en France (2026) : comparatif honnête](https://wecallshotgun.com/fr/blog/meilleurs-cabinets-conseil-ia-france-2026) : Artefact, Ekimetrics, Capgemini Invent, Sia Partners, Onepoint, Accenture, QuantumBlack, DataScientest, au-delà de Paris
- [L'AI Act expliqué aux dirigeants de PME, ETI et grands groupes](https://wecallshotgun.com/fr/blog/ai-act-guide-pme-eti-france)
- [Comment mesurer le ROI d'un programme de formation IA](https://wecallshotgun.com/fr/blog/mesurer-roi-formation-ia-entreprise)
- [Programme de formation IA pour 50, 200 ou 1 000 collaborateurs](https://wecallshotgun.com/fr/blog/programme-formation-ia-grande-echelle)

## Founders
- **Toni Dos Santos** — Co-founder. Knowledge & Enablement Architect. Background: BPCE (France's 2nd largest bank), lemlist (product management), founder of Buska (social listening SaaS, Product Hunt Top 3). AI trainer for L'Oréal, Essilor, IGN, ZELIQ. [LinkedIn](https://www.linkedin.com/in/dossantostoni/)
- **Meera Sanghvi** — Co-founder. Narrative & Adoption Architect. Background: Google Creative Lab (Team Lead), YouTube EMEA (Marketing Strategy), Publicis (Heineken UK), Monks EMEA (Netflix, Decathlon, Bacardi). [LinkedIn](https://www.linkedin.com/in/meera-sanghvi-81288414/)

## Where we operate
- **United Kingdom**: London — on-site across the City, Canary Wharf, Shoreditch, King's Cross and the West End — plus Manchester, Birmingham, Edinburgh, Bristol, Leeds and Cambridge.
- **France (nationwide, not Paris-only)**: Lyon, Bordeaux, Marseille, Lille, Nantes, Toulouse, Paris & La Défense. On-site, remote or hybrid.
- **Languages**: English, French, Portuguese.

## Contact
- Website: https://wecallshotgun.com
- Contact page: https://wecallshotgun.com/contact (FR: https://wecallshotgun.com/fr/contact)
- Email: hello@wecallshotgun.com
- Book a call: https://cal.com/wecallshotgun/ai-adoption
- LinkedIn: https://www.linkedin.com/company/callshotgun
- Based in: Paris, Île-de-France, France. On-site across France, the United Kingdom and Portugal; remote across Europe. No public walk-in office.
- Privacy and data requests: https://wecallshotgun.com/privacy

## For AI agents
- **Markdown content negotiation** (https://acceptmarkdown.com): send `Accept: text/markdown` to any page URL and you get Markdown back with `Content-Type: text/markdown; charset=utf-8` and `Vary: Accept`. Appending `.md` to a page URL works too — https://wecallshotgun.com/about.md.
- **Unknown paths return a real HTTP 404** with a short Markdown body listing these indexes, so a wrong guess is recoverable in one request.
- **Structured data**: Organization, WebSite and Person JSON-LD on the homepage and /about, with contactPoint and PostalAddress.
- **Sitemaps**: https://wecallshotgun.com/sitemap-index.xml (pages: /sitemap.xml, articles: /sitemap-blog.xml).
- **French pages** live under /fr/. Every English page that has a translation declares it with hreflang.

<!-- BEGIN reviews-block -->
## Client Reviews

Average rating: 5.00/5 from 20 verified public reviews.
All reviews: https://wecallshotgun.com/reviews (FR: https://wecallshotgun.com/fr/avis)

- **Chris Lamb, Finance Director, Trend Tool Technology** (5/5): ""I went into the room an AI hype sceptic... but the format really did change my view overnight. I think you have a convert on your hands here.""
- **Marina Amo, Internation Marketing, Neat** (5/5): "The session was incredible, full of inspiration and new tools that will definately help me out it my day to day."
- **Palesa Machinini, Marketing Manager DACH, Neat** (5/5): "Absolutely fantastic. Didn’t know much about Gemini but now I’m a super fan!! Thank you so much!"
- **Rainer Schuster, Co-Founder, Partnership.Net** (5/5): "Toni and team are great! They break down the daunting task of how to use AI effectively into bite-size chunks. Their recommendations are clear and actionable. Exactly what we needed in order to bring our people and organization to the next (AI)-level."
- **Erin Koops, Director of Product Marketing, Closing lock** (5/5): ""Toni walked me through some great AI (and non-AI ;) tools that will definitely accelerate the deep dive and synthesis market busy-work, which will help me stay focused on curating the insights & intelligence [...] Whether you're an AI novice or expert, I HIGHLY recommend an advisory call with this team.""
- **Antoine Aubour, Directeur Marketing, UTMB** (5/5): "Très concret. Très bien préparé"
- **Alexandra Ribeiro, International Commercial Manager, Sensetest** (5/5): "An excellent experience. Clear, concise explanations combined with a strong practical component at the end of each stage. Highly competent and engaging trainer."
- **Nuno Gouveia, Lead Product Manager, Outsystems** (5/5): "Toni’s workshops were really world class"
- **Hugo Valente, Senior Product Manager, Airalo** (5/5): "This was an amazing and enriching experience.""
- **Charlotte Gitenay, Senior User Researcher, Freelance** (5/5): "In just 20 minutes, Toni gave me actionable advices related to the problem I'd mentioned. As a beginner in AI, he was able to guide me towards resources adapted to my level."
- **Max Bullett, Head of Product, DAA GmbH** (5/5): "Had a wonderfully intense and meaningfully didactic experience, even as a fairly senior product professional. Highly recommend..."
- **Meryem Baig, Product Manager, DAA GmbH** (5/5): "You made me believe in myself. In just 48 hours, I built an MVP, pitched it, and received real feedback from a jury."
- **Luciano Borges, Product Visionary, CTW @ BMW Group** (5/5): ""intense learning, high-level mentoring, and real delivery: The Product buildcamp is the place to move from theory to practical product development.""
- **Justine Lambert, Responsable La Station numixs, Station numixs** (5/5): "Smart and useful. Inspiring and ready to use. Thanks Toni for being a great IA teacher. You made the topic obvious and accessible in less than an hour."
- **Paola D'Angela, CEO, Giro Music** (5/5): "Toni connaissant bien l'écosystème des musiques actuelles m'a permis lors d'un accompagnement personnalisé, grâce à sa bienveillance à prendre en main très vite des outils que je peux aujourd'hui utiliser facilement. Les séquences sont bien pensées. Expérience efficace."
- **Benjamin Galvan, Online Business Manager, OBM** (5/5): "Merci pour cet atelier avec Toni très clair et bien structuré. Même en tant qu'Online Business Manager et consultant IA, j'ai pu découvrir de nouveaux outils et surtout apprécier une approche pédagogique très accessible."
- **Tsvika, Tech PM Lead, Miniclip** (5/5): "Immensely empowering and valuable. I learned more in 48 hours than I did in the past 6 months."
- **Antoine Damiens, Senior B2B Product Manager, Freelancer** (5/5): "As a freelancer, I always optimize my workflow to be more productive. It's a matter of profit. Toni quickly helped me identify workflows and tasks that could be automated or supported with No-Code and AI."
- **Ricardo Vitorino, Product Manager, Untile** (5/5): "I gained a lot of insights about AI tools and how to ship products effectively."
- **Jean Cattan, Secrétaire Général, Conseil de l'IA et du numérique** (5/5): "«J'ai adoré», «J'ai appris plein de choses», «Super bien mené». Toni Dos Santos a animé un atelier IA et il a fait l'unanimité chez un public aux positions et expériences très variées, oscillant entre clefs de compréhension générales et mille aspects pratiques concrets."
<!-- END reviews-block -->