Free resource · Word + Excel kit
AI Innovator's Kit: 24 use cases and their prompts, in three levels
From weak signal to funding application: ready-to-use prompts to help your innovation, R&D and transformation teams develop their skills independently, step by step.
The innovator's AI kit, in short
- What it is
- A self-training kit for innovation, R&D and transformation teams: 24 everyday use cases with prompts, classified into three levels to progress without prior training, and an Excel workbook of tools.
- For who
- Innovation departments, labs, R&D, product and transformation teams of large organizations, and intrapreneurs who are leading a project.
- Format
- A 43-page Word document: a user guide, 24 one-page worksheets, a validation challenge for each level, a section for the innovation team and the lab, and six appendices. A nine-tab Excel workbook.
- What it contains
- Self-positioning in twelve questions, prompt framework in four ingredients, five safety rules specific to innovation, 24 prompts ready to copy from monitoring to funding, selection grid, hypothesis map, test cards, idea portfolio, verification checklist.
- What is it based on?
- BCG studies on AI in innovation and R&D, work from the Wharton School, Harvard and Science Advances on AI and creativity, Anthropic's AI Fluency framework, and our workshops with innovation teams from large organizations.
- Price
- Free, in exchange for a professional email address. No unsolicited emails.
- Update
- October 2026.
- What it doesn't replace
- Your organization's AI charter, the opinion of your intellectual property department, or a workshop on your own topics with the team.
- Go Further
- AI training for innovation teams, AI toolkit for project managers, AI Measurement and ROI Kit.
Most innovation teams in large organizations have a generative AI tool at their fingertips: Copilot in Teams, a ChatGPT or Claude enterprise account, Gemini in Workspace, or Vibe from Mistral. What's missing isn't the tool itself, but a structured approach. Everyone tries it on their own, ends up with a list of banal ideas or a "market research" study with unverifiable sources, and concludes that AI isn't useful for innovation. Meanwhile, a minority integrates it into their routines: automated monitoring, large-scale ideation, and concepts tested in a week instead of three months.
This page explains what research shows about AI in innovation, why a tiered progression is better than a list of prompts, what the 24 use cases in the kit contain, and gives you the complete kit, in Word format, with its Excel workbook, to get started today on your real-world topics.
1. AI in innovation: what studies show
BCG has been measuring this gap for several years. In its 2024 study of the most innovative companies, 86% of innovation teams were experimenting with generative AI, but organizations already "ready" to innovate had deployed it at scale five times more often than others; yet only 3% of companies reached this readiness zone, while 83% placed innovation among their top three priorities. In R&D, BCG anticipates a 10% to 20% reduction in time to market and up to 20% cost savings by 2025, provided that working methods are revised rather than simply adding a new tool.
Academic research clarifies what AI does well and poorly for an innovator. At the Wharton School, Girotra, Meincke, Terwiesch, and Ulrich compared product ideas from GPT-4 to those from MBA students in 2023: the model's ideas were, on average, of higher quality, produced much faster and at almost no cost, and the vast majority of the best ideas in the sample came from the model. Their conclusion: when ideas cost nothing, selection becomes the profession.
But two other studies highlight the limitations. The Harvard experiment with 758 BCG consultants demonstrated the "jagged frontier": 40% higher quality on creative tasks, but significantly worse results than without AI on an analytical task where the model confidently made mistakes. Furthermore, Doshi and Hauser published in Science Advances in 2024 that AI assistance improves individual output, especially among less creative individuals, but also makes it more similar: collective diversity decreases. This toolkit is designed to leverage the first set of findings and mitigate the second.
2. Why a progression by level, and not just a list of more prompts?
Libraries of prompts for innovation abound: consulting firms, tool publishers, and design thinking communities all publish them. They share three common flaws. They are simplistic: fifty prompts presented out of order, leaving beginners unsure where to start. They stop at the AI release: nothing on what to prepare beforehand, nothing on what to verify afterward, even though this is precisely where invented sources and plausible market figures emerge. And they ignore the context of a large organization: unregistered concepts, R&D data, confidentiality agreements with partners, selection committees, and funding applications.
The toolkit takes a different approach to the problem. It follows the three ways of working with AI described in the AI Fluency framework published by Anthropic in 2025: automation, where the AI performs a specific task on command; augmentation, where humans and AI build together in multiple rounds; and agency, where humans configure the AI to then perform tasks on their behalf. These three modes result in three levels, and the four skills of the same framework (delegate, describe, discern, and be diligent) provide the self-assessment grid. Each level covers the eight situations encountered by innovators, from exploration to funding, so the toolkit can also be used on demand: you find your situation, then the level where you feel most comfortable.
Each fact sheet finally explains what the lists don't: when to use the case, what preparation is needed, what should be achieved, the three points to check before using the result, the variations depending on the tool, and the next fact sheet. The kit also establishes safeguards specific to innovation: noting down ideas before consulting the AI, requiring diverse perspectives, and retaining a degree of human input in each selection.
3. The three levels and the six-week course
| Level | Working method | What you know how to do at the end | Indicative duration |
|---|---|---|---|
| Level 1 · Start | The AI performs one task at a time, based on what you give it. | Get in minutes a sourced weak signal note, grouped verbatim quotes, a framed challenge, fifty ideas, a value proposition, a selection grid, a hypothesis map and a pitch, which you use after verification. | 2 weeks, 20 to 30 minutes per day |
| Level 2 · Structuring | Co-constructing AI: you provide your documents, you engage in dialogue over several rounds, you make it work in stages. | Produce committee deliverables: updatable state of the art, analyzed interviews, innovation brief, ideation by analogies, scenario and concept sheet, selection committee, test protocol, business case and call for projects file. | 2 weeks, 30 to 45 minutes per day |
| Level 3 · Systematize | You configure the AI to work for you and the team: standing instructions, projects and agents, scheduled tasks, prompt chains. | Having an innovation assistant, a weekly monitoring service, an insights database, an ideation chain, a prototype in one day, a portfolio evaluator, an experimentation journal and a funding co-pilot. | 2 weeks, 45 minutes a day, then a few minutes a week |
The proposed program takes six weeks: four worksheets per week, presented in the order of the situations, and a validation challenge at the end of each level, based on a real-world topic from your portfolio. At level 1, you will produce a report on weak signals, verifying all sources, and a value proposition for a concept. At level 2, you will present a complete concept, including its testing protocol, to a committee. At level 3, you will activate your innovation assistant and a tool (monitoring system, insights database, evaluator, or journal) that a colleague uses independently. The twelve-question self-assessment indicates your entry level; the progress log and self-assessment grid in the appendices measure your progress.
4. The 24 use cases, from monitoring to financing
Eight situations, three levels, one sheet per box. The situations are based on those we work on in the AI training for innovation teams, supplemented by user insights and experimentation.
The details of each case follow: the situation, what you get, and the point to check before using the result. The complete prompt is in the kit.
Level 1 · Start: AI executes
1.1 Map ten weak signals on a theme, with their sources
Exploration and monitoring. Time saved: 1 hour per monitoring note. At the start of an exploration, when you are asked "what's happening" on a theme (decarbonized mobility, AI in customer relations, new materials) and you have neither the time to read forty articles nor the desire to arrive at the committee with three intuitions.
What you get: ten dated and sourced signals, each linked to a question for the organization; three underestimated signals with supporting arguments and a list of monitoring keywords. To check : Open each source: title, date, and content match. The models produce plausible and false references.
1.2 Group thirty verbatim comments into needs, irritants, and hypotheses.
User insights. Time saved: 2 hours per batch of verbatim comments. After a series of interviews, an internal survey, and a review of customer feedback or support tickets.
What you get: eight groups of needs and irritants named, counted and illustrated with real quotations; five hypotheses of unmet needs, formulated to be tested. To check : Each quote exists word for word in your verbatim transcripts: do a text search on two or three of them.
1.3 Reframe a vague problem into an innovation challenge and "How could we" questions
Framing the challenge. Time saved: 45 minutes per framing. When a leader or a profession entrusts you with an intention ("we need to do something about loyalty", "find us a use for AI") without a defined problem, and you have to come back with a challenge that is precise enough to be worked on and open enough to leave room for ideas.
What you get: three reformulations of the problem and three "How could we" questions of a good length; the list of assumptions hidden in the request, and those to be checked first. To check : no question "How could we" already contains the solution ("...thanks to a mobile application").
1.4 Generate fifty ideas in ten minutes with SCAMPER
Ideation. Time saved: 1 hour per ideation session. When the challenge is framed and it is necessary to open up the field before the workshop with the team.
What you get: fifty short ideas, spread across the seven SCAMPER levers, with no duplication; a first selection: five promising ideas, five original ideas. To check : compare with your ideas noted before the prompt: if the AI has not taught you anything, try again with "suggest twenty ideas that an actor from another sector would have".
1.5 Write the value proposition of an idea in one page
Concept design. Time saved: 1 hour per concept. When an idea emerges from the workshop and needs to be made presentable and debatable.
What you get: a structured page: title, value statement, benefits, adoption effort, risks, objections; the weaknesses of the concept made visible before the committee. To check : The problem solved is proven by data or verbatim, not assumed: otherwise, return to sheet 1.2.
1.6 Construct an initial selection grid: desirability, feasibility, viability
Evaluation and prioritization. Time saved: 45 minutes per grid. When several ideas are on the table and the team still chooses "by feeling" or to the benefit of the one who speaks the loudest.
What you get: a grid of up to nine criteria, each with a scale described and expected proof; a summary rule and its limits. To check : The criteria are aligned with your actual strategy, not a generic one: review them with a member of the committee.
1.7 List the critical assumptions of a concept and rank them by risk
Experimentation and testing. Time saved: 45 minutes per concept. As soon as a concept exists (sheet 1.5) and before investing.
What you get: the complete map of the concept's assumptions, in four families; the three riskiest assumptions and the lightest test for each. To check : The "level of evidence" assigned by the AI corresponds to what you actually have: correct it hypothesis by hypothesis.
1.8 Prepare a three-minute pitch for an idea
Deployment, pitch and funding. Time saved: 1 hour per pitch. Before an innovation committee meeting, an internal ideas competition, a meeting with a sponsor.
What you get: a three-minute pitch that states the problem, the evidence, the unknown, and the demand; five anticipated difficult questions, with their answers. To check : Every number in the pitch comes from you and you know where it comes from.
Level 2 · Structuring: Co-constructed AI
2.1 Construct a state-of-the-art or sourced benchmark that can be updated
Exploration and monitoring. Time saved: 1 day per state of the art. Before committing resources to a particular track.
What you get: a state-of-the-art plan validated by you, then a completed and sourced comparative grid; a synthesis that distinguishes the established, the contested, the failures and the free spaces. To check : Check at least one out of three sources, and all those that form the basis of an opportunity: date, author, actual content.
2.2 Prepare user interviews, then analyze the transcripts for opportunities
User insights. Time saved: 3 hours per campaign of five interviews. When the assumptions of sheet 1.2 need to be tested against real users.
What you get: a non-suggestive, experience-oriented interview guide with prompts; a hypothesis-based analysis with exact and numbered quotes, verdict, and confidence level. To check : The transcripts are anonymized and deposited in the authorized tool: a user interview contains personal data.
2.3 Drafting the innovation brief: stakeholders, constraints, success criteria
Framing the challenge. Time saved: 3 hours per briefing. When the challenge is formulated (sheet 1.3) and the reference document is required which commits the sponsor, frames the team and will serve as the arbiter when selecting the concepts.
What you get: a mock interview that reveals questions you didn't ask; a two-page brief in which each assumption is marked [hypothesis]. To check : Each [hypothesis] is confirmed or ruled out with the actual sponsor before broadcast.
2.4 Guided ideation: cross-sectoral analogies, then combination and enrichment of concepts
Ideation. Time saved: 2 hours per session. When the first batch of ideas (sheet 1.4) remains too close to what already exists.
What you get: five mechanisms from elsewhere, with a real example each and its transposition; fifteen combined concepts, traceable until they borrow. To check : The "real" examples of the analogies are verified: this is where the models invent companies and figures.
2.5 From concept to use case, storyboard and concept sheet
Concept design. Time saved: 3 hours per concept. When a concept is selected and it needs to be made tangible for the users, the committee and the teams who will build it.
What you get: an eight-step usage scenario that shows the frictions, not just the benefits; a storyboard ready to be drawn or modeled, with the organizational needs. To check : Each step of the scenario is based on a real insight: mark those that are imagined.
2.6 Evaluate a batch of ideas using the grid, then challenge the finalists with simulated user profiles.
Evaluation and prioritization. Time saved: 3 hours per selection committee. Before a selection committee: you have ten to thirty ideas or concepts, the grid of sheet 1.6, and you want to arrive with a reasoned pre-assessment and the objections already heard.
What you get: an evaluation table where each note cites its proof or admits "not specified"; duplicates to be merged and eliminations explained. To check : The notes are proposals submitted to the committee, never a final ranking: present them as such.
2.7 Design the experimental protocol, then analyze the results and decide
Experimentation and testing. Time saved: 4 hours per experiment. When the riskiest hypothesis is designated (sheet 1.7).
What you get: three protocols compared and a recommendation, with thresholds set before the test; an analysis that separates what the data proves from what it suggests. To check : The thresholds were correctly written before the test and have not changed: keep the document from round 1.
2.8 Write the business case and roadmap for a concept, and prepare a project proposal.
Deployment, pitch and funding. Time saved: 1 day per case. When a concept has passed its initial tests and it is necessary to obtain an internal budget or external funding (Bpifrance, France 2030, ADEME, Horizon Europe, region, foundation).
What you get: a business case in three scenarios where each hypothesis is visible and modifiable; a roadmap in three phases with passing criteria. To check : Each figure in the business case is either yours or marked [to be confirmed]: redo the key calculations by hand.
Level 3 · Systematize: AI works for you
3.1 Create your permanent innovation assistant
Framing the challenge (cross-cutting). Time saved: 2 hours per week thereafter. As soon as you find yourself piecing together the strategy, the selection grid, the brief, and the in-house vocabulary for each prompt, the innovation assistant is the foundation for the other seven sheets at this level.
What you get: an assistant who knows your strategy, your grid, your formats and your current challenges; the other 23 cards launched in one sentence (“card 1.5 on idea X”) without re-explanation. To check : The attached documents do not contain any data that your charter prohibits from being stored in the tool: unregistered concepts, partner data under confidentiality agreement, personal data.
3.2 The monitoring system: automated monitoring of trends, competitors, patents and calls for projects
Exploration and monitoring. Time saved: 3 hours per week. When the watch is done in fits and starts, before a committee meeting, and never in between.
What you get: Each week, a short, dated and sourced summary is posted in the team's channel; calls for projects are identified with their deadlines, early enough to prepare a file. To check : Every Monday, ten minutes to open the links to the facts that you will relay: the watchdog proposes, a human relays.
3.3 The living insights database: collecting, labeling, and querying everything users say
User insights. Time saved: 4 hours per month, and studies avoided. When the team has accumulated interviews, supporting verbatim, opinions, surveys, in ten different cases, and each new project starts a new study because no one can find what has already been said.
What you get: each source indexed in five minutes, tagged, linked to what already exists; sourced answers in exact quotes to "what do we know about…", with the degree of confidence. To check : All sources are anonymized before submission, and the space is that which the IT department and the data protection officer have validated for this type of data.
3.4 The ideation chain: massive divergence, regrouping, scoring, shortlist, with safeguards against uniformity
Ideation. Time saved: 1 day per challenge, instead of a week. When a challenge warrants exploring a very broad avenue before converging.
What you get: 200 ideas in ten perspectives, grouped into families of mechanisms; a classification by grid, with its areas of weakness. To check : The team's ideas were gathered before launching the channel: a study published in Science Advances in 2024 shows that the help of AI makes individual productions better and more similar to each other.
3.5 From storyboard to testable prototype: narrative mockup, test page, clickable screens
Concept design. Time saved: 1 day per prototype, instead of several weeks. When a concept needs to be tested against real users (sheet 2.7) and waiting for development is too expensive.
What you get: a two-minute script and a presentation page ready for an interest test; a clickable mockup of the key screens, with dummy data, in one day. To check : no real data, no customer names, no internal figures in the prototype: it will be seen by outsiders.
3.6 The portfolio evaluator: apply the grid to all submitted ideas, identify duplicates, and monitor the portfolio.
Evaluation and prioritization. Time saved: 1 day per month for a portfolio manager. When the team receives dozens of ideas (suggestion box, internal call, intrapreneurs) and the evaluation takes weeks, is uneven depending on the evaluator, and leaves duplicates and already tested ideas coming back every year.
What you get: Each submitted idea is checked, compared to the portfolio and pre-evaluated in minutes, with a response to the bearer; a monthly portfolio review that shows gaps, blockages and arbitrage opportunities. To check : The pre-assessment is never sent directly to a sponsor or committee: a human pair reviews and signs it.
3.7 The experimentation log: test maps, learning maps, pivot memory
Experimentation and testing. Time saved: 1 hour per week, and tests not repeated. When the team conducts several experiments in parallel and the learnings get lost in the reports.
What you get: updated test and learning maps every week, and protocol deviations reported; three learnings per week shared with the team. To check : A threshold that is modified during testing is reported and discussed, not silently corrected.
3.8 The funding co-pilot and committee: from the identified call for projects to the submitted application, and the rehearsal of the presentation
Deployment, pitch and funding. Time saved: 2 days per case, instead of two weeks. When the watchdog (sheet 3.2) has identified a relevant call, the deadline is approaching, and several contributors need to be mobilized.
What you get: a case plan aligned with the grid, distributed among contributors and dated; sections rewritten in the vocabulary of the jury, consistent figures, unsubstantiated claims flagged. To check : The intellectual property department has validated what is written about the invention and in which tool the file was written: a filed file is a disclosure.
5. What a record contains, with an example
Each sheet fits on one page and follows the same format: a heading (situation, level, compatible tools, order of magnitude of time saved), "When to use it," "What to prepare," the prompt to copy with its parts to be replaced highlighted, "What you should achieve," "Check before using," "Variations depending on the tool" (Copilot in Teams, Word, Excel or PowerPoint, Researcher agent; ChatGPT, Claude and Gemini with their advanced search; Gemini Notebook; Vibe), and "Next step." Here, to set the tone, is the prompt for sheet 1.7, the map of critical assumptions of a concept.
Here is the value proposition of a concept: [paste page 1.5]What we already know: [evidence available].
1) List all the assumptions that must be true for this concept to succeed, categorized into four groups: desirability (users want it), feasibility (we can do it), viability (it creates value for the organization), and adaptability (the context, regulations, and partners allow it). Formulate each assumption as follows: "We believe that [who] [will do / accept / pay what] because [reason] 2) Rate each hypothesis on two axes: importance (if it is false, does the concept die?) and current level of evidence (do we have facts or opinions?). 3) Identify the three riskiest hypotheses: important and poorly proven. 4) For each hypothesis, propose the quickest and cheapest test that would confirm or refute it, with the numerical result that would lead to "it's validated" and the result that would lead to "we stop."
And the three accompanying verification points: the level of evidence assigned by the AI corresponds to what you actually have, hypothesis by hypothesis; the proposed tests are feasible within your organization; the validation thresholds are set before the test, not after seeing the results. It is this discipline, repeated across 24 worksheets, that transforms a trial into a skill.
6. The five safety rules of the innovator
An innovator manipulates what an organization holds most vulnerable: ideas that don't yet exist, research results, partner information, and customer feedback. The toolkit's user guide sets out five rules, to be applied in addition to your organization's AI charter, which should never replace it.
- The authorized tool, for the data it authorizes. Internal data goes into the company tool validated by the IT department; R&D results, personal data and partner information never go into a free, consumer-grade tool.
- An unregistered concept is a secret. An invention is only patentable if it has not been disclosed before filing. Before describing an invention to an AI, publishing a prototype, or filing an application, consult your intellectual property department; the INPI's e-Soleau envelope establishes the date of your work at a low cost.
- Anonymize before pasting, and respect the agreements. Names of people, clients, partners: labels are all that's needed for the prompt to work. Information received under a confidentiality agreement remains within the scope defined by the agreement, including the tool itself.
- The AI proposes, the team decides. No grid note, no market figure, no AI-generated business case hypothesis enters committee without a human owner who knows its origin. A simulated user profile prepares for an interview; it does not replace it.
- Protect diversity, be transparent, keep track. Write down your ideas before consulting AI, ensure diverse perspectives, and retain a degree of human input in each selection. Mention the use of AI when required by the charter and keep the prompts and exits that informed a decision. Since February 2, 2025, Article 4 of the European AI Regulation requires employers to take measures to develop AI proficiency among those who use it; Omnibus Regulation (EU) 2026/1744 has made this an obligation of means ("promoting the development of AI proficiency"): this toolkit contributes to that.
7. What the kit contains
The kit is a 43-page Word document, with the sections to be replaced highlighted in yellow in each prompt. It opens with instructions: why this kit, the three levels and the six-week learning path, the twelve-question self-assessment, the prompt template with four ingredients (objective, context, sources, expectations) and the verification reflex, the five safety rules, and the key for the prompt sheets. Next come the three levels, eight sheets each, each level beginning with what you will be able to do and ending with a validation challenge and its passing criteria. The fifth section is for the innovation team and the lab: the Monday AI quarter-hour, the shared prompt library, the team's eight usage rules, the five indicators for measuring gains, and the role of the innovation manager. Six annexes complete the kit: the use case matrix and the "which tool for what" table, the self-assessment grid by level, the checklist for verifying an AI output, the progress log, the glossary and the sources.
The accompanying Excel workbook contains nine ready-to-fill tabs: the weighted selection grid, the idea evaluation with calculated scoring, elimination and ranking, the hypothesis map with its risk level, the test maps with pre-set thresholds, the experimentation log, the idea portfolio with its alerts, the user profiles for role-playing games and the progress log.
Download the innovator's AI kit (Word + Excel, free)
43 pages, 24 one-page fact sheets, three levels, six appendices, and a nine-tab binder. Written by NextStart.AI based on BCG studies on AI in innovation and R&D, work from the Wharton School, Harvard and Science Advances on AI and creativity, Anthropic's AI Fluency framework and workshops conducted with innovation teams from large organizations.
- Self-positioning, prompt framework and five innovation-specific safety rules
- 24 ready-to-copy prompts, from monitoring to funding, in three levels
- A validation challenge by level and a six-week course
- Excel workbook: selection grid, hypotheses, test cards, idea portfolio
8. Frequently Asked Questions
Who is this kit for?
For innovation, R&D, product, and transformation teams in large organizations, as well as for labs and intrapreneurs, regardless of their starting level: this twelve-question self-assessment indicates where to begin. Innovation departments can distribute it as is, as a self-training course.
Is training required before using the kit?
No. The kit is designed for self-sufficiency: each sheet explains what to prepare, provides the prompt to copy, and lists the points to check. Training on your specific topics accelerates skills development and addresses the unique needs of your organization, but it is not a prerequisite.
What tools does it use?
With Microsoft 365 Copilot, ChatGPT, Claude, Gemini and Vibe by Mistral (formerly Le Chat). The prompts are neutral; each card indicates useful variations when the tool changes the game, for example, in-depth search for a state of the art, Gemini Notebook for an insights database, Copilot in Excel for a selection grid or Claude's Artifacts for a clickable mockup.
Can an invention be described to an AI?
Not without caution. An invention is only patentable if it has not been disclosed before filing: the kit advises consulting your intellectual property department before describing an invention in a tool, publishing a prototype or filing a dossier, and using only tools authorized by your organization for this type of information.
Won't AI homogenize our ideas?
This is a risk that research has identified: AI improves individual ideas but also makes them more similar. The toolkit avoids this through three rules: write down ideas before consulting the AI, ensure diverse perspectives in the ideation prompts, and retain a degree of human input in each selection. The ideation workflow described in sheet 3.4 is built upon these safeguards.
How long does the journey take?
Six weeks, 20 to 45 minutes per day, two weeks per level, with a validation challenge at the end of each level. The kit can also be used on demand: the matrix in Appendix A allows you to directly find the sheet that corresponds to your current situation.
9. Going further
The kit allows you to progress independently; it does not replace a workshop on your own topics. AI training for innovation teams NextStart.AI's 1.5-hour or half-day training program gets your teams working on their explorations, concepts, and real-world portfolio, with 70% hands-on practice, and they leave with the prompts and agents built during the session. Level 3 worksheets also link to resources from the library app.nextstart.aiincluding the evaluator of the portfolio of ideas and the monitor of calls for projects and funding.
Read also : AI Toolkit for Project Managers: 24 Use Cases and Their Prompts, Measuring the uses, value and ROI of AI: the 2026 toolkit, AI Ambassador Network: The 2026 toolkit for building your collective, Manager's toolkit for dealing with AI: getting your team on board et AI Project Sponsor Security and Data Kit.
The entire series of free kits. To drive adoption: AI charter template, AI training plan template, AI Ambassadors Kit, Manager's toolkit for dealing with AI, Adoption Roadmap Kit, AI maturity grid, AI Measurement and ROI Kit, AI Governance Kit, Gift Sets Cowork, Copilot Studio kit, Claude Skills kit et Vibe coding kitFor the AI project sponsor: AI Act kit, GDPR kit, Security and data kit, Shadow AI kit, AI and Social Dialogue Kit, frugal AI kit, AI Sovereignty Kit et responsible AI kitBy profession: AI HR kit, AI marketing kit, AI finance kit, AI kit for lawyers, AI kit for salespeople, AI toolkit for project managers, innovator's AI kit and communicator's AI kit.
10. Sources
The figures and references cited on this page and in the kit are based on the following sources, consulted in October 2026.
- Boston Consulting Group, Innovation Systems Need a Reboot (Most Innovative Companies 2024), June 2024.
- Boston Consulting Group, Executive Perspectives: AI-Powered R&D, February 2025.
- Boston Consulting Group, Most Innovative Companies 2025, June 2025.
- Karan Girotra, Lennart Meincke, Christian Terwiesch and Karl T. Ulrich, Ideas are Dimes a Dozen: Large Language Models for Idea Generation in Innovation, Wharton School, 2023, and Managing Innovation with Generative AIMack Institute, January 2024.
- Fabrizio Dell'Acqua, Ethan Mollick et al., Navigating the Jagged Technological Frontier, Harvard Business School Working Paper 24-013, September 2023.
- Anil R. Doshi and Oliver P. Hauser, Generative AI enhances individual creativity but reduces the collective diversity of novel contentScience Advances, July 2024.
- James Brand, Ayelet Israeli and Donald Ngwe, Generative AI Is Reshaping Early-Stage Market Research, Harvard Business Review, July 2025; BI Norwegian Business School, Can AI facilitate market research and innovation?, November 2025.
- Rick Dakan, Joseph Feller and Anthropic, AI Fluency Framework, 2025 (CC BY-NC-SA 4.0 license).
- Board of Innovation, AI tools for innovation ; Alexander Osterwalder and David J. Bland, Testing Business Ideas, Wiley, 2019.
- INPI, What is a patentable invention?
- Directorate General for Enterprises, France 2030 call for projects “AI Pioneers” European Commission, EIC Accelerator.
- Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence, Article 4, Mastering AIapplicable from February 2, 2025.