Over the decades, project management has consistently leveraged technological advancements to streamline its processes and maximise the value of deliverables produced. Contemporary tools offer project managers an increased ability to accurately align schedules with resources, anticipate a wider range of risk factors, and synchronise documentation in real-time via the cloud.
However, since late 2022, a disruptive transformation has been taking place in this discipline with the emergence of increasingly powerful Large Language Models (LLMs). These systems, capable of simulating human conversations and generating original content, go beyond simple optimisation. They redefine the management of complex tasks, significantly increasing productivity and stimulating team creativity, particularly for activities such as drafting summaries or preliminary document analysis.
Personalised AI assistants: a key asset for the project manager
Personalised AI assistants are specialised versions of LLMs, designed to accomplish very specific tasks. Unlike generic models that respond to a multitude of queries, these assistants are trained to follow precise instructions and base their knowledge on contextual documents and information. This specificity allows them to be adapted to particular tasks or knowledge domains, thus offering users unparalleled relevance and accuracy in their responses.
To make these assistants truly effective, one must master user inputs. Learn about the future of Gen AI prompting and its challenges to understand this critical component.
This article will not cover the full range of LLM capabilities. It will focus specifically on Personalised AI Assistants (often called “Custom GPTs” or “GEMs”). We will define their precise role and explore why they represent a major asset in the daily life of a project manager.
Custom GPT: ChatGPT’s Personalised AI Assistant
GEMs : Gemini’s Personalised AI Assistant
Devoteam and AI: a strategy serving innovation
The integration of personalised AI assistants to boost productivity and innovation in project management is central to the AI strategy of Devoteam.
The strategic plan, “AMPLIFY,” aims precisely to solidify their leadership position by offering concrete and innovative solutions to their clients. Devoteam has already demonstrated its mastery of LLMs with projects like the deployment of a transactional chatbot for the Portuguese postal service, managing delivery tracking and adjustment requests. More recently, an AI assistant based on Gemini Pro was made available to 20,000 employees of a European hotel chain to streamline their functional calls. These initiatives highlight Devoteam’s AI expertise, which they are now directly applying to project management.
To deliver this added value, Devoteam’s project managers benefit from advanced training and privileged access to Google’s AI tools, such as Gemini. The “AMPLIFY” strategy, complemented by internal events like workshops and conferences, ensures the optimal and secure integration of these new technologies directly into project management processes. This allows them not only to maximise delivered value but also to develop new agile approaches.
Genesis of an internal AI initiative: from informal to structured
The year 2023 at Devoteam was marked by significant excitement surrounding the adoption of Large Language Models (LLMs). Spontaneously, informal working groups emerged, united by a common interest and a growing desire to collaborate. This dynamic was complemented by the creation of a club dedicated to artificial intelligence technologies. This enables valuable knowledge sharing through regular “Lunch and Learn” presentations and increased awareness of AI opportunities. This initial momentum laid the groundwork for more structured collaboration.
In 2024, these initiatives matured. The groups became more active and organised, conducting activities simultaneously in our agencies in Paris, Lyon, Marseille, and Toulouse. Continuing the “AMPLIFY” strategy, Devoteam established robust governance for these technologies, ensuring a structured and secure deployment framework. Simultaneously, our R&D branch launched ambitious AI projects, which will ultimately consolidate our competitive advantage.
This collective maturity led to a decisive step: today, in 2025, a working group specifically dedicated to project management has been formalised, endowed with clear objectives and concrete projects. Bolstered by Devoteam’s support and internal developments, our ambition is to model and develop the concept of the “AI-augmented project manager.” This article, the result of this group’s work, aims precisely to promote this key initiative and share our vision for AI-optimised project management.
A competition to amplify AI mastery
In this momentum, to deepen practical mastery and collaboration around AI, we organised the “Clash of Prompts.” This national competition challenged different groups to design and submit the most effective prompts to solve a given project problem, with the AI itself then being solicited to evaluate the proposals. This initiative was a valuable opportunity to practice the essential “patterns” for creating powerful prompts, to introduce our teams to the use of Google Gemini’s generative models, and to broaden our horizons thanks to the varied perspectives of the participating teams.
The multiple benefits of the initiative
We can identify three categories of benefits from this initiative:
- For the organisers, it’s an opportunity to play a pioneering role. By exploring new uses of AI applied to project management, they actively contribute to innovation while disseminating their expertise. Their position as “knowledge facilitators,” combined with strong pedagogical ability, allows them to effectively support their colleagues in this upskilling process.
- For the participants, it’s a concrete entry point into the world of personalised AI assistants. They were able to discover how these tools help them daily, and rethink some of their practices. Most importantly, reignite their appetite for innovation, creativity, and experimentation.
- For the organisation, the impact is threefold: strengthening its competitiveness in high-potential areas, stimulating synergies between teams spread across the territory, and fostering a company culture based on collaboration, sharing, and agility.
These cross-cutting benefits explicitly demonstrate the strategic contribution of personalised AI assistants, not only for the individual efficiency of project managers but also for the structural and cultural strengthening of Devoteam. In summary, this initiative transforms the approach to project management into a more innovative, collaborative, and high-performing process, perfectly aligned with the future challenges of the sector.
Optimising AI assistant usage: prompt engineering
For project managers and their teams looking to maximise the potential of personalised AI assistants, adopting rigorous “prompt engineering” practices and an iterative, ethical approach to AI is fundamental.
First, it’s essential to precisely define the AI assistant’s role and objectives. For example, clarifying whether the AI should act as a “senior project manager” for specific tasks (like risk analysis and summary drafting) or to achieve precise goals helps guide the AI toward more relevant responses and minimises deviations. Mastering prompt engineering techniques is crucial for the quality of instructions. It’s not just about asking a question but structuring it concisely, clearly, and specifically. The “CRAFT” approach (Context, Role, Action, Format, Tone) is a proven method for formulating precise queries:
- Context: Define the framework and key information.
- Role: Indicate which expert the AI should embody.
- Action: Explain precisely what is expected.
- Format: Define the structure of the response.
- Tone: Specify the target audience.
Providing rich and relevant context is fundamental. The more information the AI assistant has about the project, its environment, stakeholders, and cultural nuances, the more accurate and tailored its responses will be. Finally, particular attention to data confidentiality and security is imperative to prevent any disclosure of sensitive information outside the company.
An iterative approach is also essential when interacting with AI. Collaborating with an AI assistant is a multi-step process of reformulating queries and adding information to refine results. This progressive method allows for guiding the machine toward the project team’s intended goal and optimising the quality of the AI’s responses.
Generative AI could well become an indispensable co-pilot for every role. Not only in technical professions but also on the functional side.
Concrete use cases: personalised AI assistants in action
We will now illustrate some concrete use cases of generative AI assistants. What are their potential improvements for the roles of Project Manager, Product Owner, and Scrum Master?
| Project Management | Product Management | Agility |
| Rapid drafting of deliverable drafts (charter, plan, progress report) | Accelerated drafting of Epics, Features, and Stories, streamlining the product backlog | Generation of personalised and creative retrospectives |
| Meeting or workshop synthesis (via audio transcription or notes) and support for agile events | ||
| Assistance with planning (WBS, Gantt, milestones) | Generation of product roadmap drafts based on context and market (identification of key milestones) | Targeted agile coaching, suggesting new practices to experiment with based on context |
| Management of risks and opportunities (based on historical documents/projects and suggestion of response plans) | ||
| Team coaching and facilitation (proposal of activities adapted to context and needs) | ||
| Project Communication (emails, newsletters, slides, support for report translation, multilingual communications, personalisation for targeted stakeholders) | User Feedback Analysis (NPS, surveys, comments) | Agile Maturity Self-Assessment via automatic questionnaire analysis, with feedback and action plans |
| Preparation of Calls for Tenders (draft generation, technical summaries or annexes) | Competitive Benchmarking via automatic summaries (product sheets, competitor websites, customer comments) | Creation of Agile Artifacts (e.g., Definition of Done) based on team context and performance |
| Assistance with decision-making and prioritisation through scenario simulations and argumentation of options | ||
| Monitoring indicators and reporting through automated dashboard generation | Simulating personas, user journeys, story mapping, and impact mapping based on customer segments and pain points | Generating agile indicators and metrics, reducing manual reporting and promoting transparency |
| Training and upskilling (creation of personalised courses, quizzes, syntheses of best practices) | Training and upskilling (creation of personalised courses, quizzes, syntheses of best practices) | Training and upskilling (creation of personalised courses, quizzes, syntheses of best practices) |
| Co-creation of innovation via automated brainstorming on unexpected functionalities or scenarios in the face of a client problem | Co-creation of innovation via automated brainstorming on unexpected functionalities or scenarios in the face of a client problem | Co-creation of innovation via automated brainstorming on unexpected functionalities or scenarios in the face of a client problem |
In conclusion, thanks to AI assistants, we’ll be able to:
- Automate without dehumanising all repetitive, predictable, replicable, or improvable tasks.
- Improve our collaborative work by simplifying access to knowledge, exchanging direct feedback, and structuring ideas into clear action plans.
- Evolve responsibilities to prioritise deeper analysis, problem-solving through innovation, strengthened interpersonal relationships, strategic thinking, and the ability to mentor colleagues.
- Provide tailored tools to support these evolutions, especially for young project managers, including those from technical backgrounds, who need practical support to manage their projects effectively.
Next steps: ongoing AI acculturation
This initiative acts as a catalyst for continuous and profitable momentum within our organisation. In a context where generative artificial intelligence is exponentially integrating into an ever-wider spectrum of applications and tools, it becomes crucial to relentlessly reiterate and expand this technological acculturation experience.
Our fundamental objective is to continuously familiarize all our teams with the latest advancements in generative AI, as well as with best practices for utilising these innovations. This proactive approach aims to ensure a sustainable, effective, and informed adoption of these technologies, thereby enabling our employees to fully master the tools of tomorrow and seamlessly integrate them into their daily work processes. To this end, Devoteam provides several practical training courses tailored to the different areas of expertise of its consultants.
By investing in this continuous acculturation, we ensure that our employees remain at the forefront of innovation, capable of leveraging the opportunities offered by AI to optimise performance, stimulate creativity, and solve complex problems with increased efficiency. It is by cultivating this internal expertise that we consolidate our leadership position and can anticipate future market developments, thus transforming every technological challenge into a real opportunity for growth and innovation.
Risks, perspectives, and general conclusion
A deep understanding of AI’s capabilities and limitations is fundamental. Users must be fully aware of what personalised AI assistants can and cannot accomplish. For example, language models can sometimes “hallucinate” incorrect facts or references. Also, they cannot provide information on very recent events without integration with search tools. Critical evaluation and systematic verification of results are therefore essential, especially for tasks requiring human judgment.
It’s important to remember that we are communicating with an AI. Despite their sophisticated conversational interface, these tools are machines, not human beings. Anthropomorphising AI can lead to risks, such as excessive attribution of responsibility or overestimation of its capabilities. A pragmatic attitude allows for the best exploitation of AI’s technical efficiency.
Beyond these promising use cases, it is crucial to address the inherent risks in adopting generative AI. Data confidentiality remains a major concern, requiring robust solutions to prevent the leakage of sensitive information. Potential model biases, inherited from training data, can also distort analyses or recommendations, demanding constant vigilance and human verification. Finally, excessive reliance on AI could potentially erode certain human skills if not managed with discernment.
However, the perspectives are immense. Personalised AI assistants, by freeing teams from repetitive tasks, allow project managers, product owners, and Scrum Masters to focus on strategic analysis, creativity, and interpersonal relations, thus transforming their role into true catalysts for innovation. Ultimately, the thoughtful and ethical integration of generative AI does not just optimise existing processes. It fundamentally redefines project management, making it more adaptive, intelligent, and humanly enriching.
Special Thanks:
We wish to express our gratitude to our colleagues Yvan Masselin, Geoffroy Renaud Blanc Fontenille and Frédéric Delannay for their review in refining the content of this article.