Table of contents
Business challenge
Preparing a high-quality offer for a complex B2B project is a multi-stage process. It requires close collaboration between sales, presales, solution architects, and delivery teams to ensure the proposal is technically sound, commercially viable, and perfectly aligned with the client’s needs.
Our established manual process, while effective, was becoming a bottleneck. It was time-consuming and difficult to scale as the number and complexity of inquiries grew. This created several business risks:
- Time-intensive and difficult to scale: each offer demanded significant manual effort, making it hard to respond to market opportunities quickly.
- Increased risk of lost opportunities: a longer quoting cycle meant a greater chance of a potential client choosing a faster-moving competitor.
- Operational burden on expert teams: our highly skilled presales team spent too much time on repetitive, operational tasks like gathering information and formatting documents, rather than focusing on high-value activities like solution design and client consultation.
We needed to re-engineer the process to be faster, more efficient, and more scalable, without sacrificing the quality and customisation our clients expect.
Building an AI co-pilot for the entire sales cycle
Our solution was not a single, off-the-shelf tool, but a bespoke ecosystem of specialised AI components integrated directly into our presales workflow. All of the tools mentioned below (save for Napkin AI) were developed in-house and tailored to our organisation’s specific needs.
This system supports the team at every stage of the process, acting as a co-pilot that handles repetitive work and provides data-driven insights.
Automated inquiry analysis and context building
The process begins the moment a new opportunity is identified. An AI agent analyses the initial email from a salesperson, automatically identifying key information such as the client’s name, core business need, and the initial scope of the request.
It then creates a prospect record in our CRM and generates a concise summary. This provides the presales team with immediate context, allowing them to engage with a new opportunity without the need for manual data entry and information gathering.
AI-powered research and content generation
An AI assistant named Jarvis supports the team’s daily work by acting as a research and content-generation partner. It analyses client-provided documentation, summarises long reports into key points, and answers specific questions about the project’s context.
Jarvis can also search through historical proposals, scanning how similar services were scoped and presented before, so the team can build on what has already worked rather than starting from scratch.
It also drafts initial versions of standard content such as service descriptions and role profiles, and performs a final review of the offer to ensure linguistic quality and consistency.
The result is significantly less manual work and faster content preparation across the board.
Automated project structuring and estimation
Based on the client’s documentation, a tool called Scope Smith automatically generates a foundational structure for the project. This includes:
- A high-level project overview
- A breakdown of the project into functional areas and phases
- A preliminary feature backlog
- Suggestions for architectural decisions (ADRs)
- PERT-based effort estimations
- A preliminary risk register
- A Work Breakdown Structure (WBS)
- A suggested team composition
This dramatically shortens the time needed to create the initial solution concept and estimate, allowing the team to focus on refining the details rather than building the framework from scratch.
Enhanced communication through AI visualisation
To ensure complex technical concepts are clearly understood by the client, we use tools like Napkin AI to quickly generate diagrams, process flows, and other visual structures. This helps to organise requirements and architectural plans into a format that is easy to discuss and agree upon, improving communication and shortening alignment cycles.
Access to organisational knowledge
An AI agent named Lucien acts as a gateway to our entire body of organisational knowledge. It supports the presales team by providing instant access to relevant case studies, lessons learned from past projects, and contextual recommendations based on similar work. This prevents teams from “reinventing the wheel” and ensures that every new proposal benefits from decades of collective experience.
Data-driven process improvement
The AI ecosystem also provides a powerful analytics layer. It automatically processes completed offers to classify technologies, services, and business contexts, building a structured knowledge base.
This enables us to generate detailed reports, analyse sales effectiveness, and identify trends. Furthermore, an AI-driven “Client Drive Scoring” model assesses the quality of new opportunities on a 0-5 scale, analysing the clarity of needs, business context, and expected results.
This data-driven approach allows us to continuously learn and refine our sales process.
Rapid prototyping with AI-assisted development
Using tools like Replit, Cursor, and Claude, our teams can rapidly build and present working Proofs of Concept (POCs) to clients during the presales phase. This ability to demonstrate a functioning solution, rather than just talking about it, significantly increases the credibility and quality of our offers.
Benefits of the project
- 56% reduction in the time required to prepare a comprehensive sales offer
- 40% increase in the throughput of the 9-person presales team, allowing us to handle more opportunities
- 29% reduction in the average cost of preparing a single offer
- Annual operational savings of 760,000 PLN by avoiding the need to expand the team to handle the increased workload
- Over 10,000,000 PLN in potential additional annual revenue enabled by the increased number of offers prepared at the improved conversion rate
- A full return on investment (ROI) was achieved in under 12 months from the start of the project