Last October, a consultancy firm started their journey with AI. The management team met with AIAIAI to discuss the challenges facing their sector: pressure on fees, non-core tasks taking up valuable time, and maintaining control over how AI is used within the firm. Today, 60 agents are active. By the end of the year, that number will exceed 100. Most of these were designed by the employees themselves, rather than management or IT. This case study focuses on one highly valuable agent: Peter, the proposal specialist.

Writing proposals is time-consuming with uncertain returns

Drafting a proposal is a tedious and time-consuming process. The outcome is never guaranteed, making it feel like unpaid work. It almost always requires a senior team member to get the proposal right. It is precisely the kind of work that gets delayed and causes frustration.

Peter drafts proposals based on past successes

This prompted the firm to build Peter: an agent designed to draft proposals. Peter does not start from scratch; he uses existing information to build his drafts.

Peter uses successful proposals, price lists and availability as a starting point

Peter draws from the firm's current resources, including successful and unsuccessful proposals, agreements made via email, price lists, and team availability. Using this information, he drafts a proposal tailored to the client, built on what has worked before.

Figure 1: Peter drafts a sales proposal.

A second agent checks the proposal against guidelines

Once the proposal is ready, Peter shares it with Sylvia, a second agent. Sylvia immediately reviews the document against internal guidelines, functioning as a digital editor. This allows the agents to check each other's work before a team member reviews and finalises the proposal.

About the platform

Build agents yourself, no coding required.

Peter was not built by IT. In AIAIAI, team members build agents themselves. By giving the agent a name, clear instructions, and the right sources, it is ready within minutes. This is how the firm created 60 agents, designed by the people who understand the work best.

Almost 4,000 hours saved per year, and counting

The maths is straightforward. The firm writes around 100 proposals a year. Each proposal takes an average of 50 hours, totalling nearly 5,000 hours annually. Peter is not perfect, but already completes up to 80% of the proposal. The remaining work involves review and formatting by a team member. This saves approximately 4,000 hours a year. That 80% completion rate is expected to rise this summer, as Peter begins using the firm's own templates.

The next step: explore what an agent like Peter can do for your firm

If you would like to see what this could mean for your firm, please get in touch.