All Case Studies
  • Intelligent operations
  • Workflow automation
  • Proposal generation

Nexus Labs: a proposal drafted automatically when the discovery call ends

Writing proposals was eating too much time and producing inconsistent work. This is how we found where the time was going and what we built so that the proposal starts itself.

Client

Nexus Labs

What we built

Quill

Share

At a glance

Client
Nexus Labs, a production studio that generates multiple proposals a week across a range of services
The problem
The manual proposal process took too much time and produced inconsistent output.
What we found
The call already held everything the proposal needed, and the hours went into rewriting it afterwards.
What we built
Quill, an automated workflow in which a notetaking assistant joins the discovery call and a proposal draft is ready for review when the call ends.
The result
A full draft generated automatically, which a person reviews and edits before it goes out.
01

The problem

Nexus Labs is a production studio that generates multiple proposals a week across a range of service offerings. Writing them by hand was eating too much time and producing inconsistent output.

The pattern is familiar to most service businesses. You finish a discovery call with a clear picture of what the client needs and how to solve it. Then someone has to turn that conversation into a structured, professional document that the client can review, approve and sign.

That document takes two to three hours to produce on a good day. On a bad day, when three proposals land in the same week, it takes longer, looks inconsistent, and sometimes goes out with the wrong client name still in it from the last template.

The delay has a cost as well. Service businesses across Africa often lose deals not because the client said no, but because the proposal arrived four days after the call, when the energy had already dropped.

To protect Nexus Labs' commercial pricing and client information, the version of Quill shown here runs on a representative set of service packages.

02

The diagnosis

The finding was that the hours were not going into thinking, because the thinking happened on the discovery call. Everything the proposal needed had already been said out loud, and the time went into rewriting it afterwards: turning notes into a structured document, matching the right service packages and pricing, and keeping the formatting consistent. The work also started only when someone found the time, which is where the delay came from. Those are the parts a system can take over.

03

The solution

We built Quill as an intelligent operation, which means the work starts and runs by itself once a discovery call begins, and a person stays in charge of the result. It does not send anything to a client on its own. It removes the blank page and the waiting, and the thinking stays with you.

  1. 1

    A notetaking assistant joins the discovery call. It takes notes for the whole conversation, so nobody on your team has to.

  2. 2

    The call ends and Quill gets to work. It pulls out the client, industry, problem, requirements, timeline and budget signals.

  3. 3

    Quill drafts the scope of work. It matches two or three of the service packages to the project and writes all eight sections: Executive Summary, Scope of Work, Phasing and Timeline, Deliverables, Investment, Exclusions, Assumptions and Payment Terms.

  4. 4

    A person reviews and edits. Every section can be changed in place, and nothing goes to the client until a person approves it.

  5. 5

    Download a branded PDF. Choose a colour palette and a font, and the PDF comes out with a cover page, the matched pricing packages and clean formatting.

04

The results

  • Automatica full draft generated when the call ends
  • 8sections generated together
  • 7service packages matched against the pricing database
  • 80%first-draft accuracy reported by early users of the first version

We first tested Quill with a group of service business owners, in the first version where call notes were pasted in by hand. One response summed up what we were hoping to hear:

“I generate three to four proposals a week. This cut my average time from two hours per proposal to about 12 minutes. The output needed light editing but it was already 80 percent of the way there on the first run. I sent one directly to a client with only minor changes and it closed.”

The pattern across the group was consistent. The first draft was not perfect, and it was not meant to be. It was good enough that editing it was faster than writing from scratch. The automated version takes the step before that away as well, because the draft is waiting when the call ends.

What it could mean for you (an illustration, not a measured result)

If an owner or account manager spends two hours on every proposal and writes three a week, that is six hours a week. At a conservative billing rate of $50 an hour, that is about $300 a week, or roughly $15,000 a year. Early users of the first version reported about 12 minutes per proposal. It does not count the deals lost when a proposal arrives late.

05

What this means for your business

When we build this for a specific business, the pricing is replaced with that business's own packages, rates and deliverables. The assistant is set up for the calls that matter to its sales process, the extraction is tuned to the signals that matter in its discovery process, and the PDF carries its own branding.

Proposals are one of the places where work gets stuck. The Business Immersion follows your work from first enquiry to final payment and shows you where the next fix is.

How we protect your data

Quill works with confidential client conversations and commercial pricing, so protecting them shaped how we built it.

  • Nothing goes to a client until a person has reviewed and approved it, and every section can be edited.
  • Pricing is drawn from the business's own service packages, so the Investment section reflects what the business actually offers.
  • The AI drafts and the person decides, so the user stays in control of what the client sees.
  • For this case study, Quill runs on a representative set of service packages, so Nexus Labs' commercial pricing and client information are not published.

Found this useful? Share it.

Share

The Business Immersion

Find out where work gets stuck in your own business.

The Business Immersion is a one-working-day look at how work moves through your business, from first enquiry to final payment, ending with a clear plan for what to fix first.

See how the Business Immersion works
Nexus Labs: a proposal drafted automatically when the discovery call ends | Refactrd