Can an agency send 60% more proposals with the same team?
Yes. A proposal agent drawing on the agency's case library, winning proposals, and pricing matrix generates a first draft in 2.5 hours versus the previous 2 days. Volume increased by 60% with the same team. The hit rate went from 22% to 31% — equivalent to ~4.8M DKK in extra won deals per year.
What was the problem?
The agency sent ~40 proposals/month. Each proposal took an average of 2 days — the industry average is 25 hours according to Inventive AI 2026 data. Seniors were the bottleneck: they had to touch every proposal because juniors did not know the case library and pricing matrix well enough to produce consistent first drafts.
The hit rate stood at 22%. The agency lost proposals primarily because they arrived too late, were too generic, or used case references that were not relevant to the prospect's industry.
How is the system built?
- 1
Knowledge base indexing
The agency's 4 years of cases, proposals, and delivery documentation were indexed in a searchable knowledge base. The agent can retrieve precisely the relevant references for each new proposal.
- 2
Scope generation
The agent analyzes the brief/requirements from the prospect and generates scope, deliverables, timeline, and price estimate based on the pricing matrix and comparable historical projects.
- 3
Senior review
The first draft is sent to a senior who reviews, adjusts, and approves. The agent never sends to the client. This is a deliberate choice — proposals require business judgment.
- 4
Feedback loop
Won and lost proposals enter the knowledge base and continuously improve the next generation of drafts.
Concrete results (measured over 4 months)
- —Proposal time: 2 days → 2.5 hours — 87% reduction. Inventive AI benchmark is 90%.
- —+60% proposals sent with the same team during the period.
- —Hit rate: 22% → 31% (+41%). Inventive AI reports up to +50% in similar implementations.
- —Estimated annual revenue impact: ~4.8M DKK in extra won deals.
- —Senior time freed for strategy, pitches, and client relationships.
Questions about this case
- What is the benchmark for AI-generated proposals in an agency?
- Inventive AI reported in 2026 that agencies using AI reduce proposal time by up to 90% and see hit rate improvements of up to 50%. This case reached 87% reduction and +41% hit rate — in line with Inventive AI data.
- Does a human approve the proposal before sending?
- Yes. The agent generates a first draft. A senior reviews and approves everything before sending. The agent never sends autonomously — agency proposals require business judgment.
- What is included in the proposal agent's knowledge base?
- The agency's complete case library, pricing matrix by service type, historical winning and losing proposals, and standard scope blocks. The agent draws on what is relevant for each new proposal.