Automating Proposal Preparation
Proposal preparation time reduced from ~8 hours to under 90 minutes per case, with consistent output quality.
The challenge
Advisers were spending the equivalent of one full-time employee per month on research and proposal assembly that could be systematised — leaving the work that needed human judgment under-resourced.
The sourcing and proposal preparation process had four compounding inefficiency problems. First, it was entirely manual: each proposal started from scratch regardless of how similar it was to a previous one. Second, it was unstructured: yield calculations were done in spreadsheets that were not standardised across advisers, meaning the same inputs produced different outputs depending on who ran the numbers. Third, it was not connected to a data layer: neighbourhood analysis was conducted ad hoc from a mixture of Rightmove, Google, and memory — with no repeatable methodology and no record of what research had been done on which area. Fourth, the proposals themselves varied significantly in format and completeness, creating quality inconsistency that was occasionally visible to clients. When multiple simultaneous sourcing requests arrived — common during busy periods — the team simply fell behind, creating bottlenecks that delayed client responses by days.
The system
Decision system built
We designed and implemented an AI-augmented workflow covering three stages of the sourcing and proposal process: automated property sourcing (filtering against client criteria across multiple sources), structured yield and financial modelling (standardised calculation templates with pre-loaded local rental comparable data), and proposal generation (a templated output layer that assembled the sourcing findings, financial model, and neighbourhood analysis into a consistent, client-ready document with minimal manual formatting). The system was designed to produce a first-draft proposal — sourcing summary, yield calculation, area profile, and investment framing — that an adviser could review, adjust, and send, rather than build from scratch. Neighbourhood analysis was systematised using a defined scoring methodology covering transport links, school catchments, rental demand signals, and capital growth indicators.
System components
Automated property sourcing layer: multi-source filtering against client criteria (budget, area, property type, yield target, timeline) with ranked shortlist output
Standardised yield and financial modelling engine: gross yield, net yield, cash-on-cash return, and mortgage scenario calculations from a single data input set
Neighbourhood analysis scoring model: transport, schools, rental demand, and capital growth indicators compiled into a repeatable area profile
Proposal generation layer: structured template assembling sourcing shortlist, financial model, area profile, and investment framing into a client-ready document
Quality review checkpoint: adviser review and annotation step before any proposal is sent, ensuring human judgment on client fit and negotiation positioning
How we worked
Engagement scope
Process audit of existing sourcing and proposal workflows, AI tooling design and configuration, yield modelling standardisation, neighbourhood scoring methodology, proposal template design, and adviser onboarding.
Timeline
Six-week build with a two-week adviser adoption period before the workflow was treated as standard operating procedure.
Operating model
Built alongside the advisers who would use the system daily. The proposal template was drafted, used on real cases, revised, and used again before being locked — we did not ship a design that hadn't survived contact with actual client requirements.
Outcomes
Business impact & measurable results
Proposal preparation time reduced from ~8 hours to under 90 minutes per case, with consistent output quality.
Average proposal preparation time reduced from approximately 8 hours to under 90 minutes per case — a reduction of over 80%, freeing approximately 120 hours of adviser capacity per month
Yield calculation standardisation: all financial outputs now produced from a single methodology, eliminating adviser-to-adviser variance in how investment returns were presented to clients
Simultaneous case capacity increased: the team can now handle multiple sourcing requests in parallel without queue-based delays, reducing average client response time from days to hours for the sourcing stage
Proposal quality consistency: output format and section completeness are now uniform across all advisers and all cases — no client receives a materially different quality of proposal based on which adviser is handling their case
Structured neighbourhood data layer built as a byproduct: area profiles produced for sourcing purposes are stored and reused, meaning the marginal cost of each subsequent research request in a known area falls toward zero over time
Governance
Trust, collaboration & governance
Adviser review checkpoint is non-negotiable in the workflow design — the system produces a draft, not a final output, and that distinction was explained to the team and to clients
Yield calculation methodology was documented and explained, not treated as proprietary — advisers understand what the numbers mean and can answer client questions about them
The system was designed to make the adviser faster, not to replace them — every component was evaluated against whether it left the adviser more time for the work that actually requires their judgment
Reframe
Automation works when the adviser trusts the output enough to put their name on it.
Across every engagement, the goal is the same: engineer a system that makes better decisions — faster, more consistently, and at scale — than the process it replaces.
Next steps
Related services
Start a discovery
Most engagements begin with a conversation about context.
We do not send a proposal before we understand the problem. Start by telling us about your decision context — we will identify the highest-leverage intervention areas before any scope is agreed.