Decision Framework

Build, Buy, or Bespoke. A scored decision tool for private credit, real estate credit, and real asset funds evaluating AI and data tooling.

Written for fund principals at private credit, real estate credit, and real asset funds, not technology teams. The question is rarely which tool to adopt. It is which operating posture fits each workflow, the team, and the next three years of the fund. Run the diagnostic once per workflow. The answer is rarely the same twice.

Scope Screening, memo, monitoring, LP, regulatory
Audience Private credit, real estate credit, specialty finance, real assets
Horizon 12 to 36 months

What counts as a workflow

Working definition

A workflow is a repeatable, end-to-end task that today consumes a meaningful share of senior time. It has a trigger, a defined output, and an identifiable owner inside the fund.

  1. W1 Deal screening Trigger: ABL teaser, IM, or CRE teaser package with rent rolls and cap rate assumptions arrives // Output: pass, proceed, or escalate summary for IC routing Sets the top of the funnel. Senior origination time lost here compounds across the fund.
  2. W2 Credit memo and IC pack Trigger: deal progresses past screening // Output: underwriting model, comps, structuring view, IC-ready memo The document on which capital is committed. Structure is standard, inputs are bespoke.
  3. W3 Portfolio monitoring Trigger: quarterly close, covenant, collateral, or rent coverage event // Output: LTV updates, covenant checks, rent coverage and occupancy reads, early-warning flags, partner view Where losses are prevented or missed. Multi-asset, multi-jurisdiction, no vendor depth.
  4. W4 LP reporting Trigger: quarterly and annual cycle, or ad-hoc LP query // Output: NAV, exposure, attribution, SFDR, side-letter tracking A commodity output to a non-commodity audience. Errors cost institutional trust.
  5. W5 Regulatory and compliance reporting Trigger: AIFMD II filings, EU Taxonomy, concentration limits, depositary queries // Output: regulator-ready submissions Filing deadlines are fixed. The data pipeline that feeds them is not.

What you are actually choosing between

Most funds collapse the decision to build or buy. There is a third path that sits between them, and for mid-market private credit and real estate credit it is often the one that fits. The three paths below are described on their own terms, not ranked.

Buy off-the-shelf
Vendor SaaS // Licensed
WhatA packaged product you license. Configuration, not construction. The vendor owns the roadmap.
CostLow-to-mid five figures a year, scaling with users, seats, or AUM. Predictable opex.
GetWorking product from day one. Support, updates, audit trail, usually SOC2.
Give upFit to your process. You adapt to the tool. Differentiation is hard to sustain.
Time2 to 8 weeks to production for commodity workflows. Longer if data integration is custom.
Key riskThe vendor does not cover your asset class, or covers 60 percent of it, and the gap becomes yours to staff.
Build in-house
Internal Hire // Owned IP
WhatYou hire engineers and data people. You own the code, the data pipeline, and the roadmap.
CostEUR 1.2m to 1.8m fully loaded over year one for a credible team of three. Fixed overhead on the fund.
GetExact fit to your process. Proprietary IP. Full control over security and data residency.
Give upTime, management attention, and a new discipline to run. Hiring in a market where you do not compete on pay.
Time9 to 18 months to reach a production system that partners trust on live deals.
Key riskYou end up with a team you cannot retain, or a codebase only one person understands.
External bespoke build
Forward-Deployed // Co-Built
WhatA senior operator sits inside your workflow, builds against your actual deals, and hands over working systems.
CostEUR 30K to 80K per workflow, fixed scope, fixed fee. Capped engagement. Opex, not headcount.
GetFit close to in-house, delivered at off-the-shelf speed. Knowledge transfer built into the engagement.
Give upVendor-scale support. The IP question is a contract term, not an assumption.
Time4 to 12 weeks to the first production workflow. Further workflows compound on the first.
Key riskPartner dependency. The engagement must end with you running the system, not the builder.
// Cost of inactionDoing nothing has a price. It is not zero.

Six originators at EUR 800 fully-loaded per day. Two hours per teaser on manual screening. Roughly 300 teasers per year at mid-market origination volume. That is ~EUR 180K of senior origination time absorbed by first-look screening alone, before counting deals missed or mispriced due to screening fatigue.

Portfolio monitoring adds a comparable amount when covenant checks, rent coverage reads, and early-warning reviews are still stitched from spreadsheets. Any of the three paths above must clear that bar, not a zero bar.

Matching the path to the problem

Buy wins when
  • The workflow is a commodity. LP portals, fund accounting, KYC, e-signature.
  • A dominant vendor has product-market fit for your asset class at your scale.
  • Your process is standard enough that adapting to the tool is cheap.
  • Switching cost is acceptable if the vendor stalls or gets acquired.
Build wins when
  • The capability is a structural edge, not a utility. It belongs on the fund's balance sheet.
  • You already run a technology function at scale and can absorb another team.
  • Data sensitivity or regulatory posture rules out third-party code paths.
  • You have a 10-year horizon on the capability and can amortize the build.
Bespoke wins when
  • Your deals are idiosyncratic. Collateral types, structures, or jurisdictions differ across the book.
  • No off-the-shelf vendor covers your asset class with real depth.
  • You want proprietary fit without standing up a permanent engineering team.
  • You have one or two workflows where getting it right is worth the attention.

Why the generic framework underfits these asset classes

The five factors below apply to both private credit and real estate credit, with different emphasis. A commercial real estate fund will weigh collateral diversity and jurisdictional spread differently than an asset-based lender, but both feel each one.

// 01
Collateral diversity
Commercial real estate, infrastructure, renewables, receivables, specialty lending collateral, NPL portfolios. Each collateral type has its own valuation logic, covenants, and early-warning signals. Generic portfolio tools flatten this.
// 02
Jurisdictional spread
Multi-country origination means multi-language documentation, local legal structures, and different data registries. Vendors built for a single market rarely travel well.
// 03
Bespoke deal structure
Every transaction is negotiated. Waterfalls, guarantees, step-ups, rent coverage covenants, and security packages do not fit a fixed schema. Screening and monitoring must accommodate the non-standard.
// 04
Thin vendor ecosystem
Unlike CLOs or syndicated loans, there is no Intex, no Bloomberg-grade depth. A handful of credit-agnostic platforms exist. None cover asset-based lending or CRE credit end-to-end.
// 05
Regulatory load
AIFMD II for loan-originating AIFs, SFDR Article 8 and 9 disclosures, EU Taxonomy, and fund-level concentration reporting. Compliance is data work, not policy work.

Six questions, one score, one default posture

Answer each question for the workflow in front of you. Each answer maps to a path (Buy, Build, or Bespoke) and carries a weight. The running totals update live. The highest total is the default posture for that workflow. Run this diagnostic once per workflow. The posture that wins for LP reporting will lose for deal screening.

Assessing as
// Begin

Take the assessment

Six questions, answered one at a time. Each click auto-advances. You can go back. No email required to see your posture.

~3 minutes Keyboard-friendly Save as PDF when done
// Your default posture
Bespoke
Score: Buy 0 // Build 0 // Bespoke 0
Opens your browser's print dialog. Choose "Save as PDF".
// Optional

Want this in your inbox? Enter your email and we'll send a copy of your score and posture, saved so Sam can pick up the thread if you want to talk.

No spam. You'll only hear from Prospekto if you ask to continue the conversation.

One fund, five decisions

Most funds make five decisions here, not one. The posture that wins for LP reporting will lose for deal screening. Default postures below are tuned for a mid-market private credit or real estate credit fund of roughly EUR 1 to 5 billion AUM. Treat them as priors, not conclusions.

WorkflowDefault postureWhy
LP reporting Buy Commodity workflow, multiple credible vendors, regulated output. The differentiation is your returns, not your NAV template.
Regulatory and compliance reporting Buy + Bespoke integration Vendor handles filing and format. Bespoke handles the data pipeline that feeds it from your deal stack. AIFMD II deadlines do not wait.
Credit memo and IC pack Hybrid: Buy templates, Bespoke extraction Memo structure is standard. Extracting terms from bespoke deal docs across five jurisdictions and four languages is not.
Deal screening Bespoke Idiosyncratic deals, no vendor fit, highest partner-time drag. First-look triage is where senior hours leak.
Portfolio monitoring Bespoke Multi-asset collateral, multi-jurisdiction, covenant and rent coverage complexity. No vendor covers asset-based lending or CRE credit end-to-end.

What to put in the contract

The bespoke path becomes buyable for a risk-conscious partner only when the exit is written down. Five line items. None are unusual in institutional contracting. All are negotiable upfront and expensive to retrofit.

// Non-negotiables on day one of the MSA
Five clauses that make bespoke ownable
  1. IP assignment. All code, models, and configuration assigned to the fund at delivery, not licensed. No residual rights sitting with the builder.
  2. Source-code escrow or direct handover. Fund holds the repository. Builder retains no kill-switch, no hosted dependency, no runtime lock.
  3. Knowledge-transfer milestones. Defined training deliverables and documentation, tied to payment tranches. If the handover is unfinished, the final tranche is unpaid.
  4. Defined exit. The engagement has a stated end date and a stated end state. Handover, not dependency. Extension is optional, not structural.
  5. No-reliance clause. Fund can operate the system without the builder from day one of handover. Written into the contract, demonstrated in the UAT.

Run the diagnostic on a live workflow.

If the scored answer lands on bespoke for one or two of your workflows, the next conversation is a scope. Fixed fee, fixed timeline, handover written into the contract.

Happy to walk through this on specific workflows when the internal conversation is ready.  /  Sam Griek, Prospekto
v3.0 // Interactive // Prospekto.app