Trapol8Forty managers answer the same hundred questions. The responses arrive over six weeks, in whatever order the managers get to them. Somewhere in that pile: six answers that changed from last cycle, two that were answered around rather than answered, and one that does not look like anybody else's.
Finding those nine is the job. Collecting the four thousand is not, and has not been for some time.
Yet most evaluations are still run on the collection features. Vendors demo the questionnaire builder, the distribution list, the reminder schedule and the completion dashboard, because those are easy to show and every platform has them. The analysis layer, which is where the platforms actually differ, gets ten minutes at the end.
This guide inverts that. It covers what the category contains, seven criteria for the part that varies, and the questions worth asking every vendor.
What This Category Actually Contains
Platforms that sound identical in marketing copy are often built for opposite ends of the same transaction. Before comparing features, confirm which end a vendor serves.
LP-Side Diligence Platforms
Built for institutional investors and consultants who send questionnaires and need to manage the responses. These handle distribution, tracking, response storage and reporting across a manager roster. Dasseti and DiligenceVault are the established platforms in this category. If you run diligence on managers, this is the category you are shopping in.
GP-Side Response Platforms
Built for the asset managers, hedge funds and PE firms receiving your questionnaires. They draft responses from a firm's approved content and route them through internal compliance review. You are not the buyer here, but what your managers use determines the quality and speed of everything you receive.
General-Purpose RFP Tools
Loopio, Responsive and Qvidian were built for sales teams answering procurement RFPs. They appear in DDQ searches because the shape of the problem looks similar. They have no concept of a manager roster, a diligence cycle, or an ILPA or AIMA template, and are not a serious option for either side of institutional diligence.
Platforms Spanning Both Sides
A newer category: one platform serving the allocator sending and the manager answering. Trapol8 is here, having built the manager side first and added the allocator side afterwards. The relevance to a buyer is not tidiness. It is that response quality is an input to your analysis, and a platform that owns both ends can affect it.
Seven Criteria for the Analysis Layer
First, what to stop evaluating. Questionnaire creation, distribution, deadline reminders, completion tracking and response storage are table stakes. Every platform in this category does all five competently. If a demo spends its first half hour there, you are watching a presentation designed around what is easy to show rather than what will determine whether the product earns its cost.
The seven below are where platforms diverge.
-
01
AI That Is Native, Not Appended
Almost every vendor in this category now has an AI feature. The question is whether the analysis was designed with the product or added to a collection tool that already existed. The tell is where it sits in the workflow: native analysis runs on responses as they arrive, so the review is already underway when the last manager submits. Appended analysis is a separate step you remember to run afterwards, on a dataset you export. Ask when the AI features shipped relative to the platform.
-
02
Can You Tell It What To Watch For?
Diligence concerns are specific and they change. This year it is key-person provisions because two teams lost a founder. Next year it is valuation policy. A platform whose analysis is fixed at implementation cannot follow that, and a platform that requires vendor involvement to redirect it will be redirected once and then never again.
Separate two things when a vendor answers. Onboarding the platform, meaning loading your questionnaires and manager list, is legitimately a project. Configuring what the AI looks for should not be. Ask directly: if I want to start watching for a specific concern this afternoon, what happens? If the answer involves a training cycle, a rules exercise or a support ticket, the capability exists on the datasheet rather than in your workflow.
-
03
Comparison Across Cycles, At The Answer Level
Manager drift is rarely announced. It shows up as a reworded valuation policy, a risk limit revised upward, a reporting-line change buried in a paragraph about governance. Detecting it means comparing this cycle's answer to the same manager's last one, and doing it on free text rather than only on structured fields and regulatory filings.
Filing-level comparison, such as diffing successive Form ADVs, is genuinely useful and considerably easier, because the documents are standardised. Confirm which one a vendor is describing. They are not the same capability and the second does not imply the first.
-
04
Comparison Across The Manager Set
Some answers are only interesting relative to their peers. A leverage figure, a fee structure, a valuation frequency: each is unremarkable alone and revealing when it is the only one of its kind among forty. This requires the platform to hold the responses in a form comparable question by question, and to surface the answer that does not fit rather than leaving you to notice it on the third read.
Ask to see it on a real set rather than in the abstract. Cross-manager comparison is claimed more often than it is demonstrated.
-
05
Reporting You Can Shape
Analysis that cannot leave the platform has limited value, because the audience for it is an investment committee, a board, or a client. The question is not whether the platform exports, but whether the output can be shaped to the argument you are making rather than the vendor's fixed template. A report you must rebuild in a deck is a report the platform did not really produce.
-
06
What Happens At The Manager's End
This is the criterion most evaluations skip, and it determines the quality of everything above it. Your analysis operates on what managers send back. If responding is painful, you get late submissions, partial answers, and last year's text pasted forward without review. No analysis layer recovers information that was never written down carefully.
Almost every LP-side platform optimises the collector's experience and treats the manager as a data source. Ask a vendor what the responding experience is, and better, ask two of your managers what it was like the last time they received a questionnaire through that platform. It is the single most informative question in an evaluation and almost nobody asks it.
-
07
Migration From What You Send Today
Most allocators arrive with a questionnaire library refined over years, living in Word, Excel or PDF. Rebuilding it inside a new platform is the hidden cost of switching, and the reason evaluations stall after a successful pilot. Ask whether you can upload the blank questionnaires you send today and have them digitized with the questions, order and wording preserved, or whether someone on your team will be retyping for a fortnight.
Platforms Compared
Described from each vendor's public materials as of August 2026. This category is moving quickly and marketing pages lag shipped features in both directions, so treat this as a starting point for questions rather than a substitute for a demo.
- Questionnaire creation from ILPA, AIMA or your own question library
- Distribution to a manager roster with deadlines, reminders and completion status in one view
- AI variance monitoring: answers that changed since a manager's last cycle, missing or evasive responses, and outliers across the manager set
- Analysis directed in plain language, applied to incoming responses as they arrive
- Existing blank questionnaires uploaded and digitized, preserving questions, order and wording
- Built the manager side first, so responding through it is a designed experience rather than a form to fill in
- SOC 2 Type 2 certified; customer data never trains AI models
- Established LP-side diligence workflow, data collection and manager database
- Used by endowments, pensions and family offices to manage manager diligence at scale
- AI review agent for surfacing gaps and inconsistencies requiring follow-up
- Form ADV change detection, comparing successive regulatory filings
- A governed prompt library for reusing approved prompts across teams
- Also ships manager-side AI, so a manager may encounter it from the other direction
- LP-side DDQ distribution, tracking and response aggregation
- Diligence pipeline management and manager scoring
- Built around the investor collecting and evaluating responses
- Not designed for the manager drafting and routing them internally
- Answer libraries and AI drafting for sales teams responding to procurement RFPs
- No concept of a manager roster or a diligence cycle
- No ILPA or AIMA support, and no cross-respondent comparison
- Listed here because they surface in DDQ searches, not because they belong in this evaluation
How to Choose
Confirm the category first. If a vendor's customers are mostly asset managers answering questionnaires, the product was built for the other side of your transaction, whatever the allocator-facing page says.
Then run three tests, all of them on your own material rather than a curated demo dataset.
- The instruction test. Name a concern you actually have, in your own words, and ask the vendor to make the system watch for it while you are on the call. Time it. This single test separates an analysis layer you will use from one you will configure once.
- The cycle test. Load two cycles of real responses from managers you know well, and see what the platform surfaces. You already know what changed. A platform worth buying finds it, and ideally finds something you missed.
- The manager test. Ask two managers in your book what it was like to respond through that platform. If the vendor cannot name managers who have, that is itself the answer.
Implementation timelines are worth probing separately from analysis setup. A multi-month platform onboarding is not unusual for a large roster. A multi-month wait before the AI can be pointed at a new concern is a different problem, and a permanent one.