Review · risks · conditions
Project due diligence:
from questions to a decision
Due diligence is not about accumulating documents. Its purpose is to test the key assumptions of a project, find circumstances that could change the decision, and translate findings into clear terms of a transaction, financing or launch.
Anton Konnov · 21 August 2026 · 11 min read
Perimeter
Depth of review depends on the decision and the risk
Before requesting documents, record the subject: an investment, a credit, an asset purchase, a partnership or an internal launch. Specify the amount and form of involvement, the project stage, the decision criteria and the factors that could genuinely alter the decision.
A review of an early-stage project differs from that of a going concern. The perimeter must account for industry, jurisdiction, scale, data maturity and the cost of error; specialised legal, tax, technical and environmental questions require qualified specialists.
First request
Which materials to start with
Below is a starting list to adapt to the agreed scope. It is not an exhaustive professional checklist: what is needed depends on the transaction, industry and jurisdiction, and specialists will add to their own areas.
The decision and the project. A brief description of the proposed decision — amount, form of involvement and timing — the project stage and the current presentation or memorandum.
Financial model and track record. The model with an explicit list of assumptions and a version date. For an existing business, financial statements and cash flow for the periods available; for an early-stage project with no operating history, cost estimates, the basis for key assumptions and the funding plan.
Customers, suppliers and demand. Material contracts, letters of intent, the sales pipeline and pilots, showing separately what is signed and what is still under discussion.
Rights, assets and obligations. Records of rights to key assets, permits and licences, loans, pledges, guarantees and material disputes.
Implementation. Schedule, budget, team and key contractors, plus known open issues and risks.
If a document does not exist or is still in progress, say so, and give a version and date for every figure. A clearly flagged gap is more useful than data put together in a hurry to answer a request: it simply becomes a review question.
No documents are needed for a first enquiry: a short note on the decision, the project stage and the materials already available is enough. Confidential files are shared once the scope and a secure exchange method have been agreed.
Six steps
How to conduct project due diligence
Frame the question and criteria
Define what must be confirmed for a go decision, which risks are unacceptable and which gaps can be closed with conditions. Build a risk-weighted work plan rather than a standard checklist for all projects.
Set up the data room and audit trail
Maintain a request register, received versions, owners and open questions. A material finding must reference a document, system, interview or independent confirmation; the absence of data is itself a finding.
Assess the market and operating model
Match claimed demand against contracts, pipeline and customer behaviour; evaluate competition, pricing, capacity, technology, suppliers, personnel, permits, timelines and critical dependencies. A forecast must be achievable given the project's resources.
Assess finances, law and tax
Reconcile management data with primary sources, normalise one-off effects, review working capital, debt, pledges, guarantees, obligations and liquidity needs. Separately evaluate rights to assets, material contracts, disputes, licences and corporate and tax structure.
Evaluate the team, integrity and E&S risks
Check competencies, delegation of authority, motivation, dependence on key individuals and reputational or conflict circumstances. Environmental, social, health and other sector-specific risks are assessed by the scale of potential impact and the team's ability to manage them.
Translate findings into a decision
For each material question, specify the fact, source, uncertainty, potential effect and action. The outcome may be a reject, additional testing, a price or structure adjustment, a precondition, a warranty, a limit, tranches or a remediation plan with monitoring.
Areas of review
A unified picture instead of separate opinions
Commercial
Market, customers, competitors, pricing, channels and demand confirmation.
Operations and technology
Technology, capacity, resources, suppliers, timelines and permits.
Finance and tax
Quality of results, cash flow, debt, obligations and structure.
Law, team and sustainability
Rights, contracts, disputes, governance, integrity and E&S impact.
Red flags
What should change the course of the review
A key assertion cannot be linked to independent or primary evidence.
Document versions and figures change without an explainable audit trail.
A forecast requires demand, capacity, funding or permits that do not yet exist.
Material obligations, related parties or conflicts are disclosed late.
The team refuses to discuss a downside case or close a gap.
Example
How a finding becomes a decision condition
A hypothetical example, not a real client case. The model for a new production facility assumes that, from year two, most of its output will be bought by a few large customers.
Forecast assumption. Year-two sales volumes and prices, on which revenue and the funding plan depend.
What the documents show. Negotiations with prospective customers and several letters of interest; no signed contracts with volumes or prices yet.
What remains unknown. Whether orders will materialise, at what volume and price, and when.
What to request. Negotiation status for each customer, draft contracts, pilot supply terms, data on comparable sales where available and, where appropriate and agreed, direct conversations with customers.
What to test. Scenarios with lower volumes, a later start of sales and the loss of the largest customer: the point at which the project stops covering its obligations, and how much liquidity is needed to get through that period.
The next step depends on how material this demand is to the decision. If the project holds up even in the downside case, the issue can stay on the monitoring list. If not, options include a condition that key contracts are signed before the next tranche, staged financing, further market review or a revised plan with a smaller initial volume. There is no automatic approve-or-reject rule, and final transaction terms are set together with the relevant legal and financial advisers.
The report
A finding is only valuable alongside its impact and action
Separate the confirmed fact, the specialist's assessment and the unresolved uncertainty. Rank issues by probability, scale and reversibility, but do not bury a critical blocker inside an average assessment.
The final document should show a recommendation, the basis, review limitations, open questions, decision conditions, responsible parties and timelines. Appendices preserve the evidence; the main body remains readable without reconstructing logic from folders.
AI in due diligence
Faster at spotting discrepancies without losing verifiability
AI helps index a data room, reconcile document versions, extract obligations, build a query register and link every finding to its source. Specialists still determine materiality, check context and remain responsible for legal, tax, technical and investment conclusions.
Source materials
Methodological reference points
IFC: due diligence areas — integrity, tax, corruption risks and conflicts ↗
IFC: review, investment decision, conditions and subsequent monitoring ↗
This material is educational and does not replace legal, tax, technical, environmental or investment advice for a specific transaction.
First step
Frame the decision and the three key uncertainties
We will define a risk-weighted review perimeter and the format of a final finding.