AI is finding its place in real estate practice through specific tasks where it saves time and improves the quality of decisions. This is another example. The bank tender process is one application that most investment and asset management teams will recognise immediately.
Consider the following situation. The investment committee has reviewed a value-add investment opportunity and has been generally supportive. The investment team has agreed most commercial terms with the seller. Signing is four weeks away. Due diligence is almost done and the external advisors have identified no deal breakers nor significant red flag findings. Final due diligence reports are expected within days. The SPA draft is in good shape with a few points still being worked through. Things are moving well. But the weeks ahead are demanding. And through all of this, the final IC approval rests on two conditions: satisfactory completion of the SPA and documentation and a qualified term sheet for the debt financing to be in place prior signing of the SPA. A qualified term sheet in this case is not just a confirmation that a lender will provide debt. It is evidence that the market has been properly tested and that the selected structure fits the investment and is in line with the business plan. The IC is not asking for the fastest answer. It is asking for the most attractive one. AI helps get there without sacrificing the time that everything else at this stage also needs.
The bank tender and why it matters
A structured bank tender is standard practice in institutional real estate. Testing several lenders, receiving term sheets and comparing them properly is part of any credible financing process. And with a conditional IC approval in play as in this case, it is not optional.
Banks shift. Lending appetite moves with balance sheet cycles, sector concentration, regulatory pressure and internal strategy. The lender that was most competitive on a similar transaction eighteen months ago may be in a very different place today. A tender tells you where the market actually is. A well run tender also creates competitive tension. Not just on margin. On the full package. Covenants that were presented as standard may become more flexible. Fees that looked fixed may get reconsidered. Prepayment terms that seemed immovable may improve. This is where a tender earns its benefits.
What drives differences between lender offers
The margin often gets the most attention. And this is fine. But it is not the only variable.
What actually shapes an offer is a combination of factors that experienced investment teams analyze and negotiate. Balance sheet capacity and sector concentration, cost of funds and funding model, relationship value and appetite for ancillary business are some of the relevant aspects to be considered. Covenant corridors and calculation methodology across ICR, DSCR and LTV or LTC. Tenor availability and duration appetite. Hedging requirements and the total costs for the debt financing. Two banks can look at the same asset on the same day and arrive at materially different debt financing offers. Not because they see the asset differently. But because their own positions, constraints and appetites are different.
Mapping those differences accurately across lenders, on a consistent basis, and running them through the scenarios that matter for the business plan is exactly where AI adds real value. It does not just surface the differences. It helps quantify them and structure them in a way that makes the decision clearer to everyone involved.
The situation on the ground
Five lenders have submitted term sheets. The margins sit within a range of 78bps between the lowest and highest offer. All five banks are credible. One of them is a lender the team has worked with before. The instinct is to move quickly and finalise with them. But a qualified term sheet requires more than a preferred relationship. It requires a defensible comparison. And the details matter.
Income treatment during the void and lease up phase is the first area where the offers diverge. Two lenders calculate the ICR on contracted passing rent only. The other three include a proportion of estimated rental value for vacant space during an agreed stabilisation period. On an asset with meaningful vacancy at acquisition, that difference can produce a technical covenant breach under the first two structures from day one, even though the business plan is entirely credible. AI helps to model this across all five lenders against the actual leasing assumptions in the business plan, not just the headline ICR threshold.
Leasing covenants may add a further layer. Three lenders require formal consent before exchanging any material lease above or below an agreed rent level. Two require consent only outside a defined corridor. The practical difference is response time. In a competitive leasing market, having to wait for bank approval before exchanging with a tenant can mean losing the occupier entirely. AI maps consent thresholds against the letting strategy to show where each lender's structure creates operational friction.
Also, cash mechanics vary in ways that affect liquidity planning directly. One lender includes a cash sweep provision. Another requires a fully funded CapEx reserve held in a blocked account from day one. LTV covenant trigger levels vary significantly across the five offers. Modelling these against the projected cash flow at each stage of the business plan reveals which structures create liquidity pressure at exactly the moments when the refurbishment programme needs funding.
Valuation mechanics go beyond frequency. Some lenders require an annual third party valuation at the borrower’s expense. Others accept annual valuations by their internal teams, with the costs covered by the bank. More importantly, some build in automatic covenant resets triggered by valuation movements outside the agreed cycle. On a seven year value-add hold in a market where valuations can move, the difference between a lender who waits for the scheduled test date and one who can call an additional valuation if conditions deteriorate is a material risk distinction. AI can help to run sensitivity analysis on valuation declines against each lender's trigger structure to show where headroom disappears first.
Business plan governance may vary from lender to lender in ways that are easy to miss in a margin focused comparison. Two lenders require formal sign off on material changes to the letting strategy or CapEx programme. One requires annual business plan resubmission and approval. The others accept an agreed plan at outset with notification of material changes only. Over a four year hold with an active asset management programme, the governance overhead of the more restrictive structures is real.
Cure period mechanics are among the most important aspects to be agreed in value-add cases. Cure periods on covenant breaches range from sixty days to none across the five offers. Equity cure rights differ in how often they can be exercised and whether a cure resets the covenant level or simply prevents default on that specific test date. AI stress tests each structure against a downside scenario to show which lender's package gives the team the most room to manage through a difficult period without triggering a formal lender conversation at the worst possible moment.
And finally in this case, prepayment terms range from clean to structures that make an early exit expensive if the business plan delivers ahead of schedule.
None of this is buried. It is all in the documents. The issue is not knowledge as these are tasks any professional investment team cover. It is time. Mapping all of it consistently across five lenders, running it through the business plan scenarios that matter and presenting it in a form the IC can act on. AI does not make the decision. It helps ensure the decision is based on a full and consistent comparison of the term sheets.
Where AI adds value
Term sheet comparison is methodical work. It requires consistency and attention across a large amount of detail. This is where AI is strong and where humans working under pressure across multiple workstreams are most at risk of missing something.
Standardising covenant definitions across five documents onto a common framework. Running covenant calculations on each lender's methodology and restating them consistently. Stress testing headroom against downside scenarios involving higher vacancy, softer rents, or increased capital expenditure. Mapping cash mechanics against the liquidity moments that matter in the business plan. Modelling prepayment costs under different exit timings, etc.
Work that would normally take several days can now be completed in far less time. The investment team is not removed from the process. Their time shifts to reviewing, challenging, and interpreting rather than building from scratch. The time recovered goes to the questions that need experience and judgement. Which covenant package actually works for this specific business plan. Where to push back in the lender negotiation. Whether the preferred relationship is still the right choice once the full picture is on the table.
Beyond the margin
The margin is easy to present and easy to compare. It travels well through an organisation. For these reasons it tends to dominate the financing conversation even when it should not. The covenants are what the team lives with. A cash sweep that restricts liquidity at the wrong moment can cost far more than twenty basis points. A covenant methodology that produces tighter headroom than expected during a soft period constrains decisions at exactly the wrong time. A prepayment structure that makes an early exit expensive affects the whole return profile. These are the terms that shape the investment objective and the leveraged returns the strategy depends on. AI ensures they get the same rigour as the margin, within the same timeline.
A word on oversight and confidentiality.
A word on oversight and confidentiality.
Two things deserve straightforward acknowledgement
First, confidentiality. Term sheets and financing structures contain sensitive commercial information. They should not go through general purpose consumer AI tools. Organisations need to use AI solutions that meet their security and confidentiality standards. Many institutional real estate firms are already working through this. Those that have not should treat the financing process as exactly the context that requires those policies to be in place first given that not only term sheets will be provided to AI but also business plan details.
Second, oversight. AI produces a strong starting point. It is not a conclusion. A qualified term sheet requires proper professional review. Outputs need to be checked. Assumptions need to be tested. Covenant language, lender relationships, and the fit between the financing structure and the business plan all require experienced judgement. AI gets the team to that review faster and better prepared. It does not replace it. Also, the relationship calls still matter. What AI changes is how the available time gets used. More of it goes to the analysis and judgements that only experience can handle.
Created with
