Interior Design Project Timeline: Draft Phases and Lead Times With AI

Give the AI three inputs: your phase list, the date the room must be finished, and a table of lead times from your own supplier records. Ask it to work backwards to each order date and the client's decision date, add a buffer, then check every date in a spreadsheet before the client sees it.

The AI knows none of your suppliers' lead times and handles calendar arithmetic unreliably, so treat its dates as a draft to compare, never as the answer. What it does well is turn a rough brief into an ordered phase list, spot dependencies that are easy to forget (a curtain maker cannot take a final measure until the walls are painted) and write the client-facing version. The dates you promise come from your records and a spreadsheet formula.

Follow me on Instagram@sagnikteaches

Most timelines slip for a dull reason: an order was not placed because the client had not decided, not because the supplier was late. So every long-lead item below carries two dates, the latest day to place the order and the earlier day by which the client must have said yes.

Connect on LinkedInSagnik Bhattacharya

Take a living and dining room refresh with a fixed styling day on Friday 18 June 2027, five made-to-order items and one in-stock item. Every lead time is illustrative and every date was calculated with the formulas shown, so you can copy the method and replace the numbers with your own.

Subscribe on YouTube@sagnikteaches

Work backwards from the day that cannot move

A forwards plan starts from today, adds the phases in order and ends on whatever date results. It feels sensible until the longest item goes onto the chart and the end date jumps by three months. A backwards plan starts from the date that cannot move (a client's event, a lease start, a builder's handover) and subtracts. Any date that lands before today is a problem you now know about, which is the whole point.

Three dates matter for each item, and mixing them up causes most of the confusion.

The chain that reaches furthest back sets the pace for the whole project. In the example that is the sofa, which makes the sofa's decision date the real deadline for the client, whatever the plan says about paint or flooring.

A phase list that works for any room

Ask the AI for the phase list last rather than first. A vague brief produces a generic list that looks complete and hides the gaps. Give it the questionnaire summary, the room measurements and the trades you plan to use, then ask for phases with a testable finish condition: "ends when the specification is issued", not "specification phase". If you do not yet have a clean brief, the client questionnaire tutorial covers building one. These are the seven phases I'd start from.

Phases 4 and 5 overlap on every real project. The sofa order goes in during phase 4 while phase 3 is still finishing the specification for the bedrooms, and the decorator is already on site. That overlap is why the timeline has to be built from the items, not from the phases.

Build a lead-time register from your own supplier emails

The register is a table you own, with one row per supplier and item type. It is the single most useful thing a studio can keep, because it replaces "the upholsterer said about three months" with what the last three orders took. The FF&E schedule tutorial shows where the item list comes from; this table adds the time dimension. Use these columns:

• Item type and supplier, for example custom sofa, upholsterer.
• Quoted lead time in weeks, exactly as the supplier wrote it.
• Clock starts when, in the supplier's own words.
• Last three actual lead times, from order confirmation to delivery.
• Planned lead time, a formula you will set below.
• Notes: ex-works or delivered, minimum order, factory closures.

"Clock starts when" is the column that saves projects. Suppliers use phrases such as "production starts once fabric is approved and the deposit is received", "from approved drawings and site survey" or "from final measure". Two quotes that both say twelve weeks can start their clocks on dates a month apart. Ask for the exact wording and keep it.

Your order confirmations and delivery emails already hold the raw data, so use the AI to read them into rows. Do not ask it to calculate lead times. Here is a prompt that keeps it to reading.

I will paste emails from suppliers about past orders. For each order, extract one row with these columns:
item, supplier type, date order was confirmed, date delivered (if stated), the supplier's exact wording about when production starts.
Rules: copy dates exactly as written. Do not calculate durations. If something is not stated, write "not stated". If two emails disagree, list both on separate rows and add the word CONFLICT in a notes column.

Here is the kind of output you would get back (illustrative), using three past sofa orders.

Check that output against the emails before trusting it. The usual faults to look for are a delivery date taken from a dispatch notice rather than the delivery confirmation, two revisions of the same quote merged into one row, and a "weeks" figure that was really working days. Fix those by hand, then do the arithmetic in the sheet with =ROUND((C2-B2)/7,1). The three sofa orders come out at 14, 16 and 17 weeks.

From the register, set one rule and apply it everywhere: plan on the slowest of the last three actual lead times, plus one week for transport and inspection. The sofa was quoted at 14 weeks, so the plan is 17 plus 1, which is 18. For a supplier with only a quote and no history, add two weeks instead of one and mark the row as unproven so you ask the supplier to confirm dispatch dates at the halfway point.

A living and dining room refresh, dated item by item

Start with the site sequence, because it decides when each item must be in the room. The styling day is Friday 18 June 2027, and each step is placed backwards from it.

Now each made-to-order item. The planned lead time is the slowest of the last three actuals plus one week, the clock start is the needed-on-site date minus that many weeks (moved back to a working day), and the decision date is five working days before the clock start.

Read the table from the top and the project's shape appears. The sofa has to be ordered by 9 February 2027, which means the client has to choose fabric by 2 February, more than four months before the styling day and long before they are thinking about curtains. The pendant and the joinery follow in the second half of February and early March. Everything else has comfortable room.

One row hides a trap. The curtain maker's clock starts at final measure, and a final measure needs painted walls and a laid floor, which is not until 4 June. Eight weeks from there is the end of July. The way out is to ask the maker whether they will start from a measure taken off the drawings and check it on site, and to record their answer in the register. The first AI pass will not catch this unless you have written "from final measure" in the clock-start column, which is another reason to keep that column.

Spreadsheet formulas that keep the AI honest

Suppliers quote weeks, so subtract in calendar days, then step back to a working day. WORKDAY exists in Excel and in Google Sheets, and in both a negative number of days counts backwards. Microsoft's page also warns that dates typed as text cause errors, so enter them as real dates or with the DATE function. If you want a refresher on prompting inside a spreadsheet, the ChatGPT and Excel guide covers it. With the register in columns A to K, these three formulas do the whole job.

Columns: A item and supplier, B quoted weeks, C clock starts when, D to F last three
actual lead times, G notes, H planned weeks, I needed on site, J latest clock start,
K client decides by.

Planned weeks (H2):          =IF(COUNT(D2:F2)=0,B2+2,MAX(D2:F2)+1)
Latest clock start (J2):     =WORKDAY(I2-H2*7+1,-1)
Client decides by (K2):      =WORKDAY(J2,-5)

Column I holds the needed-on-site date. Add your factory-closure and bank-holiday
dates in a range and pass it as a third argument, for example:
=WORKDAY(I2-H2*7+1,-1,$N$2:$N$20)

The IF in the first formula applies the unproven-supplier rule: with no actual lead times on file, it plans on the quoted weeks plus two. The +1, -1 pair in the second formula looks odd but has a purpose: it returns the date itself when it is a working day, and the previous working day when it falls on a weekend. For the sofa, 15 June 2027 minus 126 days is Tuesday 9 February, so that is the answer. Five working days before it is Tuesday 2 February.

Give the AI the formulas rather than asking it to produce dates. A request such as "write the three column formulas for this layout and explain each in a sentence" works well, and you can paste the results into the sheet. For a longer explanation of why language models stumble on this sort of calculation, see why AI is bad at maths.

Prompt: reverse-schedule the project and show the arithmetic

When you want the AI to produce the first-draft timeline itself, make it show its working so you can compare line by line with the sheet.

You are helping an interior designer plan a project. Work backwards from the styling day.
Inputs: styling day, the trade sequence with durations (working days), and the lead-time register below, which already includes the planned lead time in weeks for each item.
For every item give: needed-on-site date, planned weeks, latest clock start, client decision date (five working days before clock start).
Show the arithmetic for each date in one line, for example "15 Jun minus 18 weeks = 9 Feb".
List every assumption you made, and list any item where the clock-start condition in my register could delay it.
Do not invent lead times. If a value is missing, say so.

The "list your assumptions" instruction is the most valuable line. It makes the AI say, in plain words, that it assumed the curtain clock started at the order date, which is exactly the sort of thing you would otherwise discover in May.

Where the AI's first sofa date went wrong

A first answer from the AI usually contains errors you can catch. The sort of slip to expect: it returns 26 May as the sofa's order date, because it subtracted 14 working days (the quoted weeks read as days, with no buffer) instead of 18 calendar weeks. The answer is fluent and confident, and the date looks plausible. The sheet says 9 February, a gap of more than three months, and the mismatch is obvious once the two are laid side by side.

Other faults to look for: an order date that falls on a weekend; an install date that is earlier than a delivery; the same trade scheduled on two overlapping days; and a "client decision" date that is later than the clock start. A thirty-second scan of the date column for those four patterns catches most of them.

Client decision dates, written so a client can act on them

Clients rarely miss deadlines on purpose. They miss them because "we need the fabric choice soon" does not say what will happen if they wait. Turn each decision date into a plain consequence and send the whole set once, early, as a short table.

A before and after shows the difference. The before is the message most clients get.

Before: "Hi, just a reminder that we need to finalise the sofa soon so we can keep to the schedule."

After: "Hi [first name], the sofa fabric and frame size are due by Tuesday 2 February. The upholsterer needs about 18 weeks from approval, so each day we wait moves your delivery, and the styling day, back by a day. I've attached the three fabric samples again with the shortlist."

Ask the AI to draft the after version from the table, then edit the tone to match how you write to that client.

Where the buffer lives

Put buffer in two places. The first is inside each item (the one extra week, or two for an unproven supplier). The second is a block of float before the styling day, which in the example is the three days between furniture delivery and styling. Do not hide a third layer inside phase durations, because you will no longer be able to see which numbers are real and which are padding.

A quick sum shows why the per-item rule matters. The sofa's quoted 14 weeks would have put the clock start on 9 March, four weeks later than the planned 9 February. If the supplier performs as it did on the last three orders, that four-week gap is the difference between a delivered sofa and an empty room. If the supplier performs at the quoted time, you gain four weeks of float that you did not promise to anyone.

That is also why the client should hear a week, not a day, until the last long-lead order is in production. Share the internal dates with the client only as decision deadlines. Keep the buffered install day in your own register, and narrow the client's install week to a day once the supplier confirms dispatch in writing.

When the client decides late: reflowing the dates

Suppose the sofa decision arrives on 16 February instead of 2 February, a 14-day delay. The clock starts around 23 February rather than 9 February, so delivery moves from Tuesday 15 June to Tuesday 29 June and the styling day to Friday 2 July. Everything else is unaffected, because nothing else sits on the sofa's chain. Ask the AI to reflow the dates, then compare against the formulas again.

The client decided the sofa 14 days late. The sofa's clock now starts on 23 Feb 2027, planned lead time 18 weeks.
Recalculate the sofa delivery date and the styling day. Then list three options for keeping the original styling day, each with its likely cost or trade-off, using only the information in my register. Do not invent supplier capabilities; where an option depends on a supplier, say "ask the supplier".

Typical options include an in-stock sofa from a supplier whose lead time is a few weeks, a rental or loan piece for the event, or splitting the room so the styling day goes ahead with the sofa to follow. The AI will list these; you decide which are real by phoning the supplier. Then send the client the choice with prices attached, not a problem with a question mark.

Check it worked: plan against actual after every project

When the project closes, add one row per order to the register with the actual lead time from confirmation to delivery. Keep only the last three per supplier. After six or seven projects most of your regular suppliers will have real figures, and the "plan on the slowest of three, plus one" rule starts to reflect how those suppliers truly behave.

The sofa in the example arrived 17 weeks after confirmation. The last three become 16, 17 and 17, the planned time stays at 18 weeks, and the row tells you the supplier has stopped drifting upwards. A supplier whose last three read 14, 16 and 19 is telling you something different: next time, plan on 20 and ask what changed.

Two further checks are worth five minutes at project close: count how many client decisions arrived after their due date, and note the item that set the critical chain. If the same type of item keeps setting the chain (custom upholstery is common), build that lead time into your proposals so clients hear it at the start. The quote-checking workflow is a good companion, because a quote that does not state a lead time or a clock start is one to query before anything is ordered.

What changes on a single-room job

A single room with no joinery and one or two made-to-order items needs the same method with less in it. Skip the phase table and keep three columns: item, planned lead time, decide-by date. Most single-room projects have exactly one long-lead item, so the plan is that item's dates plus the trade days around it.

Keep the register even then. A studio that handles ten small jobs a year builds more lead-time history than one that handles two large ones, and that history is what lets you quote a realistic install week at the very first meeting. If you also keep a chase sheet for open orders, the purchase order tracking tutorial shows a layout that works alongside the register.

Questions designers ask about AI-drafted timelines

Can AI tell me how long a supplier will take?

No. Lead times differ by supplier, season and fabric, and the AI cannot see your orders. Use the weeks quoted in writing, then compare them with how long your last three orders really took. Ask the AI to organise those figures and flag gaps, never to estimate them. For a supplier you have not used before, ask them to confirm in writing when their clock starts.

Which date do I give the client, the planned one or the buffered one?

Give the client an install week rather than a day until every made-to-order item is confirmed in production. Keep the buffered dates in your own register. Once the last long-lead item has a supplier-confirmed dispatch date, narrow the week to a day. Promising a day too early turns every small supplier slip into an apology.

What if the client's date is fixed and the numbers do not fit?

Find the item that sets the longest chain and offer options before the client discovers the problem: an in-stock alternative with a shorter lead time, a stand-in for that one piece, or a later delivery for that item only. Put the cost of each option in writing, then rerun the schedule so you can show the new dates for each choice.

How often should I rerun the timeline?

Rerun it whenever a lead time changes, a client decision arrives late or a trade moves its start date, and at every phase gate regardless. Keep one master sheet so the AI always starts from the same register, and save a dated copy each time so you can see when and why the end date moved.

Further reads

• Interior Design Proposal Template: Scope, Fees and Phases With ChatGPT — Write the phases and fee basis this timeline sits inside
• AI for Interior Design Studio Admin: Sourcing, Quotes, Updates — A wider look at sourcing, quotes and client updates
• Joinery Quotation Template: Bespoke Items, Finishes, Lead Times With AI — What a joiner's quote should say about lead times
• AI Scheduling for Small Builders: Subcontractors and Deliveries — How builders sequence trades and deliveries
• ChatGPT for Interior Designers: Ten Prompts for Briefs and Schedules — More prompts for briefs and schedules
• Best AI for Cabinet Making: Supplier Orders, Lead Times and Chasers — Chasing supplier orders when dates start to slip
• AI Supplier Management: Track Prices, Lead Times, and Risk — Keep prices, risk and lead times per supplier together
• Interior Designer Software Compared: Where AI Add-Ons Earn Their Fee — Six interior design tools priced per month, the AI credits worth buying, and a 14-day trial you can run on one finished project.
• Best AI for Interior Design Mockups: Client-Ready Images for a Studio — A six-step studio workflow from a client's room photo to an approved concept image: permissions, prompts, a scale check, labelling and a one-page sign-off.
• How to Price Interior Design Services: Compare Fee Models With AI — One three-room project priced hourly, fixed-fee and cost-plus, with margin checks, stress tests and client wording for each model.
• AI for Architects and Interior Designers: Who Does What on a Joint Job — Who should own which AI task when an architect and interior designer share a project: brief, mood boards, schedules, client updates and rights.
• Can AI Draft an Interior Design Contract? Terms a Solicitor Checks — Design-fee, revision, procurement and cancellation clauses drafted with AI, and the terms a solicitor should check before you use them.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Microsoft Support and Google Docs Editors Help pages for the WORKDAY function, read in November 2026. Every lead time, date and figure in the examples is illustrative.