The questions founders actually ask.
Not a capability list. In the calls that matter, nobody asks whether we can do the work, because the portfolio has already answered that. What they ask is how long it takes, what it costs to commit, what we need from their engineers, and what happens when the technology changes halfway through. Those are the questions below.
What do you need from us before the first working session?
A spec sheet or architecture diagram, your last investor deck, and thirty minutes with whoever can say yes. That is genuinely it. We do not send a sixty-question brand questionnaire, because the answers that matter are in the engineering and in the deck, and asking a founder to transcribe them wastes the part of the engagement they are actually needed for. If you have customer interviews or lost-deal notes, those are the single most useful thing you can hand over, and almost nobody offers them unprompted.
How much of our team's time does this actually consume?
Far less than a rebrand run by a generalist, because the technical review is shorter when the person reading your spec has built to one. Expect a kickoff, a positioning session, and then review points at narrative, design and build. The engineering time is concentrated in technical review of the copy and the 3D, which is usually two or three sessions rather than a standing weekly meeting. Founders consistently tell us the surprise is how little glossary work they have to do.
We are pre-revenue and pre-product. Is it too early to brand?
Usually not, but it changes what is worth building. Pre-product, the useful deliverables are positioning, category definition and an analyst-ready one-pager, not a forty-page brand book you will outgrow. The failure mode is spending a website budget before the story is settled, then paying twice. If your next twelve months are a raise rather than a go-to-market, we will say so and scope to that.
What happens if our technology changes mid-project?
In deep tech it frequently does, and the scope is built expecting it. Positioning and narrative are sequenced before identity and build precisely so a pivot is absorbed in the cheapest phase rather than after animation. If the change is fundamental we stop and re-scope rather than shipping a site that describes a product you no longer sell. That conversation is easier if it happens in week three than in week ten.
Can you work with our internal marketing team rather than replacing them?
That is the common arrangement. Most of our clients have one or two marketing people who are excellent and stretched. We take the work that needs specialist depth or a technical read, and hand back a system your team can run: a CMS they can update without calling us, a brand kit with real rules, and source files they own outright.
How do you handle a launch gated by a tender, regulatory or funding date?
We sequence backwards from your date, not forwards from ours. Armory's website revamp launched under NDA and UPSI, coordinated with PR and armed-forces tender announcements, which meant the disclosure calendar drove the schedule rather than the design one. If a date is immovable, say so at the audit call and it becomes the constraint everything else is planned around.
What does fixed scope actually mean in practice?
Fixed scope, fixed timeline, one price, quoted after the thirty-minute audit call. No day rates and no open-ended retainers. It means the number does not move because a phase ran long at our end. It also means a genuine change of direction from your side is a re-scope conversation rather than a silent overrun, which we would rather have out loud.
Can you start before our funding closes?
Often, and it is a common request, because a raise and a rebrand are usually chasing the same date. What we will not do is start a build on positioning we think is wrong in order to hit a roadshow. If the round is the deadline, the highest-value work is usually narrative and an analyst-ready one-pager, which is also the cheapest phase to buy first.
What is the smallest useful piece of work we can buy?
The paid discovery. It produces positioning, a category read and a written recommendation you own whether or not you continue with us. Most engagements start there, and a meaningful number stop there because the founder now has what they actually needed. We would rather be the studio that sold you the right small thing than the wrong large one.
Our competitors all sound the same. How do you find a position that is not taken?
By starting from the engineering rather than the market. Category sameness happens when everyone positions against the same analyst quadrant, which flattens exactly the differences that matter in deep tech. The defensible position is usually a constraint you have solved that others have routed around: a tolerance, a data engine, a certification, a supply relationship. Those are in the spec sheet, not in the competitive landscape deck.
Should we use the technical term or the plain-language one?
Both, in a specific order. The technical term earns credibility with the specialist and is what they search; the plain-language one lets their boss approve the purchase. The common mistake is choosing. Lead with the outcome in plain language, then name the technology immediately underneath as proof. A page that only speaks one of the two languages loses whichever reader it did not choose.
How do you name a category analysts do not track yet?
You write the definition before you write the name. A category name without an accompanying definition, a set of criteria and a named alternative it replaces is just a coined phrase. We produce the analyst-ready one-pager as a deliverable in its own right, because the people who decide whether your category exists read documents, not homepages.
Our founder is the brand. Is that a problem?
Not at all, and in deep tech it is frequently an asset: technical founders carry credibility no brand system can manufacture. It becomes a risk when the company cannot make a claim without the founder present. The work is to move the founder's judgement into the brand's voice, so the site argues the way they do when they are not in the room.
How do you write about technology that is under patent review or embargo?
By describing the effect rather than the mechanism, and by agreeing the line in advance with whoever owns the risk. We work routinely against disclosure calendars, and it is generally possible to make a claim legible without making it enabling. Where it is not, we say so rather than shipping something your counsel will pull two weeks after launch.
What should a deep tech homepage put above the fold?
What changes for the buyer, in their language, with the technology named immediately underneath as proof. Not the category, not the funding, not a rotating carousel of capabilities. The most common failure is a homepage that opens by explaining what the technology is, which asks a commercial reader to do the translation you were hired to do. Open on the outcome, then earn the detail.
Our site serves recruitment as well as sales. Can one site do both?
Yes, and in deep tech it usually must, because senior engineers read your site as a brochure for their next four years. The two audiences want the same thing from different angles: evidence the engineering is serious. A site built for a technically literate buyer generally recruits well by accident. A site built only for recruitment rarely sells.
How do we stop the site going stale between rounds?
By building the parts that change most often into a CMS a non-technical person can run, and by being honest about which parts will not be maintained. A site with three well-maintained sections beats one with twelve neglected ones. We hand over the project and train your team, so updates do not require a purchase order.
Do we need a separate site for our defence or dual-use work?
Sometimes, and it is worth deciding deliberately rather than by accident. Dual-use changes who may invest, who may buy and what may be published, and those constraints can pull the architecture in two directions. The usual answer is one site with a gated or separately-scoped section rather than two brands, but it depends on whether the two audiences would be embarrassed to meet.
Does a 3D-heavy site actually perform for SEO?
It can, when performance is budgeted at the start rather than fixed at the end. Draw-call limits, compressed textures, lazy scroll sequences and mid-range hardware fallbacks are design decisions, not post-launch optimisations. Transitry gained 58 percent on PageSpeed and PolyEnergetics 38 percent on Lighthouse SEO through builds carrying custom 3D and motion.
When is a film worth making, and when is it not?
It is worth it when the product cannot be understood from a screenshot, and when there is a specific moment it has to survive: a conference floor, a sales call, an investor meeting. It is not worth it when the positioning is unsettled, when nobody owns distribution, or when the honest answer is that a well-built page section would do the same job for a tenth of the budget.
Can you film inside a fab, clean room or secure facility?
Sometimes, and where access is impossible we build it instead. That is the practical reason so much deep tech work ends up in 3D: the interesting part is behind a door nobody is getting a camera through. Modelling from your CAD and spec avoids the access problem entirely and is often more accurate than footage would have been.
How do you keep technical claims in a film accurate?
Technical review before anything is rendered, because a correction after animation is expensive and after a shoot can be impossible. As Cloudphysician put it: if a heart rate was 120, it had to be 120, not 100. At a conference full of experts, one wrong detail costs the credibility of everything else in the film.
What does engineer-led actually change day to day?
The brief stops degrading. Ten of the studio are engineers by degree, working in strategy, 3D, delivery and the build, so the person reading your spec sheet has built to one. In practice that removes the glossary call, shortens technical review, and means the render is right about what the object does rather than only how it looks.
What if our engineers disagree with the copy?
They are usually right about the fact and usually wrong about the emphasis, and both are useful. Technical review is a scheduled part of the process rather than an obstacle to it. The rule we hold is that nothing ships that an engineer believes is untrue, and that accuracy is a constraint to design within rather than an excuse for a page nobody outside the company can read.
Do you work with companies outside India?
Yes. The studio is in Bengaluru and the work is worldwide, across defence, nuclear, semiconductors, biotech, robotics and climate. Deep tech categories are global by construction: the buyer for a counter-drone system or a reactor twin is rarely in the same country as the engineering.
How do you measure whether the brand worked?
Against what you said the problem was at the audit call. If the complaint was that procurement could not map you to their RFP language, the measure is whether that stops happening. Where the site is the deliverable there are hard numbers, such as PageSpeed and Lighthouse SEO gains. Where the deliverable is positioning, the honest measures are qualitative and we will say so rather than claiming credit for your pipeline.
Ask the one that is not here.
Thirty minutes, no deck. Bring the spec sheet and the date you are working towards. More on the vocabulary, film and who does the work.