Short answer: A key card costs more than its purchase price suggests once printing, replacement, admin time and the security gaps of lost or lent cards are counted. Facial recognition avoids most of those costs for everyone who enrolls. Edge.AI by Proptech.AI pairs face access with app-less QR codes and RFID fallbacks, a cost-effective mix that keeps cards only where they earn their place.
Key takeaways
- A card’s purchase price is only the first cost; printing, replacement, admin time and the risk of lost, lent or copied cards follow.
- Facial recognition has no physical credential to buy, print, replace or collect, and app-less self-enrollment moves setup to the user.
- Cards still belong in the mix as a fallback, and Edge.AI supports RFID cards alongside faces and QR codes.
- Integrating via Wiegand lets faces and existing cards run on the same panels, the approach used at 101 Greenwich.
- Early in the year, with last year’s card orders still fresh, is a good time to total your credential spending.
What does a key card really cost over its life?
The card itself is the cheapest part. Every credential also carries printing or encoding, a chance of being lost, and admin time each time it is issued, replaced or switched off.
Buying and replacing cards
Blank cards or fobs, badge printing, supplies and encoding all cost money, and someone has to keep stock on hand for the next move-in or new-hire class. Cards crack, fobs fall off key rings and both get left at home, so each replacement repeats the purchase and issuing time or becomes a manual override at the desk.
Issuing, deactivating and recovering
Someone has to enroll each person, activate the card and, when they leave, deactivate it and try to get it back. In offices that means new-hire classes and departures; in apartment buildings it means fob handovers at every move-in and move-out.
Security gaps that stay off the invoice
A card proves that a card was presented, not who presented it. Lost cards stay active until someone reports them, cards get lent to friends or coworkers, and many older card formats can be copied. Each gap is a risk cost that only surfaces when something goes wrong.
How does facial recognition change the cost per person?
It removes the physical credential for everyone who enrolls. With Edge.AI facial recognition access control, there is nothing to buy, print, ship, replace or recover for those users, and the person does most of the setup.
Enrollment needs no app: an email arrives, the biometric privacy consent comes first, a selfie gets uploaded, and access starts with the person’s next arrival. Proptech.AI specifies 99.9% accuracy from the device’s multiple sensors, and spoofing or fraud attempts trigger alerts, which makes a borrowed or copied credential much harder to use.
| Cost line | Plastic cards and fobs | Facial recognition on Edge.AI | QR code or RFID fallback |
|---|---|---|---|
| Credential purchase | Paid per person, and again for every replacement | No physical credential | QR: nothing to buy; RFID: paid per card, for a smaller group |
| Replacement | Lost, broken or forgotten cards reissued | Nothing to lose or leave at home | QR: nothing physical to lose; RFID: replaced as today |
| Issuing time | Staff enroll, print and activate | App-less self-enrollment by email | Light admin for the remaining card holders |
| Offboarding | Deactivate and try to recover the card | Remove access in the console, nothing to collect | Deactivate, and recover RFID cards where possible |
| Lending and copying | Cards can be lent, and older formats copied | The credential is the enrolled person, backed by spoofing alerts | Similar lending risk to cards, so set clear rules |
What do credentials cost a 600-person office building each year?
The cards plus the time to manage them, and in this example the time costs more.
Illustrative example: assume an office building with 600 badge holders. These are assumptions for illustration, not benchmarks.
- New cards: assume 25% annual turnover. 600 x 25% = 150 cards.
- Replacements: assume 12% of holders lose or break a card each year. 600 x 12% = 72 cards.
- Card cost: assume $8 per printed card, including supplies. (150 + 72) x $8 = 222 x $8 = $1,776.
- Admin: assume 20 minutes of admin per card, covering issuing it and switching off the old one, at $35 per hour. 222 x 20 minutes = 4,440 minutes, or 74 hours. 74 x $35 = $2,590.
Total: $1,776 + $2,590 = $4,366 per year, or about $19.67 per card event ($8 + 20/60 x $35). That excludes the cost of any incident involving a lost or lent card. Apply the per-event cost to your own card volumes and line the result up against a subscription quote from Proptech.AI.
These costs renew with every arrival, departure and lost card, which is what makes AI access control the most cost-effective proptech upgrade for credential-heavy buildings. For towers with thousands of badge holders, see the guide to access control for commercial real estate.
When do key cards still make sense?
As a fallback rather than the default. A sound proptech credential strategy gives every person a way in that suits them, and cards still serve several groups well:
- People who opt out of facial recognition and prefer something in their pocket.
- Short-term contractors, although a QR code on their phone may be simpler.
- Doors not yet upgraded during a phased rollout.
- Locations where local rules restrict facial recognition.
Because faces, QR codes and RFID cards all run on the same Edge.AI platform, nobody is forced into one method. For the wider trade-offs, see the comparison of facial recognition, mobile credentials and key fobs.
How do you move from cards to faces without new panels?
Start integrated. Wire Edge.AI into the panels you already have via Wiegand: the old system keeps running, today’s cards still open doors, and faces, QR codes and analytics come on top.
You could also pilot it in parallel, mix integrated and standalone doors, or move everything onto Edge.AI and switch the old platform off. Cloud and on-prem hosting are both options, and the retrofit guide for adding facial recognition to existing systems weighs each route.
That is how 101 Greenwich, a JLL-managed New York City building of 480k RSF over 25 floors, home to 1,500 tenants, made the move. Before the upgrade, people were counted by hand and the access control system was limited; Proptech.AI then fitted its six turnstiles with six Edge.AI sensors tied into the existing card-based system, which gave the building 24/7 secure access.
What privacy steps come with switching from cards to faces?
A face is personal data in a way a card number is not, so treat enrollment as a consent process. Edge.AI puts biometric privacy consent first, offers QR and RFID alternatives to anyone who declines, deletes data upon request, and keeps PII separate from both the device and the management console.
Write a retention policy, post signage where local rules require it, and check the laws that apply to each site. Under GDPR Article 9, biometric data that uniquely identifies someone falls into the special categories of personal data, and Texas regulates the capture or use of biometric identifiers in Business and Commerce Code chapter 503. Proptech.AI designs Edge.AI to help you comply with these frameworks. This is general information, not legal advice.
Frequently asked questions
Facial recognition avoids most of the recurring costs that key cards carry: buying and printing cards, replacing lost ones, staff time to issue and deactivate them, and the security risk of lost or lent cards. Whether the total is lower for a given building depends on its volumes, so compare your annual card costs with a subscription quote for Edge.AI by Proptech.AI, which supports faces, QR codes and RFID cards together.
Key cards and fobs carry costs beyond their purchase price: badge printing and encoding, replacements for lost or broken cards, admin time to issue and deactivate every card, cards that are not returned when people leave, and security gaps when lost cards stay active or cards are lent or copied. These costs recur every year, which makes the true cost of access credentials easy to underestimate.
Yes. Edge.AI by Proptech.AI plugs into existing access control panels through Wiegand, which lets current cards carry on working while faces and QR codes are added. The 101 Greenwich building in New York City, for example, was fitted with six Edge.AI sensors on six turnstiles, integrated with its existing card-based system. Edge.AI also supports RFID cards directly when it runs as a standalone access control system.
Key cards remain useful as a fallback: for people who opt out of facial recognition, for doors that have not yet been upgraded during a phased rollout, and for sites where local rules restrict facial recognition. Short-term contractors and visitors can use a QR code on their phone instead. Edge.AI by Proptech.AI supports faces, QR codes and RFID cards, so each person can use the method they prefer.
Add the purchase and printing cost of one card to the admin time needed to issue it and deactivate the card it replaces. For example, if a printed card costs $8 and admin takes 20 minutes at $35 per hour, each card event costs about $19.67. Multiply that figure by your yearly new and replacement cards, then compare the total with a subscription quote for Edge.AI from Proptech.AI.
Price your cards against a quote
Run the credential math for your property, then ask Proptech.AI for an Edge.AI subscription quote or a demo of face, QR and RFID access side by side. Send questions through the contact form at the bottom of this page or by email to sales@proptech.ai.
