Artificial intelligence is no longer a future consideration for quick service restaurants. It is here,
and the pressure to adopt it is real. But according to Brooks Thompson, Director of Emerging
Technology and Partnerships at Telaid, the harder question is not whether to adopt AI. It is
knowing which AI is actually real and which is just noise.
Where AI Is Gaining Real Traction
QSR has always adopted new technology, Thompson notes, but it has rarely led the way.
That is starting to shift as AI computer vision and generative learning, already proven in other
industries, find their footing in restaurants. The technologies gaining the most traction are the
ones touching the customer directly: chatbots, drive-thru communication tools, and
personalized suggestions. Large language models, by contrast, are still mostly working in
digital channels rather than in-store.
On the operations side, Thompson points to production planning as an area with real upside:
understanding what is in the prep area, predicting drive-thru volume, and balancing speed of
service against waste. He also highlights simpler technology making a difference, like
systems that verify a bag's weight against the order to catch missing items before they reach
the customer.
The Gap Between the Lab and the Store
A consistent challenge is the gap between what works in a lab with abundant resources and
what works in an actual restaurant with limited bandwidth and infrastructure. A solution built
for a 100,000 square foot facility does not automatically translate to a QSR location. That gap
is part of why cloud-based architecture, decoupling data and shifting compute elsewhere, is
becoming a bigger part of the conversation.
Voice: The Use Case Everyone Is Watching
Thompson sees voice technology as the one poised for a breakout moment. Early chatbot
and AI ordering experiences generated excitement, but adoption has been cautious since
brands do not want to risk the customer experience, and most people have had a frustrating
call center encounter that makes the risk obvious. Once voice is figured out at scale,
Thompson expects it to spread quickly, unlocking upselling and more personalized
interactions. The technology and the appetite are there. The execution is not quite ready yet.
Choosing the Right Partners
Thompson's first piece of advice for any QSR company considering AI is to focus on
partnerships before anything else. Who is helping deliver the system, and does the
architecture actually hold up? It is easy to get excited about use cases without understanding
the infrastructure required, only to discover later that cost or model limitations make the plan
unworkable. An experienced partner brings outside perspective that helps validate pain points
before they become expensive surprises.
A few red flags are worth watching for. Any partner who claims they can do everything without
acknowledging real challenges is an immediate warning sign. So is a partner who never
mentions training or customization, since every AI model needs to be trained for its specific
environment, and that takes real time.
Start With the Problem, Not the Technology
Thompson's most counterintuitive advice is to flip the usual approach. Rather than starting
with the technology and searching for a place to apply it, start with the actual pain points in the
business and determine whether they are worth solving. Validating how real and how costly a
problem actually is tends to surface the right AI use case naturally.
Integration Is the Dream, Rarely the Starting Point
Full integration with existing IT systems is the ideal outcome, but Thompson sees two
common paths there. Some organizations try to integrate everything from day one, which is
often slow due to technical debt. Others prove the technology independently first, then face a
long road toward broader adoption, which tends to generate more frustration once leadership
is invested and asking why it is not yet delivering results.
The better outcome is a blend of the two: proving the concept while staying engaged with the
teams who will eventually support it. To bridge that gap, Telaid runs workshops that bring in
the project owner alongside every supporting team, operations, marketing, and IT, so
everyone understands the plan early and can build it into their own roadmaps.
Real AI vs. Marketing Buzzwords
With "AI" attached to nearly every product on the market, Thompson offers a simple test. If a
tool just hands you data points and leaves the decisions to you, that is analytics, not AI. True
AI gathers that data and produces recommendations or makes decisions on its own, whether
at the edge or in the cloud, depending on the speed required.
Where Deployments Actually Break Down
Most AI projects do not fail because the technology stops working at scale. They fail in the
decisions and execution that happen before deployment. A solution might work perfectly in its
own controlled environment, but once it starts interacting with other teams and systems,
internal politics and miscommunication often stall it out.
Thompson is clear that most failures are not really about the technology. Projects typically
have proper metrics and real-world testing behind them. What breaks down is communication
between teams and a failure to connect the technology to genuine value for the people using
it. His closing advice: getting stakeholders to sign off is not the same as getting real buy-in.
Without genuine change management at the store level, even a well-designed solution will
struggle to deliver the adoption and ROI it was built for.
Ready to take the next step? Contact us below to see how Telaid's Solution Group can partner on your next project.