Mobile internet connects more people than ever to critical services such as healthcare, education, e-commerce and financial services and provides income-generating opportunities. Yet the benefits of connectivity are not being realised equally.
There are two ways people can be 'unconnected'; either they live in an area not covered by mobile internet broadband, or they live in an area that is covered, but they do not use mobile internet.
Some 39 per cent of the global population lives within mobile broadband coverage but are not using it, while 4 per cent of the world's population (around 350 million people) live in areas without mobile broadband coverage and, thus, mobile internet access.
As AI adoption accelerates worldwide, one major group is being left behind - those without reliable internet. teks is solving this by pioneering AI messaging at the lowest connectivity levels, expanding into lower network infrastructures such as LoRa and integrating verification layers for AI-driven agents.
An SMS-driven solution
Wanting to create a solution now, rather than waiting for bureaucrats, is teks. This German-Fillipino startup has developed SMS-based technology that allows access to any AI model and provides easy information retrieval through simple text messaging.
By transforming text messages, a basic feature with 92 per cent global coverage, into a powerful tool for internet navigation and content delivery, teks addresses the critical challenge of digital exclusion in areas where advanced connectivity infrastructure — like 3G, 4G, or broadband — is scarce or non-existent.
The startup aims to connect 3.8 billion unconnected people worldwide.
I spoke to two of teks' founders, Carlo Baltazar Canda (CTO) and Livia Dolle (CEO)
The duo previously founded takle and exited in early 2024.
Canda is based in the Philippines, specifically Mindanao, in the south.
He shared:
"The idea for this project started last year, during the peak of ChatGPT's popularity.
My rice farmer uncle asked me about AI and whether he could access it. He wanted to check weather forecasts and farming techniques, but internet access is limited in rural areas, and not everyone has a smartphone.
So we wondered — how can we bring AI to these areas?”
The challenge of life without connectivity
In the example of farming, while government websites share information about things like farming and agriculture, they are impossible to load with low connectivity.
“Governments build massive web portals for their citizens, but they don’t realize people can’t even open them,” explains Dolle.
"Government websites are impossible to load on low-bandwidth connections. And even if a farmer reaches a website, navigating it without digital literacy is another challenge."
Further, plantations are typically owned or subsidised by large conglomerates that supply seeds to the farmers. However, if farmers experience problems, they have limited means of accessing information.
Dolle detailed, "Usually, one farmer gets the answer to a question and passes it on via word of mouth. This is hugely problematic because problems in the field can greatly diminish a farmer's harvest and income."
What is the cause of the lack of internet connectivity?
According to a global report by GSMA, a worldwide industry organisation that represents the interests of mobile network operators, the growth rate at which people are adopting mobile internet has remained flat, and significant digital divides persist.
Canda detailed:
"I've been in the engineering industry before with an electrical engineering background, including a company responsible for setting up internet connections in these rural areas.
The process takes a long, long time."
According to Canda, the government is responsible for infrastructure, but implementation is slow:
"Projects take years due to procurement and bureaucracy. Even when the internet reaches rural areas, people often can't afford to use it."
Dolle elaborated:
"The internet penetration rate here is just 1.8 per cent per year - it's painfully slow. Satellite internet options like Starlink exist, but remain unaffordable for most users. Many NGOs try to provide connectivity, but they fund it for a year and leave, making it unsustainable."
Worse, even if the internet infrastructure exists, an accessibility gap exists due to economic capacity and how-to knowledge.
The 2G text-based solution
By transforming SMS into an interface for AI models, teks removes the need for mobile data, apps, or even smartphones.
teks uses 2G SMS infrastructure to transform internet-based tokens into 2G-compatible text. It also enhances prompts and trim responses to fit within text message limits.
Canda notes:
"Many people here don't know how to use Google but know how to send a text. AI can take natural language queries, process them, and send back a response via SMS."
How teks works: expanding AI access on the weakest networks
teks enables AI-powered natural language interactions over SMS, eliminating the need for users to learn complex search techniques:
-
Text-based query processing converts natural language queries into 2G-compatible messages.
-
Network adaptation adjusts response delivery to SMS, USSD, and LoRa-based messaging, ensuring compatibility across all low-connectivity networks.
-
Agent verification integration means that AI interactions are expanded to include verified knowledge agents that ensure information reliability.
-
Response compression and optimisation restructure AI-generated responses into SMS-sized, readable formats for seamless delivery.
This structured pipeline removes the need for traditional web-based AI interactions, making it possible to access AI anywhere with just a text message.
teks has already gained traction purely through word-of-mouth, now handling over 150 daily AI queries via SMS.
Students are one of the biggest direct beneficiaries of teks — can you imagine being a student in 2025 without reliable internet connectivity?
According to Canda:
"A big chunk of our users are students without internet access. Previously, if they wanted to do internet research for assignments or school projects, they would have to travel to the city centre to an internet cafe.
But they have access via text, and that's what motivates us every day."
The opportunity for AI-messaging expansion
Upgrading from 2G to more advanced networks is a slow and costly, particularly in less developed areas. Many regions might not see significant infrastructure improvements for years, making TEKS a temporary solution and a long-term necessity.
The combined market opportunity for telehealth, digital agriculture, and remote education is projected to reach $1 trillion by 2029.
Unlike many "AI for Good" initiatives, teks does not approach this as a philanthropic project - it sees AI accessibility as a global requirement.
As Dolle notes:
"The data from people without internet access is valuable. It can help refine AI models for low-connectivity environments, and eventually, we'll monetise it responsibly."
While teks started as an SMS-based AI access layer, the platform is expanding into LoRa and other ultra-low bandwidth networks, making AI even more widely available.
The company is currently focused on developing verification layers for AI-powered agents, expanding AI access beyond SMS to LoRa and ultra-low bandwidth networks, and scaling AI messaging to global markets beyond just rural areas.
Would you like to write the first comment?
Login to post comments