A telecom operator can know almost everything about its customers and still miss the useful bit. The clicks, app opens, plan changes and browsing trails pile up by the billion. A dashboard can describe yesterday beautifully. But if a customer starts shopping for a new handset at breakfast and the relevant offer arrives on Friday, the data has become expensive archaeology. Intent HQ, a London software company founded in 2010, is built around that narrow, valuable gap between noticing and acting.
Its software studies streams of behavior and turns them into signals about what an individual may want, when that desire is strengthening and which action a business should take. The buyers are not small shops hoping to tune an email subject line. They are telecoms, banks, retailers, loyalty programs and consumer apps with millions of customers, complicated data estates and compliance teams that would prefer not to discover a privacy problem in production.
Intent HQ says its stack now manages more than 300 million customer profiles and processes about 250 billion events every day. It reports signal latency below 120 seconds. Those figures explain the company's natural habitat: enterprises where a tiny improvement, repeated across a vast customer base, can pay for the machinery.
The send button for customer data
The company has used a good analogy for the problem. Enterprise data without a path to activation is like an email without a send button. Intent HQ supplies the missing motion. Deep Signal cleans and organizes raw behavioral events. Intent AI looks for the shape of a decision forming. Privacy Twins create useful behavioral representations without exposing the person. Marketing Agents turn those signals into possible actions. IntentOne, launched in May 2026, provides the operating surface where teams can manage data, audiences, activation and governance.
From noise to a useful moment
This is not a replacement for every piece of marketing software. A customer data platform can unify records. A campaign tool can send a message. A lakehouse can hold and process enormous datasets. Intent HQ sits between those systems, selling interpretation and timing. Its current phrase for that layer is “Opportunity Intelligence” - ranked, decision-ready situations rather than another field in a customer table.
The distinction matters because the company's toughest competitor is often not a single vendor. It is the client's own stack: Salesforce or Adobe, a Databricks environment, homegrown models, a decision engine and enough analysts to keep the assembly moving. Intent HQ's case is that the connective tissue, behavioral models and privacy engineering are difficult enough to buy rather than build.
Privacy moves from promise to plumbing
Personalization has an image problem. The more uncannily accurate an offer becomes, the more likely its recipient is to wonder who was watching. Intent HQ's answer is Edge AI: some inference happens on the customer's device, where raw personal signals can remain. The company says its architecture can classify thousands of behavioral interests locally and send out derived, privacy-safe intelligence instead of the underlying trail.
“Intelligence without privacy is surveillance.”Intent HQ's technical credo
That is the sharpest part of its positioning, and the part most worth interrogating. Privacy cannot rest on a slogan. It depends on what an implementation collects, which permissions exist, what leaves the device, how profiles can be joined, how long data remains and whether a person can opt out. Intent HQ points to SOC 2 Type II, ISO 27001 and ISO 27018 certifications, and says 18 of its 49 granted patents concern privacy-preserving AI. The architecture offers stronger tools; the client still has to use them properly.
The edge capability came through a strategic change. In November 2023, Intent HQ merged with Anagog, an Israeli on-device AI company. Intent HQ already knew how to turn immense enterprise datasets into customer intelligence. Anagog supplied the missing vantage point: behavior understood locally, close to the moment it occurs. The merger moved the proposition from cloud analytics with privacy controls toward a split architecture spanning device and cloud.
The first product did not look like this
Intent HQ began in a looser, more permissive era of the web. When Oxford Capital led a £4.75 million investment in 2014, the product was described as creating a “fingerprint” of a website visitor from social media data and online behavior. Publishers and consumer brands could tailor content and advertising. Named commercial relationships included The Telegraph, Sky Italia, Food Network and IDG UK.
The underlying ambition - understand interests well enough to improve relevance - survived. The methods and market language changed. Regulation tightened. Browser tracking became less dependable. Consumers became more suspicious. Telecom operators and other data-rich businesses wanted the upside of individual prediction without casually moving sensitive data around. Intent HQ responded with privacy engineering, on-device models and a much heavier enterprise operating layer. What changed its mind was not one public failure; it was a market that made the old bargain steadily less workable.
The cost is easier to see at company level than customer level. Intent HQ has no public price list. Its terms allow recurring subscriptions, usage fees, one-time charges and professional services, which is the vocabulary of a negotiated enterprise contract. The supplied private-company data puts total funding near $21 million, though public databases disagree. The one well-documented early round is the £4.75 million financing in 2014. In other words, this is bought like infrastructure, not downloaded like a team app.
Proof arrives in dollars, not adjectives
The company's most effective sales document is a case study about an unnamed Tier One North American telecom operator. The client did not lack data or analytical talent. It lacked a fast, shared route from insight to action. Intent HQ says the 90-day program produced more than $40 million in incremental revenue: roughly $20 million from real-time engagement optimization and more than $20 million from a personalized mobile-plan experience. A new cross-sell strategy returned three times its cost.
Reported incremental revenue for a Tier One North American telecom operator. The lesson is less “collect everything” than “remove the days between insight and action.”
Other examples are smaller but more legible. Verizon reported a 51 percent uplift in marketing return on investment in an audience program and, in a separate churn project, an improvement of up to 3.5 percent in prediction accuracy with a forecast value of $180 million over five years. REV, the Cable Bahamas operator, used behavioral insight to anticipate churn and decide which television content customers actually wanted; Intent HQ says the content decision saved more than $1 million. Property operator HB Reavis tested the best moment to contact shopping-center visitors and recorded 287 percent higher click-through rates for a selected campaign.
These are vendor case studies, not universal constants. Their useful feature is specificity: a control, a measured intervention, an economic result. Intent HQ's Human Guides - data scientists, customer strategists and delivery leads - help choose those use cases. That services layer is not an embarrassment hidden behind the software. At enterprise scale, deciding what to model and getting legal, data and marketing teams to act together is part of the product.
Where the machine fits - and where it stalls
Intent HQ is best suited to a company with a huge customer base, frequent digital behavior, valuable repeat decisions and enough activation channels to do something immediately. Telecom is the obvious proving ground because churn, upgrades, plans, devices and service problems all leave signals. Banking, insurance, loyalty, retail and large consumer apps share parts of that pattern. The company has partnerships with Databricks for technical distribution, Seed Group for entry into the Middle East and Orange for a Spanish market-insight product built on anonymized profiles.
The fit weakens when customer events are sparse, purchase cycles are very long, the reachable audience is small or the organization cannot act in real time. A model that correctly spots intent is wasted if campaign approval takes three weeks. On-device intelligence also needs distribution through an app or SDK, user permission and enough local behavior to interpret. A business without those conditions may get more value from better data hygiene, simpler segmentation or basic lifecycle automation.
The copyable part
Do not begin with “Where can we use AI?” Begin with the customer decision your company notices too late. Set a commercial measure, preserve a control group, define the privacy boundary and shorten the interval between signal and action. The patented machinery is Intent HQ's. The discipline is available to anyone.
The company itself has had to keep shortening that interval. Revenue grew from a reported €2.4 million in 2019 to €12.3 million in 2022, helping Intent HQ appear in the Financial Times FT1000 for three consecutive years. It merged to acquire an edge capability, joined the Databricks ecosystem, expanded through Dubai and collected a run of British analytics and AI awards in 2024 and 2025. Its team is now roughly 120 people by private-company estimates, spread across London, New York, Barcelona, Lisbon, Tel Aviv and other regional posts.
IntentOne is the latest attempt to make all that technology feel like one product. Its challenge is familiar: enterprise buyers already own many platforms that promised a single view of the customer. Intent HQ has a more pointed claim. The view is not valuable because it is single. It is valuable because it is current, explainable, governed and attached to a decision. If the company can keep proving those four things, it is selling something rarer than another AI dashboard: useful timing.