For roughly a century, the corporate press review performed a dependable trick. It gathered the morning's clippings, placed the important ones near the top and delivered proof that the world had, once again, mentioned the company. The ritual survived radio, television, the internet and social media. Then somebody at Deutsche Telekom asked an impolite question: what if the press review did more than report that the press existed?
The answer became the COM Morning Briefing. It opened with context, not just links. It tried to tell employees what the day's media picture meant and why it deserved attention. The idea met the usual resistance. A press review is a press review, skeptics said. An executive asked to write the opening commentary reportedly answered, in effect, forget it. He slept on it. The team tested dummy editions. On February 28, 2025, the new version went live. In 2026, the briefing received a Deutscher PR-Preis.
AITASTIC helped build and operate that briefing. The episode explains the company better than a feature matrix could. AITASTIC is not trying merely to find every reference to a brand. Its proposition is that the pile of references should become an instrument: one that helps a communications chief detect a risk, measure whether a message traveled, or understand what an audience actually heard.
“We help you to find trends you were not looking for.”AITASTIC's description of its analysis service
The machine is called Harry
AITASTIC was founded in 2021 by Bernd Hitzemann, Dr. Rainer Mathes and Holger Oftring, three veterans of media analysis, communications research and technology. They named their core system Lasswell AI after Harold Lasswell, the political scientist whose famous communications formula asks who says what, through which channel, to whom and with what effect. Around the office, the system is simply Harry.
Harry scans paid, earned, shared and owned content, plus traditional media. The company says it works across text, images and audiovisual material in more than 50 languages and 70 markets. Natural-language processing and knowledge graphs turn unstructured coverage into topics, entities and relationships. From there, AITASTIC can measure familiar things such as sentiment and share of voice, as well as message transfer, salience, agenda setting and what it calls Semantic DNA matching.
The software is only half of the offer. The AITASTIC App gives teams a shared portal, editorial workflow, permissions and distribution tools. Analysts and consultants then configure measurement programs, curate daily or weekly briefings and interpret the results. This places the company between a software subscription and a research consultancy. Clients are not simply buying a login. They are buying a maintained way of seeing.
What Volkswagen wanted to know was not on the dashboard
At the 2025 IAA mobility show, Volkswagen introduced a naming strategy around the ID. Polo. A traditional monitoring job could count the articles, score their tone and rank the outlets. Useful, but incomplete. AITASTIC's media analysis was combined with daily audience-perception research from Civey in six core European markets. The joined data showed both the output of the announcement and its outcome: how strongly the story appeared, how positively it was discussed, who noticed it and how the relevant audiences evaluated the change.
Coverage is what left the building. Perception is what arrived.
- Media analysis measured resonance and tone around the naming news.
- Live polling measured reach and reaction among target audiences.
- Volkswagen's communications team received the findings while the event was still happening.
That is the difference AITASTIC wants to own. Meltwater, Cision, Brandwatch, Talkwalker, Signal AI, Onclusive and UNICEPTA all occupy parts of the same broad market. Several offer enormous source coverage and capable AI. AITASTIC's argument is narrower: proprietary analysis, an editorial workflow and experienced consultants should operate as one system. The advantage is strongest when the question is bespoke, multilingual and important enough that a false positive can reach the boardroom.
The price of trustworthy reading
AITASTIC does not publish customer prices. That is normal for a tailored enterprise product, though inconvenient for a buyer hoping to compare plans before lunch. Contracts appear to blend platform access, continuous monitoring, custom analysis, editorial briefings and consulting. The bill will depend on markets, languages, sources, users, service levels and the amount of human work involved.
One disclosed cost sits upstream. In September 2026, AITASTIC announced a long-term, multi-million-euro licensing agreement with PMG, Germany's press-licensing organization. The deal permits AI analysis of licensed journalistic content from thousands of publications. The exact figure was not released. The direction was clear: if an enterprise wants answers grounded in professional journalism, the right to process that journalism is part of the infrastructure, not a footnote.
That move also reveals what could fail first in AI media intelligence. It is not necessarily the model. It is trust in the input: missing rights, weak sources, bad taxonomies, or a generic prompt that ignores how the client defines risk. AITASTIC's human-in-the-loop promise is designed for precisely this weak seam. A consultant can recognize that a sarcastic headline is not positive sentiment, or that a small trade publication matters more to a regulator than a viral post.
A playbook worth borrowing
A team does not need AITASTIC's infrastructure to copy the central discipline. The lesson from the Telekom and Volkswagen work is to design backward from a decision. Before monitoring anything, name the choice the data should improve. Then separate exposure from effect, let machines compress the repetitive work and assign a person to challenge the conclusion.
- Start with a decision, not a keyword list. “Should we change the launch message?” is more useful than “Track our brand.”
- Measure arrival as well as output. Coverage can be loud while audience understanding remains faint.
- Write an editorial layer. A short explanation of why the morning's signals matter will beat another unranked feed.
- Keep a human veto. Automate collection and first-pass classification, but make someone accountable for context.
- Secure the inputs. Source quality and usage rights determine how confidently the result can travel inside an organization.
Where the fit gets awkward
Strong fit
A global enterprise with several markets, high coverage volume, executive briefings, sensitive reputation questions and a need for tailored measurement.
Weak fit
A small team in one market that wants instant self-service setup, transparent monthly pricing and simple mention alerts without consulting.
The model also depends on client participation. If leaders cannot agree on the question, the audience or what success means, more sophisticated analysis will produce a more sophisticated argument. And if speed is the only requirement, a lightweight alerting tool may be cheaper and perfectly adequate. AITASTIC makes most sense when interpretation has economic value and when somebody inside the client organization is prepared to act on it.
Four years after its founding, AITASTIC says it works with more than 50 Fortune 500 companies and almost 40 percent of DAX companies. Its public footprint spans New York, Mainz, Ann Arbor, London, New Delhi, Tokyo and Shanghai. Its automotive ties remain visible through Volkswagen, Porsche, the World Car Awards and German Car of the Year. Yet the most revealing artifact is still the humble morning briefing. It arrives at the edge of the workday and tries to perform one valuable conversion: from everything that happened to the few things worth knowing.