We offer a free 30-minute ideation workshop to help you get started. You don’t need to have the perfect question upfront—AI makes it easy to explore broad directions and refine them as we go. Whether you're curious about the future of energy storage, materials innovation, or emerging disruptions, that's already specific enough for us to begin.

No. All we need is your question. We can support you in refining it during the initial workshop. You don’t need to share proprietary data to get value from our platform.

No. We build AI systems that work with human experts, not instead of them. AI handles the grunt work—like processing and analyzing massive datasets—so you can focus on strategic decisions. That said, experts who do use AI tools will move faster and outperform those who don’t.

We use LLMs too—but we’ve built a platform that optimizes them specifically for technology scouting and evaluation. Think of it like the difference between an off-the-rack suit and a tailored one. Plus, ChatGPT can’t forecast. We can.

You get major efficiency gains. The platform lets you scout, compare, and landscape technologies fast. It summarizes patents, supports deep dives, and lets you interact with the data directly to answer complex questions.

We use global patent data. While individual patents aren't predictive, the system as a whole forms a massive, structured dataset of technological progress. The hidden metrics in the patent citation network help us forecast which technologies are improving fastest—and those always win.

Not really. While there’s up to an 18-month delay between filing and publication, most (55%) publish within a year. More importantly, trends in patent data show up years—sometimes decades—before the general public catches on.

Because they measure popularity, not performance. Popular tech isn’t always good tech. Our analysis focuses on improvement rates hidden in patent citations—this is what lets us spot winning technologies before they take off.

Not yet. Scientific publications are harder to standardize and access. We plan to expand in this area, but for now, patents give us more than enough lead time for forecasting disruptive change.

We use LLMs to surface relevant technologies, even from large or unfamiliar spaces. You can start with a broad question, and our system identifies what’s emerging, improving, or being adopted across domains.

Yes. We use TIR (Technology Improvement Rate) metrics derived from the patent citation network to compare how fast different technologies are improving. This is core to how we forecast.

TIR measures how quickly a technology is getting better, based on citation patterns within patent networks. Faster-improving tech is more likely to dominate in the future.

We map out the ecosystem around a particular technology—who’s working on it, what subdomains exist, where the activity is happening, and how fast things are moving.

No. We take data security seriously:

For chat features, we use OpenAI's Enterprise API—inputs are never used for training and are only stored for 30 days for security screening.

All calls to LLMs are routed through our own API, anonymized and randomized. Chats are encrypted at rest and only accessible by the creator.

Forecasting and search features run entirely within the GetFocus infrastructure—no external APIs involved.

Our contracts include a 5-year confidentiality clause, even after termination.

No problem. Our Vector Search and Forecasting tools don’t use LLMs, so you can still work with us while staying compliant. You’ll still get valuable insights, just without the enhanced chat features.

If absolutely no patents exist in your area, our approach won’t help. But that’s extremely rare. We haven’t encountered a project yet where we couldn’t extract relevant insights—even from lightly-patented domains.

When an entire domain of emerging technology has zero patent activity, we cannot measure a signal. However, in practice, this virtually never happens. Different companies will make different choices on whether to patent or not, but generally speaking, many of them will choose to patent. We do not need all innovation to be captured in patents in order to measure a reliable improvement rate signal, we just need a representative sample. Similar to when we want to know the average height of people in a city, as long as we can measure a representative sample, it doesn’t matter if some of the residents are hiding that day.

Because the pace of innovation is accelerating. More technologies are emerging than ever before, and your bandwidth to keep up hasn’t changed. Our system helps you cut through the noise, avoid blind spots, and make sense of what’s coming—before it arrives.

The model that estimates technology improvement rates has zero generative AI in it, it is a completely deterministic, black and white, and fully explainable model based on the citation network that underlies global patent data. Therefore, there is no risk of hallucination when it comes to our improvement rates.