SensibleSearch is the smartest, fastest and most secure Enterprise Search solution.
Based on meaning instead of keywords, it lets anyone find, process and connect information by voice or text, with perfect accuracy and typo tolerance. From any mix of structured and unstructured data, in virtually any format.
SensibleSearch is blockchain- and Web3-ready (any token, any smart contract) and can be deployed *on your infrastructure*. Your data and documents are secured throughout their lifecycle, with industrial strength, full scalability and 100% GDPR / CCPA compliance.
Data used to be a problem. Now it's a source of profit, performance and customer satisfaction.
Discover why adopting SensibleSearch is the most sensible decision you can make today...
While traditional, keywords-based solutions can't cope, SensibleSearch delivers field-tested solutions.
Time
Problem: You don't have time to prepare, normalize or tag data for enterprise search or applicative integration.
Solution: SensibleSearch understands data and documents without preparation or modification. Your time is saved, data integrity is preserved and information becomes immediately available.
Costs
Problem: 80% of data goes "dark", i.e unknown or unused. Loss of monetization compounds with ever increasing production, management and storage costs.
Solution: By making your data and documents easily consumable, SensibleSearch turns them into profitable, competitive assets. By delivering them effortlessly to their intended users, it inverts the absurd cost/benefit ratio.
Silo organization
Problem: People in your organization, not to mention users, have different data needs and requirements.
Solution: SensibleSearch unifies data and documents into a single knowledge base, adding public data if necessary. This aggregation bridges organizational gaps, fosters synergies and multiplies information value.
Access rights
Problem: You need to control who sees what, and when.
Solution: SensibleSearch connects to enterprise directories, user profiles APIs and implements access control by any custom categories.
Lifecycles
Problem: Your data and documents are dynamic by nature. Some are discarded while others get enriched or updated (very) frequently.
Solution: SensibleSearch can automatically access, digest and make new data available to users as it is produced.
Compliance
Problem: GDPR, DPA, CCPA... You have to ensure compliance. Your organization is legally bound and it's your professional responsibility.
Solution: Because it is compliant by design with privacy and other business-specific data regulations, SensibleSearch takes care of everything. It can also process data as it ingests them, thereby enforcing any kind of specific regulation on the fly, including pseudo/anonymization.
More solved pain points...
Compatibility / Interoperability
Problem: From legacy to mobile, your data is diverse in structure and formats. The more apps and sources (corporate, business, production, office, emails, social networks...) the bigger the problem.
Solution: SensibleSearch can ingest a large variety of data formats, structured or not. Reconciling technicalities restores consistency, which simplifies information management and allows valuable data connections.
Scalability
Problem: Your data and user base grow exponentially. Your current data applications don't scale up.
Solution: SensibleSearch is container-based so horizontal scalability is included by design -- in capacity and performance. Code is parallelized while transaction payloads are optimized for bandwidth saving, carbon footprint reduction and fast results delivery. As HPC Veterans, we know what scalability means.
Multilingualism
Problem: You have data and documents in several languages. Traditional data solutions are monolingual.
Solution: SensibleSearch's meaning-based technology is not language-constrained, contrary to keywords or hardcoded intents and entities. Answering questions on data in another language (or several of them) empowers users and maximizes information consumability.
Business specificity
Problem: Your data contains a lot of corporate and business terms, phrases and acronyms.
Solution: In addition to built-in machine-learning, SensibleSearch supports business and corporate vocabularies so that it is fully operational from day 1. Human moderation / supervision is included.
Custom processing
Problem: Your data requires in-house or external processing before it is read or consumed.
Solution: A SensibleSearch solution is always a pipeline of orchestrable workers assembled for your specific use case. Industry standard connectors allow custom plugins from any origin at any processing stage. These plugins can be local code packages or remote services. SensibleSearch doesn't have to know what the code does: sensitive black boxes are welcome.
Locality
Problem: Moving your data to the Cloud solves basic IT problems but increases data traffic, latency of results and loss of control.
Solution: SensibleSearch backends can reside on premises, in private, public or hybrid clouds... wherever is closest to your data and users. For better performance and lower network costs.
Security
Problem: Your data has value but there's ever more breaches and abuses. Storing it out of premises, making it available for mobile use and opening it for external processing only increase the risks.
Solution: SensibleSearch does not use or transport your data and documents. It reads them, understands them and uses only a vectorized digest, meaningful only to your own SensibleSearch application. This, in addition to secured network protocols and shielding from public access, guarantees total, state-of-the-art data security.
tips_and_updatesSee how Total S.A. uses it
In this video*, Total S.A. presents Smarty, its HSE (Health, Safety & Environment) assistant built entirely with LEXISTEMS SensibleSearch.
Smarty's mission is to facilitate access to the rich variety of HSE resources available to all employees within the Group.
Based on LEXISTEMS' Sensible.ai® technology, SensibleSearch works by meaning instead the competition's plain keywords. As a result, Smarty understands both users questions and the target multilingual documents, however complex. It therefore returns results that are way more pertinent and actionable compared to other benchmarked solutions.
The Smarty project was imagined and developed from A to Z in less than 6 months by a small team of HSE people at Total and data engineers at LEXISTEMS. None of the thousands of documents (Office files, PDFs, intranets...), some of them with hundreds of pages, has been modified in any way. Everything was ingested "as is" by SensibleSearch, regardless of format, structure and language.
* Subtitles available in 7 languages.
(White paper)Leveraging enterprise data with automatic Artificial Intelligence and Natural Language Processing...
In a recent use case, LEXISTEMS, NetApp and APY tackled a business problem typical of today’s document-rich organizations by automating AI-based data processing in natural language.
The key actors: LEXISTEMS’ SensibleSearch[1] solution and NetApp’s DataFabric AI-based storage combined in a global AI-bound solution designed and built by APY’s AI Lab.
The result: data and documents become a sustainable source of profit, performance and customer satisfaction - providing smarter competitive insights and actively helping users zero in on unrealized business opportunities.
[1] At the time of this white paper publishing, SensibleSearch® was named SensibleData®.
Problems solved - Enedis, France's main power grid operator, needed a search / data matching engine capable of answering technical questions and detailing corporate processes from a significant documents base (Office and PDF files, mainly). The SensibleSearch solution was delivered with two user interfaces : one is an interactive bot, the other is a more usual enterprise search engine except that it features a number of user-definable parameters (see the sidebar) to fine-tune results. Both implementations were operational as initially delivered, with backends and data residing on premises. Both implementations support vocal input so that users get results hands-free.
Data types - The source knowledge base is a repository of technical documents covering all aspects of Enedis' internal procedures including field interventions, hardware information, security guidelines and HR-related processes. None of the original documents had to be modified or otherwise prepared beforehand.
Enedis - Cinke
Problems solved - This is a particular SensibleSearch use case. Enedis, France's main power grid operator, wanted to be able to predict intervention times for their field technicians, in real time and on a 24/7 basis. To deliver the desired results with strategic precision, SensibleSearch matches a huge history of past incident reports with the specificities of the current intervention: site layout, types of equipments, failure locality, etc. The predictions are further refined with contextual conditions like traffic and weather data, served by SensibleSearch's ancillary public sources. A particular attention was given to data input: speech-enablement (via LEXISTEMS' SensibleSpeech) and AI-based autocompletion enable call centers' respondents to give time estimates immdiately upon alerts.
Data types - The data history used for predicting mainly consists of Excel files summing up incident reports. None of the original documents had to be modified or otherwise prepared beforehand.
Total - HSE
Problems solved - Total S.A., one of the six "Supermajor" oil companies in the world, has a very strong commitment to health, security and environment (HSE) matters. To raise awareness and share internal + regulated best practices among employees, they needed to facilitate access to their global treasury of HSE resources. Smarty, based solely on SensibleSearch, was the solution: a multi-purpose bot capable of answering any HSE-related question, text or speech, with direct access to reference documents. Smarty is shown here in its English user interface, with a question asked in English returning results regardless of the documents' language.
Data types - The available HSE resources include Office files, PDFs, corporate intranets and public websites contents, in several languages. The tables shown in the screenshot are rendered from a PDF table. None of the original documents had to be modified or otherwise prepared beforehand.
Société Générale - Compliance
Problems solved - Société Générale, a leading global 'universal' bank, faces complex compliance regulations on most of its operations. For SensibleSearch, the objective is simple: to help everyone find the right, actionable information as if they could ask a team of experts available 24/7 in all timezones. Since regulations may vary by countries or regions, this use case returns relevant results regardless of the language of the question and of the target documents. The screenshot shows results in English and French to a question in English, with SensibleSearch's result evaluation widget enabled.
Data types - The compliance resources included in this use case consist of an intranet with explicative pages leading to regulatory PDF documents, either public or of SocGen origin. None of the original documents had to be modified or otherwise prepared beforehand.
Service-Public.fr
Problems solved - Service-Public.fr is the global portal of the French public services. It exposes a comprehensive collection of information, resources, reference texts and official forms relating to legal matters for individuals, businesses and non-profits. At the request of a major insurance company, SensibleSearch was missioned to build a 24/7 legal Q&A system based on Service-Public.fr. The objective: to help the company save millions a year in level-1 customer assistance. In this application, SensibleSearch gives access to a wealth of knowledge via a natural-language interface (to be later integrated to the company's website and mobile apps). The screenshot shows a sample question and its final answer plus the offical references and a number of machine-learnt suggestions to help users find more on the current topic.
Data types - Service-Public.fr is an open data website. Its entire content is just read, daily and automatically, for direct availability in SensibleSearch.
tips_and_updatesSee a commented SensibleSearch story
This animation comments a demo of SensibleSearch on an open dataset (Wikidata).
Feel free to compare with your favorite search engine or digital assistant...
Start the animation
Based on LEXISTEMS' Sensible.ai technology, SensibleSearch queries data by meaning. Here is an example of a tricky question and a perfect answer returned (BTW: that's an excellent test for search engines and other personal assistants...). Each answer element has a clickable title which links to the relevant data in the target knowledge base (Wikidata here).
Sliding the answer of the previous screenshot... Note the "Idea" zone in the demo app that offers 1-click question buttons (for digging deeper on the current topic, on other topics or just for fun). The questions relate to Donald Trump, not his children. That's because...
...SensibleSearch's 1-click questions are in fact relevant to context. In this screenshot, we clicked on "place of birth", and got the corresponding answer. In addition to being very handy, this shows SensibleSearch's smart use of context to return results.
Sliding the answer of the previous screenshot...
This screenshot shows a simple question with an answer containing a full text and an image, both extracted by meaning from the target knowledge base. Note the "street" phrasing.
Scrolling down the previous screenshot, we see more contextual 1-click questions. These are related to the topic of the question (Bob Marley) rather than to the topic of the answer (Reggae). This adaptive process inherent to SensibleSearch can be configured easily. Let's click on "place of birth"...
Clicking on "place of birth" in the previous screenshot returned a compound answer (text + image). The context-related buttons return the same results as if the question had been asked "completely". If data is available in the target knowledge base, it will be aggregated according to API or user settings.
Another example of SensibleSearch' technological advance. The question might look silly but it is interesting from a data science standpoint as it is conceptual, poorly phrased and has misleading keywords. NB: SensibleSearch also offers links to more websites related to the topic (the star is for Donald Trump's official website).
Another example of answer returned from a SensibleSearch 1-click contextual question, with the aggregated elements and more context buttons.
SensibleSearch accepts colloquial entity names, even foreign.
Of course, SensibleSearch can return simple, factual answers to simple, factual questions. Three examples of data types here: one encyclopedic, one mathematic (intentionally simple), one geographic.
Using the 1-click context questions is an efficient way to dig deeper into something. Ok, regardless of the phrasing, most people know that William Shakespeare wrote Hamlet but...
Clicking on "characters" gives you the Dramatis Personae of the play in 1 click...
...as well as the narrative location...
...as well as more information about the Tragedy genre - and the list goes on. Is there any quicker and more transverse way to browse a topic?
Another example of a composite, listed answer on a simple question that most assistants and search engines won't answer directly.
Sliding the answer of the previous screenshot...
The API is permissive regarding questions phrasing. Note also that there are several famous people named "Bach". Here, SensibleSearch is configured to return only the most consulted result, but many other conditions or statistical scorings are applicable (recency, update, length, additional elements...).
Since SensibleSearch always include its "sources" in the results it returns, adding a verification or a sources consulting process right into SensibleSearch powered applications is just one parameter away.
Your SensibleSearch ROI
Enterprise search should be about maximizing data value and competitive edge.
From discovery to analysis, connection and access, SensibleSearch creates immediate value on the entire data value chain.
Simplicity
SensibleSearch understands your data and documents[1] without modification. That's a lot less time and efforts compared to the competition.
Connectivity
SensibleSearch connects your and public data on any number of general or business criteria.
Consumability
With meaning, text or speech, SensibleSearch makes data more accessible and consumable, regardless of the application.
Elasticity
Your data grow and change. SensibleSearch adapts and scales without disruption.
Security
SensibleSearch keeps your data in your datastores, using only a protected, vectorized version through secured network protocols.
Compliance
SensibleSearch enforces all Data and Privacy regulations, current and future.
[1] Restrictions may apply.
handymanTechnical requirements
SensibleSearch API
SensibleSearch is a green code, highly optimized API. So are its technical requirements.
Deployment - For maximum deployment flexibility, the SensibleSearch API and its use case-specific plugins are delivered as a fully-functional, self-sufficient Docker image with customizable volumes and shielding from network access. As such, SensibleSearch fits right in any orchestration / production environment, and scales infinitely with state-of-the-art security.
Monitoring - SensibleSearch supports all custom or standard platforms (Kibana, Graphana, Prometheus...) with async, non-blocking interactions. An optional dedicated monitoring platform is available, based on Kibana.
Hardware platforms - Depending on the specificities of your SensibleSearch API and the connected plugins, the servers to host it may or may not require a compute GPU. Other hardware specifications are of commodity grade (e.g. Intel Xeon E / AMD Epyc 7F32+, 32 GB RAM) unless concurrent transaction volumes or custom features require otherwise.
Please consult us for details.
SensibleSearch Server Appliances
SensibleSearch server appliances enclose SensibleSearch pipelines (API + use case-specific plugins) within a dedicated server. They come ready to plug and serve data and documents queries.
Custom configurations - SensibleSearch server appliances are typically built to order according to customers' production requirements, especially user base, estimated concurrent queries and optional data specificities.
Vendor agnostic - SensibleSearch server appliances can be designed from any mix of OEM sources. We favor cost-effectiveness, scale provisioning and energy efficiency with recommended architectures, but any setup is possible.
Please consult us for details.
We're here to help
Any questions about any of the above? Get in touch with us: a SensibleSearch specialist will give you all the answers you're looking for.
When you're ready, we'll setup and coordinate a call with the Alliance Member nearest you. No strings attached.
verifiedCertifications
Alternatively to customer's premises or private infrastructures, SensibleSearch APIs and server appliances can be deployed in select datacenters (of E.U. or regional, non-US jurisdiction) that guarantee compliance with the highest industry standards:
●
Datacenters tier class - Depending on the region, our infrastructure partners are of Tier IV level (or Tier III when unapplicable) according to the Uptime Institute's standard classification. Both classes require concurrent maintainability, with redundant distribution paths to serve the critical environment. Tier IV adds *full* fault tolerance to the Tier III topology. This results in typical downtimes no longer than about 30 minutes per year which, with redundancy, translates into virtually no downtime at all.
●
Information Security - In line with our commitment to total data security, our infrastructure partners are ISO 27001 cerfified. ISO 27001 specifies requirements for implementing, maintaining and continually improving information security management systems that protect data assets.
●
Energy efficiency - In line with our commitment to carbon footprint reduction, our infrastructure partners are ISO 50001 cerfified. ISO 50001 requires improving energy-related performance and energy efficiency continuously, as well as identifying and implementing energy reduction opportunities. ISO 50001 is modelled after the ISO 9001 Quality Management System and the ISO 14001 Environmental Management System; the 2018 version has clauses modular with both.
Availability - Pricing - Support
Availability
SensibleSearch is available worldwide through the
SENSIBLE Alliance.
Pricing
SensibleSearch's pricing plans depend on the type of implementation (API or server appliances).
They're available as subscriptions or one-off deals, and adjust to volumes, regions and functional specificities.
SensibleSearch incurs no hidden costs (cloud subscription, data egress, third party services...).
Support
Worldwide, world-class support is provided by the SENSIBLE Alliance regardless of the country of purchase.
Specific support needs escalate directly to LEXISTEMS' SensibleSearch engineering team.
Due to SensibleSearch's specific nature, don't hesitate to enquire about local or remote support on your applications, infrastructure and/or devices.
We're here to help
Any questions about any of the above? Get in touch with us: a SensibleSearch specialist will give you all the answers you're looking for.
When you're ready, we'll setup and coordinate a call with the Alliance Member nearest you. No strings attached.
Meet SensibleSearch
Your data assets have huge unrealized value.
In a one-to-one session or an interactive video call, see for yourself how SensibleSearch will benefit your business...
This animation comments a demo of SensibleSearch on an open dataset (Wikidata).
Feel free to compare with your favorite search engine or digital assistant...