Reliant AI

Generative AI software for extracting, structuring, and analyzing scientific and regulatory information for biopharma.

Website: https://www.reliant.ai/

Cover Block

Open sources

Field Value
Name Reliant AI
Tagline Generative AI software for extracting, structuring, and analyzing scientific and regulatory information for biopharma.
Headquarters Montréal, Canada
Founded 2023
Stage Exited
Business Model SaaS
Industry Deeptech
Technology AI / Machine Learning
Geography North America
Growth Profile Venture Scale
Founding Team Co-Founders (3+)
Funding Label Seed
Total Disclosed ~$13,500,000 [BetaKit, August 2024]

Links

Open sources

What an Investor Needs First

PUBLIC Reliant AI builds generative AI software that extracts, structures, and analyzes scientific and regulatory information for biopharma teams, a timely proposition because the company emerged from stealth with fresh seed financing in August 2024 and was later acquired by Cohere in May 2026, according to public reporting [TechCrunch, August 2024] [Reliant AI, August 2024] [AIwire, May 2026]. The founding story is straightforward but material: the company was started in 2023 by Karl Moritz Hermann, Marc Bellemare, and Richard Schlegel, combining AI research pedigrees tied to DeepMind and Google Brain with a commercial profile that includes EY-Parthenon experience [TechCrunch, August 2024] [Reliant AI, August 2024].

Its first product, Reliant Tabular, is positioned as an AI workbench for life-science analysts, aimed at querying large document sets and pulling structured data from scientific materials that are often inconsistently labeled or unstructured; TechCrunch reported that the system combines Llama 3.1 with proprietary extraction and analysis techniques, which is the clearest public signal of technical differentiation beyond a generic model wrapper [TechCrunch, August 2024] [Reliant AI, August 2024]. Public evidence on traction remains selective rather than comprehensive, but one cited case study said a client reduced time spent per document from roughly 120 minutes to 20 minutes across 51 studies, which is directionally useful even if not independently verified to institutional diligence standards [TipRanks.com].

The team is a central part of the underwriting case. Hermann is described as a former senior research scientist at Google DeepMind, Bellemare as a former DeepMind researcher and Canada CIFAR AI Chair at Mila, and Schlegel as coming from EY-Parthenon, giving the company a credible mix of model, research, and buyer-side workflow familiarity [BetaKit, August 2024] [Reliant AI, August 2024]. On financing, Reliant announced an $11.3 million seed round in August 2024 led by Tola Capital, while BetaKit reported total funding of about $13.5 million USD including prior backing from investors such as Inovia Capital, Mike Volpi, Amit Agarwal, and Yoshua Bengio; the business model is SaaS, aimed at for-profit life-science customers able to support recurring software spend [Reliant AI, August 2024] [BetaKit, August 2024] [TechCrunch, August 2024].

Over the next 12 to 18 months, the key public watchpoints would have been proof of repeatable enterprise adoption, named customer validation, and evidence that the product could expand from document extraction into higher-value decision workflows. That standalone path is now overtaken by the reported outcome: Cohere acquired Reliant AI in May 2026, with public sources indicating the transaction brought Reliant's research team, customer relationships, and biopharma-focused product capabilities into Cohere's stack [AIwire, May 2026] [Baker McKenzie, May 2026] [Morningstar, May 2026].

Partially corroborated -- Core company, founder, funding, and acquisition facts are corroborated by multiple public sources, but product performance and several customer-related claims rely partly on company or single-source reporting.

Taxonomy Snapshot

Axis Value
Stage Exited
Business Model SaaS
Industry / Vertical Deeptech
Technology Type AI / Machine Learning
Geography North America
Growth Profile Venture Scale
Founding Team Co-Founders (3+)
Funding Seed, total disclosed about $13.5M

Inside the Company

PUBLIC

Reliant AI emerged in 2023 with a fairly specific starting point: using generative AI to extract, structure, and analyze scientific and regulatory information for biopharma teams [Reliant AI, August 2024]. The company is headquartered in Montréal, Canada, according to the company’s launch materials, and its public website presents the business as built for life-science research workflows rather than general enterprise search [Reliant AI, August 2024] [Reliant AI, retrieved 2025].

The early chronology is concise but legible. In August 2024, Reliant AI announced its launch from stealth and disclosed an $11.3 million seed round led by Tola Capital, with the financing framed around expanding the product and hiring across Europe and North America [Reliant AI, August 2024]. By the company’s homepage as captured in 2025, Reliant was also publicly stating that it had "joined forces" with Cohere to support enterprise AI work across biopharma and healthcare, indicating a later change in corporate status after its initial launch phase [Reliant AI, retrieved 2025].

The founding team, as presented by the company, consists of Karl Moritz Hermann, Marc Bellemare, and Richard Schlegel [Reliant AI, August 2024]. Public company materials tie the business to a mix of machine learning research and life-science commercialization intent, which helps explain why the company’s first public positioning was narrow by function and vertical rather than broad by model capability [Reliant AI, August 2024] [Reliant AI, retrieved 2025].

Claim stands unchecked -- This section relies primarily on company website and company announcement materials, with no independent state filing or Crunchbase corroboration used here.

Under the Hood

MIXED

Reliant AI is selling a document-heavy research workflow rather than a general-purpose chatbot. Public materials describe its first product, Reliant Tabular, as an AI workbench for life-science analysts that can query large collections of scientific documents and extract structured information even when source material is inconsistently labeled or unstructured [Reliant AI, August 2024] [TechCrunch, August 2024]. The stated use cases are narrow and commercially legible: automated systematic reviews, asset scans, comprehensive analyses, and customizable evidence workflows for biopharma teams [Reliant AI, August 2024] [Airframe, retrieved 2026].

The technical claim worth watching is not model novelty on its own, but the combination of a foundation model with workflow-specific extraction logic. TechCrunch reported that the system uses Llama 3.1 alongside proprietary techniques for document extraction and analysis [TechCrunch, August 2024]. Reliant's own materials add that the platform can be configured for enterprise requirements, including custom data-pipeline integration, and that its underlying approach includes "densified Knowledge Graphs" intended to map relationships between trials and mechanisms [Reliant AI, retrieved 2025] [Reliant AI, retrieved 2026]. Those knowledge-graph and deployment claims come primarily from the company, so they are best read as product positioning rather than independently validated performance.

What can be said publicly about efficacy remains limited but directionally interesting. One cited case study, surfaced via TipRanks from a Medicus Pharma announcement, said a user cut review time per document from roughly 120 minutes to 20 minutes across 51 studies when using Reliant Tabular [TipRanks.com]. A separate company claim says the product saved a user a week of work on primary market research preparation [Reliant AI, retrieved 2026]. Neither datapoint establishes broad repeatability, but both are consistent with the product's core pitch: compressing the manual work of turning messy scientific literature into structured, decision-ready outputs.

Partially corroborated -- Core product description is corroborated by Reliant AI and TechCrunch, while technical architecture details and performance claims rely partly on company materials and a single case-study source.

Market Research

PUBLIC

The market matters now because biopharma research teams are under pressure to process more scientific and regulatory material with fewer manual hours, and Reliant AI is positioned directly on that workflow bottleneck [TechCrunch, August 2024] [Reliant AI, August 2024]. The public record does not provide a cited third-party TAM, SAM, or SOM for Reliant AI specifically, so the more defensible read is to anchor on the workflows it is trying to replace or accelerate: systematic literature review, evidence synthesis, asset scanning, competitive landscaping, and related scientific analysis inside life sciences organizations [Reliant AI, August 2024] [Airframe, retrieved 2026]. That narrows the discussion from general-purpose enterprise AI to a domain software market where the buyer values traceability, structured extraction, and integration into research processes more than broad chatbot utility [TechCrunch, August 2024] [Reliant AI, retrieved 2025].

Demand signals in the cited material are practical rather than abstract. Reliant launched around the claim that life-science analysts still spend substantial time collecting, organizing, and inspecting complex scientific data, and TechCrunch described the company as addressing the "data drudgery" of paper-heavy research work [TechCrunch, August 2024] [Reliant AI, retrieved 2025]. The strongest public workflow datapoint remains limited and should be handled cautiously: TipRanks reported that a client using Reliant Tabular cut time per document from about 120 minutes to 20 minutes across 51 studies, which suggests a meaningful labor-saving use case but is still a single sourced case study rather than a broad market benchmark [TipRanks.com].

A small sizing table is still useful here, but only for what can actually be cited.

Market lens What the evidence supports
Core market AI software for biopharma evidence synthesis, document extraction, and research analysis workflows [Reliant AI, August 2024] [TechCrunch, August 2024]
Adjacent market Scientific and regulatory information management for life sciences teams [Reliant AI, August 2024] [Airframe, retrieved 2026]
Substitute spend Analyst labor, outsourced research support, and manual literature-review processes inside pharma and biotech organizations [TechCrunch, August 2024] [Reliant AI, retrieved 2025]
Geographic footprint implied by go-to-market Europe and North America, based on hiring and expansion plans disclosed at launch [Reliant AI, August 2024]

from the public evidence is that Reliant was not trying to sell generic AI capacity. It was targeting a narrower but more budgeted slice of life-sciences knowledge work where software can compete against high-cost manual process and fragmented internal tooling [TechCrunch, August 2024] [Reliant AI, August 2024].

Adjacent and substitute markets matter because they shape both upside and competitive pressure. Reliant's product description overlaps with e-discovery style document extraction, enterprise knowledge management, regulatory intelligence tooling, and medical affairs research software, even if the company framed its wedge around biopharma analysis first [TechCrunch, August 2024] [Reliant AI, retrieved 2025]. That creates a larger possible spend pool over time, but it also means the company would need to prove that domain-specific models, structured outputs, and workflow support justify buying a specialized platform instead of extending existing search, data, or consulting workflows [IntuitionLabs] [Reliant AI, retrieved 2026].

Regulatory and macro forces also point in the same direction, even if the supplied sources do not quantify them. Drug development remains documentation-heavy and evidence-sensitive by design, which tends to favor software that can produce structured, reviewable outputs over consumer-style AI interfaces [TechCrunch, August 2024] [Reliant AI, August 2024]. The company's own emphasis on enterprise deployment options, custom data pipelines, and data lake integration suggests that procurement in this segment is shaped by data governance and system compatibility at least as much as by model quality, a pattern consistent with slower but stickier enterprise adoption if the product clears validation hurdles [Reliant AI, retrieved 2026].

Partially corroborated -- Section is grounded mainly in TechCrunch and Reliant AI materials, with limited independent market sizing and one lightly corroborated workflow case study.

Competition and Substitutes

Positioning

MIXED Reliant AI appears to sit in a narrow but commercially legible slice of applied AI, selling document-heavy biopharma research workflows rather than a general-purpose foundation model or a broad horizontal knowledge-management tool [TechCrunch, August 2024] [Reliant AI, August 2024].

The competitive map is therefore best read by segment, not by a single direct peer set. At the model layer, Cohere and Meta are upstream enablers rather than clean substitutes, since Reliant Tabular was described as combining Llama 3.1 with proprietary extraction and analysis techniques tailored to scientific documents [TechCrunch, August 2024]. At the workflow layer, the practical alternatives are still a mix of internal analyst teams, consultants, and fragmented software used for literature review, competitive landscaping, and evidence synthesis, which Reliant explicitly targeted in biopharma research processes [Reliant AI, August 2024] [Airframe, retrieved 2026]. The result is a company competing less on raw model novelty than on whether it can make regulated, messy scientific information usable inside an enterprise workflow.

That positioning gives Reliant one visible edge today: founder-market fit at the intersection of advanced AI research and a specific commercial use case. Public reporting ties the founding team to DeepMind, Google Brain, Mila, and EY-Parthenon, a mix that is unusual for a young vertical AI company and relevant to both product design and enterprise conversations in life sciences [TechCrunch, August 2024] [BetaKit, August 2024]. The edge looks strongest where credibility matters, namely persuading biopharma teams that a model-driven system can parse inconsistent scientific source material with enough structure to support downstream analysis [Reliant AI, August 2024] [TechCrunch, August 2024]. It is also perishable. If the product advantage rests mainly on workflow design and prompting around third-party models, larger platform companies or well-resourced pharma software vendors could compress that lead once the use case is proven.

Exposure is clearest in distribution and platform control. Tola Capital's public post and multiple trade reports indicate Reliant was acquired by Cohere in May 2026, with customer relationships and product capabilities folded into Cohere's healthcare push [AIwire, May 2026] [Baker McKenzie, May 2026] [Morningstar, May 2026]. That changes the competitive frame materially. As a standalone company, Reliant did not appear to own a broad channel into pharma at scale, and the public record did not name a large set of verified customers before the acquisition [BetaKit, August 2024]. Inside Cohere, it gains a stronger enterprise distribution vehicle but also stops existing as an independent contender. In that sense, the most relevant named competitor is Cohere itself, not because it was a like-for-like rival in 2024, but because enterprise AI platform vendors with capital, compute relationships, and existing procurement paths are best positioned to absorb narrow workflow products.

The most plausible 18-month scenario, viewed from the August 2024 starting point, was that the winner would be Cohere if enterprise buyers preferred a full-stack, procurement-ready AI vendor that could package vertical functionality on top of a broader model and deployment platform [AIwire, May 2026] [LinkedIn, retrieved 2026]. That is broadly consistent with the reported outcome. The loser if a different condition held would have been Reliant as a standalone vendor, if customer demand shifted from point solutions toward integrated enterprise AI suites before Reliant could establish enough direct distribution and verified reference accounts to defend independence [BetaKit, August 2024] [TechCrunch, August 2024]. The public evidence supports a view of Reliant as a strong product wedge in a market where owning the channel mattered at least as much as owning the workflow.

Partially corroborated -- Competitive positioning is supported by TechCrunch, BetaKit, and Reliant's launch materials, but the section lacks named standalone direct competitors in the source set and relies partly on analytical inference from acquisition evidence.

Opportunity

PUBLIC

The prize here is not another point tool for literature search, but a credible shot at becoming core research infrastructure for biopharma teams that need scientific evidence turned into usable decisions at machine speed [TechCrunch, August 2024] [Reliant AI, August 2024].

The headline opportunity is straightforward. If Reliant AI can move from a workbench for document extraction into a system of record for scientific and regulatory evidence, it could become the default analytical layer for biopharma diligence, competitive intelligence, and trial planning workflows [Reliant AI, August 2024] [TechCrunch, August 2024]. That is reachable, not merely aspirational, because the company paired a narrowly defined first market, biopharma, with a product built for messy source material and evidence-heavy tasks, and it did so with founders whose backgrounds span DeepMind, Google Brain, and EY-Parthenon [TechCrunch, August 2024] [BetaKit, August 2024]. The public record also shows early investor support from Tola Capital and Inovia Capital, plus participation from Mike Volpi, with total disclosed funding of about $13.5 million, enough to suggest the company had room to build beyond a prototype before its reported exit [Reliant AI, August 2024] [BetaKit, August 2024].

Three scale paths stand out from the evidence.

Scenario What happens Catalyst Why it's plausible
Evidence workflow standard Reliant Tabular becomes the default interface for systematic reviews, asset scans, and scientific evidence extraction across pharma strategy and R&D teams A large pharma deployment proves that time savings on one workflow can be repeated across many evidence-heavy functions Reliant positioned the product around automated systematic reviews, comprehensive analyses, and customizable data workflows for life sciences, while a cited user example reported reducing time per document from about 120 minutes to 20 minutes across 51 studies [Reliant AI, August 2024] [TipRanks.com]
Enterprise data layer for biopharma AI The product shifts from analyst workbench to embedded enterprise layer, feeding structured scientific data into internal knowledge systems and data lakes Successful enterprise integrations make switching costs higher and widen the buyer from individual teams to platform owners Reliant said its technology could support enterprise requirements including custom data pipelines for data lake integration, and TechCrunch reported a mix of Llama 3.1 with proprietary extraction techniques rather than a generic chat interface [Reliant AI] [TechCrunch, August 2024]
Strategic platform inside a larger AI stack Reliant's biopharma specialization becomes the domain wedge inside a broader enterprise AI vendor serving regulated industries Acquisition or deep distribution through a larger model and infrastructure company accelerates go-to-market and customer access Cohere acquired Reliant AI on May 19, 2026, according to AIwire, and public reports tied that transaction to customer relationships including GSK, Medicus Pharma, and Kyowa Kirin, which suggests the domain layer had strategic value beyond a standalone seat-based tool [AIwire, May 2026] [Baker McKenzie, May 2026]

The compounding dynamic would come from structure, not from raw model novelty alone. Every successful extraction workflow creates cleaner internal datasets, more reusable templates, and more confidence in automating adjacent workflows such as competitive landscaping, evidence synthesis, and trial-related analysis [Reliant AI, August 2024] [Reliant AI]. If the system is also connected to enterprise data lakes, as the company claimed, then each new deployment can become harder to displace because outputs stop being one-off reports and start feeding recurring internal processes [Reliant AI]. Public evidence that this flywheel had begun is limited, but there are early signals: BetaKit reported recurring platform usage with select partners, and the company described use case-specific support and customizable workflows, both consistent with a services-aided land-and-expand motion in a high-value vertical [BetaKit, August 2024] [Reliant AI].

The size of the win is best framed through the exit already on record. Cohere's acquisition indicates that a specialized life-sciences evidence platform could matter enough to a larger foundation-model company to buy rather than build, particularly when the target brings domain datasets, customer relationships, and a technical team [AIwire, May 2026] [The Healthcare Technology Report]. With 41 employees at the time of acquisition, according to PitchBook, Reliant appears to have reached strategic relevance while still relatively small in headcount [PitchBook]. If the strongest scenario had played out as a scaled standalone platform rather than an early strategic sale, the plausible value could have been meaningfully higher than a talent-and-technology acquisition, potentially in the range associated with vertical AI infrastructure assets in regulated enterprise software (scenario, not a forecast). The evidence on public comps is too thin here to anchor a precise valuation band conservatively, but the reported buyer interest from Cohere is enough to establish that the upside was not theoretical [AIwire, May 2026] [Morningstar, May 2026].

Partially corroborated -- This section relies on independent reporting from TechCrunch, BetaKit, AIwire, Baker McKenzie, Morningstar, and PitchBook, but several product and workflow claims remain company-sourced or based on a single reported case study.

Sources

Open sources

  1. [BetaKit, August 2024] Ex-Google DeepMind leaders bring Reliant AI out of stealth with $15.4-million CAD seed round | https://betakit.com/ex-google-deepmind-leaders-bring-reliant-ai-out-of-stealth-with-15-4-million-cad-seed-round/

  2. [TechCrunch, August 2024] Reliant’s paper-scouring AI takes on science’s data drudgery | https://techcrunch.com/2024/08/20/reliant-ai/

  3. [Reliant AI, August 2024] Seed Announcement | https://www.reliant.ai/seed-announcement

  4. [AIwire, May 2026] Cohere Acquires Reliant AI to Advance Scientific Discovery in Pharma, Biotech and Healthcare | https://www.aiwire.net/2026/05/19/cohere-acquires-reliant-ai-to-advance-scientific-discovery-in-pharma-biotech-and-healthcare/

  5. [TipRanks.com] Medicus Pharma Partners with Reliant AI to Transform Clinical Trial Analysis | https://www.tipranks.com/news/company-announcements/medicus-pharma-partners-with-reliant-ai-to-transform-clinical-trial-analysis

  6. [Morningstar, May 2026] Cohere Acquires Reliant AI to Advance Scientific Discovery in Pharma, Biotech and Healthcare | https://www.morningstar.com/news/globe-newswire/9441365/cohere-acquires-reliant-ai-to-advance-scientific-discovery-in-pharma-biotech-and-healthcare

  7. [Baker McKenzie, May 2026] Cohere acquires Reliant AI | https://www.bakermckenzie.com/en/newsroom/2026/05/cohere-acquires-reliant-ai

  8. [Reliant AI, retrieved 2025] Reliant AI - Complex research made simple | https://www.reliant.ai/

  9. [Airframe, retrieved 2026] Reliant AI | https://www.airframe.com/company/reliant-ai

  10. [IntuitionLabs, retrieved 2026] Reliant AI | https://www.intuitionlabs.ai/startups/reliant-ai

  11. [PitchBook, retrieved 2026] Reliant AI | https://pitchbook.com/profiles/company/reliant-ai

  12. [The Healthcare Technology Report, retrieved 2026] Cohere Acquires Reliant AI to Advance Scientific Discovery in Pharma, Biotech and Healthcare | https://www.thehealthcaretechnologyreport.com/cohere-acquires-reliant-ai-to-advance-scientific-discovery-in-pharma-biotech-and-healthcare/

  13. [LinkedIn, retrieved 2026] Karl Moritz Hermann - Cohere | LinkedIn | https://www.linkedin.com/in/karlmoritz/

Articles about Reliant AI

View on Startuply.vc