AI solutions — est. 2016

WAIW Bringing AI to the energy industry

Browser-based seismic interpretation, agentic assistants and applied machine learning — built by a team that's been shipping AI since 2016.

Book a demo Explore products

01 Products

Two products, one goal:
faster, better subsurface decisions

GeoSurreal
A platform for AI-assisted seismic interpretation, turning weeks of work into hours.
GeoAssistant
A team of specialist agents working across large industry datasets, seismic among them, in geophysics and outside it.

GeoSurreal

The AI platform for seismic interpretation

GeoSurreal runs the full structural-interpretation pipeline in one environment, with AI tools that speed up every step. It opens in any browser — no installation — and deploys on any cloud or on-prem.

  • AI-powered horizon and fault interpretation
  • Three horizon approaches — seeded, one-click, or fully unsupervised
  • Fast SEG-Y access, powered by our proprietary Rust engine
  • QC maps created instantly for every horizon
See it on your data
A seismic volume in GeoSurreal with AI-predicted horizons and faults rendered in 3D
  1. 01

    Upload

    SEG-Y at exceptional speed — no local copies, no re-indexing.

  2. 02

    Predict

    Advanced ML models, tested by subject-matter experts on real data.

  3. 03

    QC

    Compare the output of multiple models and look for discrepancies.

  4. 04

    Export

    In all the industry standard formats seamlessly.

GeoAssistant

An agentic runtime inside your perimeter

GeoAssistant (GAS) is a team of agents that plan and coordinate their own work with your data. It runs on any model, reasons over large volumes of industry-specific data, and connects with the infrastructure and agents you already run.

YOUR PERIMETER GAS Cloud models Azure ∘ Bedrock ∘ direct APIs Your agents & tools Local models vLLM ∘ Ollama Corporate systems & documents file shares ∘ internal services OSDU ∘ knowledge base Your web apps Embedded GAS chat Embeddable apps Chat multi-modal Viewers, editors library of twelve UI extensions sandboxed, from MCP Agentic core Fast answers quickly General interprets intent Think plans, executes Platform services Deployment Compose, Kubernetes Security your IdP, OPA Operations logs, console, CLI Four-layer information model Workspace ∘ Artifacts ∘ Knowledge ∘ Memory A2A MCP Agents come with built-in tools for reading, writing, computing and plotting. Your own MCP servers and A2A agents plug into the same place and become first-class to every conversation. OpenAI, Claude, Gemini and others connected without any code changes Every file and fact the agents work with is stored and catalogued inside your perimeter, and remains yours at all times. GAS provides you plug in Your agents & tools Agents come with built-in tools for reading, writing, computing and plotting. Your own MCP servers and A2A agents plug into the same place and become first-class to every conversation. Your web apps Embedded GAS chat Embeddable apps Chat multi-modal Viewers, editors library of twelve UI extensions sandboxed, from MCP Agentic core Fast answers quickly General interprets intent Think plans, executes Platform services Deployment Compose, Kubernetes Security your IdP, OPA Operations logs, console, CLI Four-layer information model Workspace ∘ Artifacts Knowledge ∘ Memory GAS Local models vLLM ∘ Ollama Corporate systems & documents file shares ∘ internal services OSDU ∘ knowledge base Every file and fact the agents work with is stored and catalogued inside your perimeter, and remains yours at all times. YOUR PERIMETER Cloud models Azure ∘ Bedrock ∘ direct APIs OpenAI, Claude, Gemini and others connected without any code changes GAS provides you plug in

  • Visible, self-correcting plans

    Updated and adapted live as the work evolves; on genuine ambiguity agents stop and ask.

  • Reliable throughout

    Every step is reviewed and nothing fails quietly: you get what’s finished and a full account of the rest.

  • Domain specialists you register

    Add your agents for industry-specific workflows: the system finds them and delegates.

Specialist agents you can trust

> “Provide a geophysical interpretation of the attached seismic section”

Seismic section annotated by GeoAssistant with eight numbered interpretation markers 1 · Seafloor / water-bottom reflector Sharp, flat, continuous high-amplitude event marking the top of the sedimentary section. Uniform character suggests a marine survey with a quiet, low-relief seabed. 2 · Shallow parallel-bedded unit Package of continuous, parallel to sub-parallel reflectors with alternating high-amplitude events — undisturbed hemipelagic / low-energy deposition; minor lateral amplitude variation. 3 · Rugose high-amplitude unconformity Irregular, bright (red-blue doublet) erosional surface with local relief; truncates underlying reflectors. Likely a major sequence boundary separating two structural regimes. 4 · Incised feature (channel / canyon) V-shaped erosional depression cut into the unconformity on the right side of the line — an incised valley or canyon, later infilled by onlapping strata. 5 · Dipping / rotated strata with faults Interval of inclined, semi-continuous reflectors with small-offset normal faults and local chaotic texture — possibly rotated fault blocks or slumped / mass-transport material beneath the unconformity. 6 · High-amplitude layered package Strong 'railroad-track' parallel reflectors (interbedded hard/soft lithologies, e.g. carbonates, cemented sands or volcanics) offset by a series of steep normal faults with clear step-like throws. 7 · Isolated bright-spot anomaly Localized high-amplitude red/blue anomaly detached from the layered package — candidate DHI (gas accumulation) or intrusive body (sill tip); worth AVO / frequency screening. 8 · Deep low-S/N zone Progressive loss of amplitude, continuity and frequency content with depth — attenuation below the faulted package; imaging limit of this dataset.
Automated geophysical interpretation — real agent output. Hover or tap the markers.

02 Services

Beyond the products

The same team that builds GeoSurreal works directly with energy companies — on models, data and people.

Bespoke ML models, AI agents and tools, built in-house and delivered to your spec.

We also build to order. For clients across the energy industry and beyond we have delivered ML models, tools and agentic systems, deployed in the cloud, on-prem or fully air-gapped.

The work has covered seismic, petrophysics and lithology among other fields, and it has left us with a deep toolkit that lets us take that expertise into new ones. Some of it is public, in the open-source libraries on our GitHub. Some went into published research, and some stayed with the client.

Applied AI and ML for upstream: upgrading existing workflows, introducing new ones, and getting value out of data you already have.

Besides our own products, we offer consulting services. Our engineers come into existing projects and help teams adopt the latest AI and ML technologies, planning, guiding and supporting new pipelines across the energy industry.

In one of our previous projects we took on overpressure prediction, using open-source tools to automate the transfer of ML algorithms into the client’s production pipeline and streamline an existing solution, predicting mud weight in carbonate formations from seismic data.

You can read more in this paper.

Every project is different, but our team remains the same: the people who build our products, with over 12 years of applied AI behind them.

Hackathons, workshops and courses, designed and run by our own team. Real industry problems on real data, for newcomers and experienced practitioners alike.

Over the past six years we’ve designed and run hackathons, workshops and courses across the energy industry, many of them with EAGE. Sessions tackle real industry problems on real data with modern AI and machine learning, and they’re hand-crafted rather than packaged — twelve years of AI and data-science experience, shared with newcomers and veterans alike.

Online and on-site, across the Middle East, Europe and the US. We’ve served on EAGE technical committees since 2020.

03 History

Around since 2016

  1. 2016

    Founded as an AI research group working on healthcare — ECG analysis and CT imaging at scale.

  2. 2019

    The technology stack moves into oil & gas: seismic processing, interpretation, petrophysics.

  3. 2021

    Operating as DAC, with US incorporation and expanded NVIDIA and AWS partnerships.

  4. 2023

    DAC becomes WAIW Software LLC — you'll still find the old name on some legacy materials.

  5. Now

    Scaling in MENA, North America and Europe, with GeoSurreal deployed and GeoAssistant growing fast.

04 Research

Our research through the years

Nine years of published work, open-source libraries and long-form write-ups. Throughout our history we have shared what we find with our peers in the industry, in journals, at conferences and on GitHub.

05 Partners

Our partners

06 Contact

Book a demo

See GeoSurreal and GeoAssistant on your own data — or talk to us about development, consulting and training.