# Shift Technology account plan

> Shift Technology is a Paris-founded AI software vendor for insurers, covering fraud, claims, subrogation and underwriting, with about 600 people.

## At a glance

Shift Technology, founded in Paris in 2014, sells AI software to insurers that automates and optimises decisions from underwriting to claims handling; claims fraud detection is its origin product. It works with more than 120 insurers in P&C, health, life, prévoyance and travel insurance in more than 25 countries, with customers in Europe, the US and Japan, and says its fraud tools serve almost 70% of the US P&C market. The company had about 600 people in 2025, including more than 200 data scientists. It raised a total of EUR 320 million between 2016 and 2021 from investors including Accel, Advent International, Bessemer Venture Partners, General Catalyst and Bpifrance; no revenue figure is given in the documents.

## Business units

- **Fraud detection**: The flagship product flags potentially fraudulent claims at claim, group-of-claims and provider level for insurers' special investigation units and claim handlers. Shift says its solutions have analysed close to 3 billion claims and help insurers save up to EUR 5 billion of claims fraud a year.
- **Shift Claims (agentic claims automation)**: Launched on 16 September 2025, Shift Claims is an AI layer on top of insurers' claims systems with agents that assess complexity, prioritise and assign, assist handlers and automate claims from first notice of loss to closure. Early adopters, including AXA Switzerland, report 30% faster claims handling and a 60% overall automation rate (company-reported).
- **Subrogation detection**: Detects recovery opportunities insurers miss; the CEO says it has taken off and is more impactful in the US than in Europe, with integration with arbitration forums as the next step.
- **Underwriting risk assessment**: Flags misrepresentation, returning fraudsters and broker fraud at policy subscription, extending the platform from claims to underwriting.
- **Insurance Data Network (IDN)**: A decentralised network on which carriers share claims data across more than 50 fields; the CEO says almost 70% of US claims data is shared on it. It is free to members on the Shift platform and Shift does not own or resell the data.

## Strategic priorities

- **Agent-first product strategy**: Since Shift Claims in September 2025, Shift sells AI agents that execute workflows such as intake, investigations and liability determination, with human-in-the-loop checkpoints; its products are being rebuilt so a fraud capability becomes a callable fraud agent.
- **Orchestration, monitoring and the insurance intelligence layer**: The CEO names these three areas as investment priorities, aiming at capabilities that fit into insurers' IT stacks with clear ROI and explainability, without rip-and-replace projects.
- **Growing the Insurance Data Network**: Continued investment in IDN is described as directly benefiting the agentic strategy, since agents can call the cross-carrier data as a tool.
- **Multi-product expansion**: Shift has widened from claims fraud to subrogation, underwriting fraud, health fraud, waste and abuse, and anti-money-laundering, and set national cross-insurer data platforms as a development axis from 2024.

## Investments and deals

- **Venture funding**: EUR 320 million raised in several rounds between 2016 and 2021 from Accel, Advent International, Bessemer Venture Partners, Bpifrance Large Venture, General Catalyst and Iris Capital.
- **Generative and agentic AI**: Generative AI with insurance-trained language models was integrated into the products in 2024, followed by the launch of agentic Shift Claims in September 2025.

## Challenges and risks

- **Insurers building their own agents**: The Chief Product Officer argues against in-house builds, conceding the opportunity to build is real but citing data architecture, tool integration, dedicated R&D and domain modelling as hurdles.
- **Accuracy and reliability demands**: Insurance flows can require accuracy above 99%, which Shift addresses with multi-agent review, drift monitoring and calibrated abstention with human escalation.
- **Market education and change management**: The CEO says a lot of education is needed in the market and that adoption requires change management and guardrails; AI is still feared as a job killer.
- **Fast-moving model landscape**: Model jumps, notably in February, changed what agents can do within six months, so architectures need continuous maintenance as models change.

## What leadership says

> "We see continued investment in IDN as directly benefiting our agentic strategy." — Jeremy Jawish, Co-founder and CEO, Four questions with Jeremy Jawish: Why agentic? Why now?, 21 April 2026

> "I think we can make people 10 times more efficient" — Jeremy Jawish, Co-founder and CEO, Profiles in Risk podcast, Ep. 727, recorded 14 July 2025

> "It took us at least three years to find the first customer." — Jeremy Jawish, Co-founder and CEO, Profiles in Risk podcast, Ep. 727, recorded 14 July 2025

> "Prompt engineering alone is simply not enough" — Eric Sibony, Co-founder, Chief Scientist and Chief Product Officer, Four questions with Eric Sibony: Shift's agents, April 2026


## Where a supplier can help

- **Agent orchestration and monitoring tooling**: Orchestration and monitoring are named investment priorities, so suppliers of LLM infrastructure, evaluation and drift-monitoring tools fit the agent-first strategy.
- **Integration partners for insurer IT stacks**: Shift Claims sits on top of insurers' claims and core systems; systems integrators and claims-platform partners can help deliver deployments without rip-and-replace projects.
- **Change management and training for insurers**: The stated need for market education and change management opens a role for consultancies that help insurers adopt agentic claims automation.
- **Secure data-sharing infrastructure**: Growing the Insurance Data Network across carriers calls for secure, privacy-preserving data infrastructure and compliance expertise.

_Tamtam's reading of the priorities and challenges above, not the company's own words._

## Key documents

### [Shift Technology enters its tenth year (company press release)](https://prnewswire.com/fr/communiques-de-presse/shift-technology-entame-sa-dixieme-annee-dactivite-et-poursuit-son-evolution-grace-au-soutien-des-assureurs-et-de-lecosysteme-francais-302007573.html)

Report · 2023-12-06

- Shift Technology (Paris, founded 2014) sells AI-based decision automation and optimisation software to insurers, from underwriting to claims handling; fraud detection on claims is the origin product.
- Works with more than 120 insurers (P&C, health, life, prévoyance and travel) in more than 25 countries.
- Its solutions have analysed close to 3 billion claims and help insurers save up to EUR 5 billion of claims fraud a year, about EUR 10 million per billion of written premiums (company figures).
- Product scope widened to subrogation detection, underwriting fraud, and anti-money-laundering / counter-terrorist-financing (AML-CFT); portfolio adapted to health and prévoyance (personal lines).
- 200 data scientists in Paris, a third of headcount, described as the largest data-science team dedicated to insurance in the world.
- Raised a total of EUR 320 million between 2016 and 2021 from Accel, Advent International, Bessemer Venture Partners, Bpifrance Large Venture, General Catalyst and Iris Capital; Next40 member for a 4th consecutive year.
- Goal for 2024 onwards: keep expanding, a multi-product approach, and national cross-insurer data platforms ("plateformes nationales de croisement de données").

### [Shift Technology launches Shift Claims to power claims transformation with agentic AI](https://shift-technology.com/resources/news/shift-technology-launches-shift-claims-to-power-claims-transformation-with-agentic-ai)

Report · 2025-09-16

- Shift Claims, launched 16 September 2025 (Boston and Paris), is an agentic AI product that assesses, prioritises, assists and automates insurance claims from first notice of loss (FNOL) to closure.
- It sits as an AI layer on top of insurers' existing claims and core systems; agents are trained on the insurer's own processes and insurers keep control.
- Early adopters report 3% lower claims losses, 30% faster claims handling, a 60% overall automation rate and greater than 99% accuracy in claims assessment.
- AXA Switzerland is an early adopter (quote from its COO and Head of Customer Operations, Markus Keller).
- Four agent roles: assess claim complexity, classify/prioritise/assign, assist and advise handlers, and an STP agent automating claims or tasks.
- Builds on the Gen AI Shift integrated into its solutions in 2024 (insurance-trained LLMs for extraction, summarisation, classification, liability determination for subrogation).
- Shift now positions itself as "the leading AI platform for insurance", combining generative, agentic and predictive AI across underwriting, claims, and fraud and risk.

### [Jeremy Jawish, CEO at Shift Technology - Profiles in Risk Ep. 727](https://www.youtube.com/watch?v=eOkxvmKNhY8)

Video · Jeremy Jawish, Co-founder and CEO · 2025-10-29

- Shift is a "vertical AI company" for insurers; its flagship fraud detection product serves "almost 70% of the US PNC market".
- The Insurance Data Network (IDN), built at carriers' request, lets insurers share claims data in a decentralised, secure way; "almost 70% of the United States" claims data is shared on it, more than 50 fields; free to members but requires being on the Shift platform; Shift does not own or resell the data.
- Subrogation detection is the product that has "really taken off", more impactful in the US than in Europe; next step is integration with arbitration forums.
- Other products: underwriting risk assessment (misrepresentation, returning fraudsters, broker fraud) and, brought from Europe, agentic claims automation.
- With agentic AI, automation rates rose "from 20% to in some cases 60%" while accuracy stayed at 99% on automated tasks.
- Company: more than 200 data scientists and 600 people; took "at least three years to find the first customer"; recorded in Boston on 14 July 2025.

### [Four questions with Jeremy Jawish: Why agentic? Why now?](https://www.shift-technology.com/resources/reports-and-insights/four-questions-with-jeremy-jawish-why-agentic-why-now)

Executive talk · Jeremy Jawish, Co-founder and CEO · 2026-04-21

- Shift's latest evolution, started with Shift Claims in September 2025, productises "scalable, secure AI agents" executing insurance workflows such as intake, investigations and liability determination, with human-in-the-loop checkpoints.
- Agents reuse Shift's existing models, scoring engines, APIs and "insurance intelligence layer", so customers avoid "rip-and-replace projects".
- Investment priorities: orchestration, monitoring and the insurance intelligence layer.
- Strategy is "agent first": productise deeper capabilities and differentiated services.
- Pitch to insurers centres on the workforce: agents bridge the industry's talent gap rather than kill jobs.
- Three forces make agentic AI viable now: model capability, proven use cases in claims and fraud, and operational pressure (costs, labour shortages, customer expectations).
- The Insurance Data Network (IDN) gives members cross-carrier claims data and acts as a tool agents call; continued IDN investment directly benefits the agentic strategy.

### [Four questions with Eric Sibony: Shift's agents](https://www.shift-technology.com/resources/reports-and-insights/four-questions-with-eric-sibony-shifts-agents)

Executive talk · Eric Sibony, Co-founder, Chief Scientist and Chief Product Officer · 2026-04-21

- Shift moved from "AI as core" to "agents as product": agentic reasoning injected into the engine plus an integration/packaging layer so agents act as callable services.
- "Selling agents" means, e.g., a claims fraud capability becomes a "fraud agent" that investigates with multiple tools, escalates to humans and can be invoked by other agents.
- Model jumps (notably in February) pushed agents from short micromanaged tasks toward "sustained, multi-day reasoning"; agentic reasoning is becoming the core of products across claims, fraud and subrogation.
- Argues against insurers building their own agents: data architecture, tool integration, long-horizon state, dedicated R&D, and insurance domain modelling are hard; accuracy needed can be ">99% in some flows".
- Calibrated abstention ("I don't know") becomes more important with agents: multi-agent review, drift monitoring, risk-adjusted yes/no/I don't know with human escalation.

## Questions about Shift Technology

### What does Shift Technology do?

Shift Technology sells AI software to insurers for fraud detection, claims automation, subrogation detection and underwriting risk assessment. Its newest product, Shift Claims, uses AI agents to automate claims from first notice of loss to closure.

### How many employees does Shift Technology have?

The CEO said in 2025 that Shift has about 600 people, including more than 200 data scientists.

### What is Shift Technology's strategy?

Shift follows an agent-first strategy: it packages its fraud, claims and subrogation capabilities as callable AI agents layered on insurers' existing systems, and invests in orchestration, monitoring and its Insurance Data Network.

### Who leads Shift Technology?

Jeremy Jawish is co-founder and CEO; Eric Sibony is co-founder, Chief Scientist and Chief Product Officer.

_Written by Tamtam from the company's own published documents listed above, read in full. Figures are as those documents state them, for the periods they name. (2026-11-17)_
