Sensay
Sensay is an offboarding platform where Sophia, an AI interviewer, captures a departing employee's know-how by text or voice and turns it into a knowledge base the team can question from Slack, Microsoft Teams, the web or an intranet.
What is Sensay?
Sensay is an offboarding platform: its stated job is to capture, transfer and retain a team's critical know-how before it walks out of the door. At the center sits Sophia, an AI interviewer who talks to employees who are leaving or changing roles. The conversation runs by text or by voice, over one or several sessions, at the employee's own pace, and it is not a checklist. Sophia is built to probe for the nuances of a role and the tacit knowledge that never reaches a handover document, and the files and links the employee drops in are tied back to the answers.
What comes out is structured three ways: a full timestamped record of the interview, a summary of issues, risks and priorities, and a conversational knowledge agent. Colleagues then question that knowledge base whenever they need it, from Slack, Microsoft Teams, the web interface or the company intranet, and can ask for tailored documents to be generated on demand. Administrators customize the interview scripts, follow departures on a real-time dashboard and manage who has access to what.
The home page argument is built on numbers: $31 billion lost every year to knowledge attrition, 650 hours a year spent hunting for information the company already holds, six months to replace an expertise, 30 to 200 percent of an annual salary to replace a person. Sensay claims 95 percent knowledge retention, onboarding 40 percent faster, two days from setup to first interview and a threefold return on training and search time. It says more than 200 companies use it, and its About page reports 500,000+ daily chatbot interactions, 25,000+ monthly active users and 10,000+ chatbots created.
Technically, Sensay announces AES-256 encryption at rest, TLS in transit and role-based access control, alongside a fully open REST API, JavaScript and Python client libraries and webhooks. The publisher is Sensay Pte Ltd, a private company registered in Singapore under number 202516659M, founded by CEO Dan Thomson after a concussion showed him how fragile memory is. One piece of vocabulary is worth noting: the site treats the knowledge bases it produces as digital replicas of people, with a dedicated policy for taking down unauthorized ones.
What it does
- Interview departing employees through Sophia, an AI interviewer, by text or by voice
- Turn each interview into a knowledge base the team can question in plain language
- Answer questions inside Slack, Microsoft Teams, the web app or the company intranet
- Produce three outputs from one interview: full timestamped record, summary of issues, risks and priorities, and a conversational knowledge agent
- Attach files, links and context so answers stay tied to real documents
- Generate tailored documents on demand from the captured material
- Track departures and interview progress on a real-time admin dashboard
When to use Sensay / When not to
A quick filter to help you decide if Sensay is the right fit.
When to use Sensay
- HR and people teams handling resignations, retirements and internal moves who need departures to leave something behind
- Organisations with high turnover, or where critical expertise sits with only two or three individuals
- Companies trying to get a replacement productive faster, a gain Sensay puts at 40 percent quicker onboarding
- Teams going through mergers, acquisitions or several transitions at once, which the custom plans are built for
- Slack and Microsoft Teams shops, and technical teams happy to wire capture in through the REST API
When not to use Sensay
- Individuals and freelancers: the terms reserve the service for personal, non-commercial use unless Sensay agrees otherwise in writing (ToS 2.1 and 3.5)
- Buyers with no budget line: no free plan and no free trial is published, and the only way in is a sales demo booked through the HubSpot calendar
- Procurement teams that require a published subprocessor list or a customer-facing data processing agreement, since neither exists
- Organisations that must name the country their data sits in: no hosting country or region is disclosed and data residency is an enterprise-only option
- Security buyers who need SOC 2 or ISO 27001 evidence, and anyone under 13, who is not allowed to use the service at all (ToS 2.2)
How to use Sensay
A typical end-to-end flow, from setup to results.
- Book a demo or a sales call through Sensay's HubSpot calendar, since no self-service sign-up is offered
- Create the organization's account: four steps, with a guided setup announced as a matter of minutes
- Build a knowledge base and invite the contributors who will feed it
- Declare who is leaving, and let Sensay schedule their interview
- Let Sophia run the interview by text or by voice, in one sitting or across several sessions
- Have the employee add files, links and context as the conversation goes
- Collect the three outputs: full timestamped record, summary, and conversational knowledge agent
- Question the knowledge base from Slack, Microsoft Teams, the web app or the intranet
- Ask for a tailored document to be generated when you need a written deliverable
- On the admin side, customize interview scripts, follow departures in real time and manage access; developers can go further through the REST API, the JavaScript and Python SDKs and webhooks
Pros & Cons
Pros
- Tackles a problem that is rarely tooled at all: the tacit knowledge that leaves with the person
- A conversation rather than a handover form, by voice or text, at the leaver's own pace
- Output lands where the team already works, in Slack, Microsoft Teams or the intranet, rather than in yet another portal
- Three levels of restitution, full record, summary and conversational agent, instead of a single document
- One readable price with no tiered grid to decode, and a fast claimed start of a few minutes of configuration and two days to the first interview
- Documented ability to refuse model training on your data (privacy policy 10.12), and an open REST API with JavaScript and Python SDKs for technical teams
- A 30-day money-back guarantee (ToS 2.7) and a dedicated take-down policy for unauthorized replicas
Cons
- No free plan and no free trial are published, and the only way in is a commercial demo
- The price is displayed as $500 with no currency code, and billing is per knowledge base with no published definition of what counts as one
- The official API documentation at docs.sensay.io returned a 404 on 24 August 2026, even though the About page links to it
- No subprocessor list, no customer-facing DPA, no SOC 2 or ISO 27001 certification, and end-to-end and client-side encryption are announced with a 'where possible' reservation while fully homomorphic encryption is still forthcoming
- No hosting country or region is named publicly, and data residency is reserved for enterprise customers
- The license granted on User Content is unusually broad: perpetual, irrevocable, sublicensable and exploitable for commercial purposes (ToS 4.3)
- The positioning is inconsistent, the Investors page describing a 24/7 sales and support chatbot rather than the offboarding platform sold on the home page; the terms also carry clauses inherited from another product, with streaming courses and a health and wellness disclaimer, and PDPA requests and consent withdrawal point to info@sensay.ai, a different domain from the site
Pricing & Plans
No free plan and no free trial are published. The single published price point is $500 per knowledge base per year, shown on the pricing page without a currency code, so buyers should confirm with the vendor whether the figure is in US or Singapore dollars. Each knowledge base includes AI-led interview sessions, document integration, secure encrypted storage available 24/7, conversational search through Slack, Microsoft Teams or the web, and full administrative control. Custom plans covering multiple transitions, post-merger integration or continuous organization-wide capture are quoted on request through Talk to sales. A 30-day money-back guarantee applies to the first purchase (ToS 2.7); beyond that no refund is granted, subscriptions renew automatically and periods already started are not refunded (ToS 3.3). Cancellation is handled by email to support or from the account settings, with access running to the end of the paid period. Taxes are borne by the customer, Singapore GST included (ToS 3.2), and Sensay reserves the right to change its prices, with notice for existing subscriptions.
- $500 per knowledge base per year. Each knowledge base includes AI-led interview sessions run by a domain-expert AI interviewer
- document integration covering files
- links and context
- secure encrypted and compliant storage available 24/7
- conversational search through Slack
- Microsoft Teams or the web
- and full admin control over progress tracking
- interview customization and access management.
- quoted on request for multiple transitions
- post-M&A integration or continuous organization-wide capture. Both blocks lead to the same HubSpot calendar
- under Book demo and Talk to sales.
Data, GDPR & hosting
A consolidated view of how Sensay handles your data.
GDPR overview
Sensay's privacy policy, last updated 5 May 2025, claims a dual regime: Singapore's PDPA and the GDPR for users in the EEA and the UK. A data protection officer is named, Marco Bettiolo, at gdpr@sensay.io. Section 16 lists the rights to information, access, rectification, erasure, restriction, portability, objection, protection against automated decisions and withdrawal of consent, with a one-month response window extendable by two. Transfers out of the EEA and the UK rely on the European Commission's standard contractual clauses or the UK addendum, plus supplementary measures. An EU representative has been appointed but is not named. DPIAs, a processing register and regular audits are described, and breach notification to Singapore's PDPC is promised within three calendar days. Two limits: the policy still says it applies Privacy Shield principles despite the framework's invalidation, and consent withdrawal and PDPA requests are routed to info@sensay.ai, a different domain.
Who owns the data?
On paper, ownership stays with you: the terms require that you hold all rights to your User Content, and the privacy policy states that users keep the intellectual property they contribute to their replicas. The license granted alongside it is far wider: non-exclusive, worldwide, royalty-free, transferable, sublicensable, perpetual and irrevocable, covering copying, adaptation, publication, distribution, storage, modification and full exploitation across all media, commercial purposes included. Feedback and suggestions are non-confidential and freely usable. Sensay declares itself data controller; on enterprise accounts, some data is shared with the organization managing the account. Export in a machine-readable format and deletion on request are both documented.
Reuse rights
The policy sets out what Sensay does with the material rather than restricting what you do with your own. Stated purposes are running and improving the service, personalization, feature development, research and debugging, and training and improving the AI models, including the models that simulate a person's behavior, knowledge and personality. Sensay also cites service notifications and support, newsletters and marketing with consent, surveys, security, fraud prevention, identity verification, audits and legal obligations, aggregated anonymized analytics and internal reporting, product R&D and work with research partners on anonymized data, and personalized advertising with consent. The legal bases are the Singapore PDPA, through consent, deemed consent and legitimate interests, and GDPR Articles 6 and 9 for users in the EEA and the UK. Data may be disclosed to affiliates, service providers covering cloud, payment, support, analytics, marketing and IT, commercial partners where consent is given, public authorities, and an acquirer in the event of a business transfer. The control left to the user is real but narrow: section 10.12 lets you decide whether your data feeds model training and withdraw from specific types of training, and you can export your data in a machine-readable format from any privacy layer, or ask for its deletion, including under the right to be forgotten, across all layers and backups.
Data retention & training
Hosting summary
Sensay is headquartered in Singapore and says it uses cloud infrastructure and service providers around the world. No hosting country and no hosting region is named anywhere in the privacy policy, and the only geographic commitment on offer is a data residency option in specific regions, reserved for enterprise customers, with no region listed. Where data leaves Singapore, Sensay says it secures protection comparable to the PDPA through an enforceable transfer agreement, binding corporate rules, certification or a legal exemption; EEA and UK data relies on the European Commission's standard contractual clauses or the UK addendum. On the technical side the policy describes AES-256 encryption at rest, TLS in transit and regular key rotation, data centers under 24/7 surveillance with biometric access control, and encrypted backups stored in geographically separate locations. Client-hosted Sensay nodes running fully homomorphic encryption are presented as forthcoming, not available. One caveat on independent signals: the marketing site's public IP resolves to an Amazon anycast node in Amsterdam, which says nothing about where the application and its knowledge bases actually run.
Where Sensay works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting Sensay.
- The license you grant on User Content is perpetual, irrevocable, sublicensable and usable for commercial purposes (ToS 4.3). Read that clause before feeding in anything sensitive, because it survives the end of the contract.
- The output is framed as a digital replica of a person. Think about consent, about what happens to that replica once the colleague has gone, and about the temptation to question a bot instead of talking to the people still in the building. A take-down procedure for unauthorized replicas exists, but it has to be triggered by hand by email.
- The positioning is inconsistent: the Investors page sells a 24/7 sales and support chatbot for real estate and e-commerce, not the offboarding platform on the home page. Confirm which product you are actually buying.
- Pricing is under-documented: $500 with no currency code, from a Singapore publisher, billed per knowledge base with no published definition of what one counts as.
- Documentation and legal pages are unreliable in places. docs.sensay.io returned a 404 on 24 August 2026 despite being linked from the About page; the terms carry clauses inherited from another product, with streaming courses and a health and wellness disclaimer; the footer says 2025 while the take-down and CSR pages date from 18 November 2024; PDPA requests and consent withdrawal are routed to info@sensay.ai, a different domain.
- Security claims need qualifying: no subprocessor list, no customer-facing DPA, no SOC 2 or ISO 27001, no hosting country or region named, data residency reserved for enterprise customers, end-to-end and client-side encryption promised only 'where possible', and fully homomorphic encryption still forthcoming.
- Any dispute falls under Singapore law with SIAC arbitration in Singapore, a real cost and distance for a European or American buyer.
Setup & Integrations
Technical difficulty
Low on the administrative side. Creating the organization's account is a four-step guided setup announced as a matter of minutes, and a published customer testimonial says the rollout to employees took minutes. Sensay claims two days between installation and the first interview. Administrators declare who is leaving and Sensay schedules the sessions, with no technical skill required. Employees simply hold a conversation, by text or voice, with no format imposed. The Slack and Microsoft Teams integrations need enabling for day-to-day querying. A technical route exists but stays optional: REST API, JavaScript and Python SDKs, webhooks and an authentication flow.
Deployment
Integrations
Behind Sensay
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Sensay actually do?
How much does Sensay cost?
Is there a free plan or a free trial?
How do I sign up?
Where does the team consult the captured knowledge?
Is there an API?
Is my data used to train the models?
How long is data kept?
Is there a minimum age?
Who publishes Sensay, and how is an unauthorized replica removed?
Should you pick Sensay?
Sensay aims at a real problem that almost nobody tools: the tacit knowledge that walks out with a person. Its answer, a conversation with an AI interviewer, by voice or text, at the leaver's own pace, is genuinely different from a handover document written in a hurry on the last Friday, and the choice to serve the result back inside Slack, Microsoft Teams or the intranet rather than in yet another portal is a sound one. Three outputs per interview, full record, summary and conversational agent, give the material more than one way to be useful.
The pricing is legible but thinly documented: $500 per knowledge base per year, with no currency code on the page and no published definition of what a knowledge base counts as. There is no self-service route, so any evaluation goes through a sales conversation. The security posture reads seriously on paper, with AES-256, TLS, role-based access control and a documented training opt-out, but no certification is claimed, no subprocessor list is published and no hosting country is named. One contractual reservation deserves a lawyer's eye: the license granted on User Content is perpetual, irrevocable and sublicensable, commercial purposes included.
The publisher is young, Sensay Pte Ltd in Singapore, company number 202516659M, funded by a $3.4 million pre-seed in April 2024. And one inconsistency is worth naming before signing anything: the Investors page describes a 24/7 sales and support chatbot, a different product from the offboarding platform sold on the home page. Ask which one you are buying, in which currency, and where the data will sit.
- Choosing a selection results in a full page refresh.
- Opens in a new window.