Opendoor
AI agent for qualifying and converting inbound leads
In the second quarter of 2026, Opendoor states that its AI assistant Chloe, embedded in the sales teams and available 168 hours a week, converts three times more than in January 2026 at a third of the cost.
Key points
- Opendoor embedded its AI assistant Chloe in the sales teams that handle seller leads.
- Chloe converts 3 times more than in January 2026, at a third of the cost.
- She is available 168 hours a week and explicitly declares herself a bot.
- Evidence level A: figures announced by the CEO on the Q2 2026 earnings call (August 4, 2026).
Objective
Convert more seller leads without growing the sales team, by handing an AI agent the continuous handling of inbound contacts and keeping human salespeople on the cases that require them.
The deployment
Chloe is Opendoor's AI assistant. On the second quarter 2026 earnings call, CEO Kasra Nejatian presents her as a full member of the company's sales teams, no longer as a side tool. She handles inbound contacts continuously: the figure communicated is 168 hours a week, the entire week. She tells the person she is talking to that she is a bot. For the quarter, Opendoor states that Chloe converts three times more than in January 2026, at a third of the cost. The comparison is about Chloe's own performance between January and Q2 2026, not about a gap with a human salesperson: Opendoor publishes neither the volume of leads handled, nor the absolute conversion rate, nor the base used to compute the cost. The same quarter saw Opendoor push several other AI components (an auto-underwriter on roughly one valuation in five, Ops Hub, Opendoor 2.0 for agents), but the conversion and cost figures cited here are attributed to Chloe alone.
Results Proof A
Figures stated by Opendoor's CEO during the Q2 2026 earnings call of August 4, 2026, within financial reporting, and available in verbatim transcripts from two independent publishers. The figures are explicitly tied to the Chloe deployment, not to the company's overall results.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- outil Chloe Opendoor's AI assistant, described by the CEO as a full member of the sales teams. The system explicitly declares itself a bot to the people it talks to.
How it runs, concretely
For ops teams-
1Seller lead arrives customer
A homeowner contacts Opendoor to sell their property.
-
2Chloe takes the case AI
The AI assistant takes the contact and states that it is a bot. Handling does not depend on the working hours or the absences of a human team.
-
3Moving the case toward the sale AI
Chloe moves the lead forward the way a member of the sales team would; that path is what gets measured as conversion.
-
4Performance tracking data team
Opendoor tracks the conversion and the cost of the deployment over time, and compares the current period with January 2026.
The conversion of inbound seller leads. That is the metric Opendoor uses to measure Chloe (three times more than in January 2026). Without tracking what happens to each lead, the gain is neither steerable nor demonstrable.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- History of inbound leads with their outcome, to measure conversion before and after
- Qualification data on the property and the seller, readable by the agent
- Traceability of exchanges so that what the agent said can be audited
Org prerequisites
- Sales team ready to share its lead pipeline with an AI agent rather than receive it filtered
- An accepted transparency rule: the agent declares itself a bot
- A clear handoff rule toward a human
- Tracking of the deployment cost relative to conversion, not to message volume
Possible stack
- Conversational agent wired to the inbound contact channel
- CRM or lead database readable and writable by the agent
- Conversion measurement layer by lead cohort
The plan, step by step
- Step 1Establish the baseline conversion of inbound leads, by source and by time of arrival.Deliverable: Quantified baseline against which any gain will be measured.
- Step 2Put an AI agent on a subset of leads, with an explicit declaration of its bot nature.Deliverable: Agent in production on a restricted scope, with a log of exchanges.
- Step 3Define the handoff point to a human salesperson and what the agent passes on at that moment.Deliverable: Handoff rule written and applied in the tool.
- Step 4Compare conversion and cost on the agent scope with the baseline scope, then expand.Deliverable: Conversion and cost read by cohort, decision to extend.
First step: Measure the current conversion of inbound leads by time slot, to isolate what waiting outside business hours costs before putting an agent on it.
The same pattern, proven in other industries
Amazon
genAI shopping assistant
Amorepacific
genAI beauty advice agent built on a skin diagnostic
Booking.com
genAI trip planning assistant
Sources
- S1 Opendoor (OPEN) Q2 2026 Earnings Call Transcript Primary archive pending
- S2 Earnings call transcript: Opendoor Q2 2026 misses estimates as shares fall after hours Secondary archive pending
- S3 Opendoor 2Q26 Financial Open House on August 4th, 2026 Primary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.