The right AI lead-qualification partner for a real estate agency is one that responds to inbound enquiries quickly and consistently, connects to the systems you already use, handles the languages your buyers speak, and can show you a documented GDPR and EU AI Act position before you sign. Everything else — dashboards, "proprietary AI," polished demos — is secondary. This checklist gives you the specific questions to ask and what a good answer looks like, so you can compare vendors on the same terms and decide with confidence.
The 10-Point Buyer's Checklist
Use this as a scorecard. Ask each shortlisted vendor the same questions and record the answers side by side. The systems named below are examples of what an agency may want to connect, not a statement of what any vendor supports.
| # | Evaluation criterion | What "good" looks like | Disqualifying answer |
|---|---|---|---|
| 1 | Speed to first response | A response time the vendor commits to in writing and you can verify in a trial, covering the channels you use (web forms, WhatsApp, portals) and out-of-office hours | "Within a few hours" or business-hours only |
| 2 | CRM integration depth | Two-way sync with the CRM you already use (for example HubSpot, Salesforce, Pipedrive, Zoho or Inmovilla), with field mapping you can inspect | "We can export CSVs" |
| 3 | Portal & channel coverage | Lead capture from the portals you actually receive enquiries from (for example Idealista, Immobiliare, SeLoger, ImmoScout24, Rightmove or Funda), plus WhatsApp, Meta and Google forms, confirmed for your own set-up | Website chatbot only |
| 4 | Multilingual qualification | A clear list of supported languages, tested by a native speaker on your own sample enquiries | English-only, or machine-translated scripts nobody has reviewed |
| 5 | Scoring transparency | You can see and edit the criteria behind each score | "Our model is proprietary" |
| 6 | GDPR & AI Act posture | Clear answers on where data is hosted, a data-processing agreement, a sub-processor list, how Art. 22 is handled, and how the AI disclosure is made | Vague reassurance, or no clear answer on where data is stored |
| 7 | Human handoff logic | Configurable rules on when a human agent takes over, with the conversation context passed on | Bot runs until the lead drops off |
| 8 | Reporting you can act on | Lead-to-appointment rate, response time, source-level conversion, cost per qualified lead | Vanity metrics: "conversations handled" |
| 9 | Implementation plan & ownership | A written project plan with dated milestones, a named implementation lead, and documented workflows you own | An open-ended timeline or undefined scope |
| 10 | Commercial terms | An itemised written quote, and clear answers on minimum term, notice period, and data export and deletion on exit | Unclear charges, or no data portability |
Why Speed to Lead Matters
The best-known evidence on response speed is a 2011 Harvard Business Review study (Oldroyd, McElheran and Elkington, "The Short Life of Online Sales Leads," HBR, March 2011). The researchers sent test web leads to 2,241 US companies. Among the companies that responded within 30 days, the average first response took 42 hours. Only 37% replied within an hour, 24% took more than 24 hours, and 23% did not respond at all. Companies that contacted a lead within an hour were nearly 7 times as likely to qualify it as companies that took longer, and more than 60 times as likely as companies that took 24 hours or more.
Two cautions apply. This is US data from 2011, not European or real-estate data, and it measures the odds of qualifying a lead, not sales. Whether the same pattern holds in your market is something you can test yourself.
Response speed is also a practical evaluation tool. When you run a vendor trial, submit test enquiries at 22:40 on a Saturday and at 07:15 on a Monday. Measure the actual response time, not the promised one, and agree in advance what response time you would consider acceptable for your agency.
How the Options Actually Compare
There are several types of provider on the market, and each comes with different trade-offs. The table below compares them qualitatively; your own quotes and trial results should decide.
| Model | Strengths | Weaknesses |
|---|---|---|
| Generic chatbot SaaS | Low barrier to entry, quick to install, self-serve | May offer limited qualification logic and CRM depth, and you build and maintain the set-up yourself |
| Offshore call centre / BDC | Human nuance, phone-first | May be limited by opening hours and language coverage; script consistency depends on training, and staff turnover can affect quality |
| Real estate point solution | Industry vocabulary built in | Workflows may be less flexible, and portal coverage may be narrower outside its home market |
| Consultancy-led custom build | Can be built around your CRM, portals, languages and handoff rules, with workflows documented and handed over | Needs discovery time up front; not a same-day install |
Price is best compared on your own numbers. Ask each vendor for a written, itemised quote covering setup, monthly fees, usage-based charges and exit terms, then compare on cost per qualified appointment using your own enquiry volumes and conversion figures rather than on headline monthly fees. Which type fits best depends on your enquiry volume, the number of languages and offices you operate in, and how much of the set-up you want to own.
Compliance Questions to Ask (EU Specifics)
These are questions to raise with any vendor and with your own data-protection adviser. This section is general information, not legal advice. Get the answers in writing during procurement, not after.
GDPR Article 22 — automated decision-making. Article 22 restricts decisions based solely on automated processing that have legal or similarly significant effects. Ask the vendor how scoring is used in your set-up: does a human review before a lead is disqualified, and is that review logged? Whether a particular use falls within Article 22 is a question for your own adviser. One design option is to route low scores to a nurture sequence rather than deleting them.
Data residency and transfers. Confirm where conversation transcripts and personal data are stored and which sub-processors (including LLM providers) touch the data. Ask for the sub-processor list and, if any processing happens outside the EEA, the transfer mechanism relied on (for example, Standard Contractual Clauses). Under GDPR Art. 83(5), fines can reach up to €20 million or 4% of total worldwide annual turnover, whichever is higher.
EU AI Act transparency. The AI Act entered into force on 1 August 2024. Since 2 August 2026, Art. 50(1) requires providers of AI systems designed to interact directly with people to make clear that the person is interacting with an AI system, unless that is obvious from context. Breaches of Art. 50 can be fined up to €15 million or 3% of worldwide annual turnover. Ask any vendor how this disclosure is handled in its product, and look closely at any vendor that suggests presenting the bot as a human agent.
Consent and marketing. Rules on contacting people by WhatsApp, SMS and email depend on the country, the channel and how consent was obtained. Ask the vendor how consent and opt-out are recorded, and whether that record syncs to your CRM, and take advice from your own adviser.
Red Flags in Vendor Demos
- The demo uses their data, not yours. Ask for a pilot on a sample of your own historical leads, handled under a data-processing agreement, so you can compare scores against known outcomes. This involves personal data, so agree the terms with your own data-protection adviser first.
- No definition of "qualified." A qualified lead should be defined concretely: budget range confirmed, timeline stated, financing status known, property type and area specified, contactable. If the vendor can't state their definition, their metrics are hard to interpret.
- Scoring you can't audit. You need to see why one lead scored high and another low. Opaque scoring makes it hard to improve, and harder to explain if someone asks how a decision about them was reached.
- No named owner on your side of the project. It helps to have one person in your agency who owns lead-flow decisions. Vendors that don't ask for this may be planning to deliver software rather than outcomes.
- Pricing per conversation with no cap. Portal spam and duplicate enquiries can inflate volume-based invoices. Ask about deduplication logic, usage-based charges, and whether a monthly cap is available.
Frequently Asked Questions
How long should implementation take?
It depends on scope: the number of CRMs, lead sources, languages and markets involved. Ask for a written project plan with dated milestones, for example discovery and qualification criteria, then build and CRM integration, then a live pilot on one lead source. Be cautious of timelines that are open-ended or scope that is undefined.
How should we compare prices?
This article doesn't quote price figures, because they vary widely between vendor types and between vendors. Ask each vendor for a written, itemised quote covering setup, monthly fees, usage-based charges and exit terms. Then evaluate cost per qualified appointment using your own numbers, rather than the monthly licence fee alone.
Will this replace our agents?
That should not be the goal, and be wary of vendors who imply it will. The realistic aim is to reallocate agent time: qualification, follow-up chasing, and out-of-hours triage can move to automation, while viewings, negotiation, and relationship work stay human. Set your own targets and measure them against your own baseline.
What if we already have a CRM we don't want to change?
Ask whether the vendor can work with it. Being required to replace your CRM as a precondition for lead automation is a red flag, because it widens the project scope and the risk.
How do we measure success after launch?
Agree a review date, then track four metrics against your pre-launch baseline: median first-response time, lead-to-appointment conversion rate, percentage of leads with complete qualification data, and cost per qualified lead. Agree these targets in writing before signing.
The Decision Framework in One Sentence
Shortlist a small number of vendors, score them against the ten criteria above, run a time-limited pilot on a single lead source with your own historical data, and choose the partner who can demonstrate fast response times in your trial, transparent scoring, and clear answers on EU data handling — not the one with the most impressive slide deck.
If you'd like to see whether automation makes sense for your agency, you can take the short, free fit check on the Altamira site.