We built an AI research platform that monitors fine-art auction listings, gathers the important details, scores each opportunity against set criteria and delivers a focused shortlist for review.
Client: Optimal AssociatesIndustry: FinTech & Financial ServicesService: AI & Automation
Overview
Client
Optimal Associates
Specialist fine-art investment research business
Engagement snapshot
The challenge
Finding good opportunities meant visiting auction platforms every day, reading through large numbers of listings, researching artists, checking context and comparing prices.
It all relied on manual attention. In busy auction periods, strong lots could be missed simply because there wasn't time to review everything to the same standard.
With so many listings, it was hard to give every relevant lot the same attention.
Manual research varied with the time available and with whichever information the researcher found first.
Much of the day went on collecting and organising information before any real analysis could start.
The solution
The platform gathers the key details from auction listings automatically, so nobody has to copy and track them by hand.
AI reviews the available information using defined research rules, such as artist context, style, historical relevance and pricing signals.
The strongest matches go into a clear daily report, with the reasoning a person needs to review each opportunity.
Key features
AI checks every lot against research rules on artist, style and pricing, then gives it an opportunity score.
Business impact
100% listing coverage
The strongest lots go into a daily shortlist that explains each pick and links to the source listing.
Business impact
Daily opportunity reports
AI scores each listing against a repeatable set of business and market criteria, rather than writing generic summaries.
Business impact
More consistent analysis
Each shortlisted lot comes with the context behind the recommendation, so the user can decide which ones deserve a closer look.
Business impact
More useful shortlists
Collection, analysis and reporting run on a schedule, which removes much of the repetitive preparation.
Business impact
Less manual research
Before and after
Before
Hours of manual review
Researchers opened listings and gathered supporting information one at a time.
After
Automated first review
Listings are collected and analysed before the researcher starts the day.
Before
Depended on time available
Research was less thorough in high-volume periods.
After
The same criteria every time
Every listing is checked against the same defined research rules.
Before
More listings, more work
Covering more listings meant more hours of manual research.
After
Wider automated coverage
More listings can be reviewed without research effort growing at the same rate.
Results
90% Less Research Effort
Automatic collection and first-pass analysis replaced most of the daily manual review, so research takes 90% less effort and runs 90% faster.
Before: High Manual EffortAfter: 90% Faster
100% Automated First-Pass Review
Every collected auction lot is checked against the same research criteria before a person reviews the shortlist.
Before: PartialAfter: 100% Coverage
Easier Multi-Source Expansion
The workflow was built so more auction sources can be added without rebuilding the platform.
Before: Manual ExpansionAfter: Reusable Process
Testimonial
“Our analysts used to spend hours searching auction listings. Now we get a daily shortlist of scored lots, each with the reason it was flagged and a link to the source. Research effort is down 90%.”
Richard SanchezHead of Research, Optimal AssociatesKey results
Trust and control
Credentials used for automated research are stored and handled securely.
SECUREEach shortlisted lot stays linked to its source information and the reasoning behind it.
TRACEABLEScheduled jobs, retries and monitoring keep the daily research running.
RELIABLEThe client owns the custom research workflows, automation and code built for the project.
100% OWNEDFaster delivery
Because these parts were already proven, the team could spend its time on the research criteria and the daily report, and launch sooner.
Tested credential handling lets the platform use the research accounts securely during automated collection.
Our analysis templates were adapted to score lots on artist context, style, historical relevance and pricing signals.
Monitoring with automatic retries keeps the scheduled daily collection running and shows when a job needs attention.
A structured data model turns each listing into consistent fields, which is what makes new auction sources easier to add.
The platform collects the details of each auction listing, then uses AI to score every lot against defined research criteria: artist context, style, historical relevance and pricing signals. The strongest matches go into a daily shortlist, and each one includes the reasoning behind it, so the user can decide what deserves a closer look.
The platform cut research effort by 90%. Collecting listings and doing the first-pass analysis used to take hours of manual work each day. Both now run automatically before the researcher starts, and 100% of collected listings get the same first-pass review, not just the ones there was time for.
The user makes every investment decision. The platform only does the first pass: collecting listings, scoring them against the criteria and explaining why each lot was flagged. Source information and reasoning stay linked to every shortlisted lot, so the user can check the evidence before deciding which opportunities to pursue.
If your team reads through large numbers of listings to find the few worth a look, we can automate the collection and scoring and send you a shortlist you can check.