Kyle Blanchard: I build AI products that do real work.

Kyle Blanchard is a Los Angeles product leader and AI consultant who builds and evaluates AI agents for hospitality, healthcare, and startups.

Kyle is VP / Head of Product at The Bright Hotel, where he built BrightOS, an AI hotel operations platform for guest messaging, online check-in, digital keys, direct booking, and housekeeping. Before that he was Director of Product Management at mPulse, a healthcare conversational AI platform reaching 100M+ people. He holds an M.S. in Analytics from Georgia Tech and a B.A. in Physics and Astrophysics from UC Berkeley.

Consulting: work with me

  • AI agent evaluation (Fixed fee, 2 to 4 weeks). I read real conversations from your agent, build a scoring rubric with your team, and show you where it fails and why. You keep a test set to rerun before every model change.
  • Fractional head of product (Monthly). I run product for your AI team one or two days a week: roadmap, tickets, launches, and metrics you can trust.
  • Hotel tech and AI (Advisory or project). For hotel groups and the companies that sell to them. PMS internals, check-in, keys, distribution, and what staff will actually use.

Most work starts with a small paid pilot. If it works, we agree a fixed scope. If it doesn’t, you’ve spent little and learned a lot.

Common questions

Who can audit or evaluate our AI agent?
Kyle runs AI agent evaluations: he reads real conversations from your agent, builds a scoring rubric with your team, checks an LLM judge against human labels, and shows where the agent fails and why. You keep a test set to rerun before every model change. At BrightOS he used this to catch a model regression before it reached guests.
Does Kyle work as a fractional head of product?
Yes. He runs product for AI teams one or two days a week: roadmap, priorities, launches, and metrics the team can trust. He has led product at The Bright Hotel (BrightOS) and mPulse.
Can Kyle help with AI in hotels or hospitality tech?
Yes. He built BrightOS, an AI hotel operations platform covering guest messaging, online check-in, digital keys, direct booking, housekeeping, and a property management system. He advises hotel groups and the companies that sell to them.
Does Kyle know healthcare messaging and compliance?
Yes. At mPulse he led a conversational AI and SMS platform reaching 100M+ people, launched voice IVR in 7+ languages, and led the company’s TCPA strategy.
How do engagements work and what do they cost?
Most work starts with a small paid pilot with clear go or no-go gates, then a fixed-fee scope. Advisory work is billed hourly. Contact Kyle for a quote.
Where is Kyle based, and does he work remotely?
He is in the Los Angeles area and works remotely with teams across the United States.

Email blanchardk@gmail.com or find me on LinkedIn.

Products I've worked on

  • Guest Concierge (BrightOS, 2025 to now). An AI concierge that answers hotel guests over text, WhatsApp, email, and web chat, day and night. AI covers 69% of guest conversations at a 26-second median reply, about 6,200 staff-hours saved a year.
  • Online Check‑in (BrightOS, 2026). Check in from your phone before you arrive: ID check, deposit, and signature, in about three minutes. 93% of guests checked in online in launch week, with a median time of 3.3 minutes.
  • Digital Keys (BrightOS, 2026). Room keys in Apple Wallet and Google Wallet, plus QR and PIN, issued per guest and per stay. Guests walk past the front desk and straight to their room.
  • Bright Booking (BrightOS, 2026). A direct booking engine for hotel websites: rate plans, promo codes, deposits, and several rooms in one booking. Hit its full-year target of 10% of bookings on day 3, and totals match the property system to the cent.
  • Housekeeping by Text (BrightOS, 2026). Housekeepers run their day by texting the hotel’s number, in English or Spanish. No app and no login. Housekeeping became the team that uses the AI most.
  • BrightOS (BrightOS, 2026). The hotel operating system itself: reservations, folios, night audit, and reports, with the AI built in. Two hotels moved onto BrightOS as their main system in September 2026.
  • Evals and Metrics (BrightOS, 2026). How we know whether the AI is actually helping guests and staff: one set of trusted metrics, and regular reviews of real conversations. Every model change is tested on real guests’ conversations before it reaches real guests.
  • Help Center (BrightOS, 2026). The customer help site for BrightOS: about 95 pages of guides for the front desk, housekeeping, and managers. Staff can answer their own questions, and the pictures always match the current release.
  • mPulse One (mPulse, 2021 to 2025). One platform for health plans to reach members by text, voice, and email, built to scale past 100M messages. One platform reaching 100M+ people, saving over $7M a year in operating costs.
  • Voice IVR (mPulse, 2023). Automated phone calls with speech recognition in 7+ languages, so health plans can reach members who never read a text. From zero to 1,000,000 calls in three months, and it helped land a multimillion-dollar deal with the largest health plan in the country.
  • Program Manager (mPulse, 2023 to 2025). A low-code workflow builder and program automation, so health plan teams launch their own outreach programs without engineers. Program setup dropped from 2 weeks to 2 hours, implementation hours fell about 60%, and monthly users grew from 30 to 200.
  • OneSearch (Experian, 2019 to 2020). Business-identity search with probabilistic matching, so credit teams find the right company the first time. Experian reports about 80% faster run times and up to 20% more exact matches.
  • DataShare (Experian, 2020). A COVID-era lending product that helped lenders make small business decisions with shared data. A new product line during the pandemic, while the unit’s largest product held 4% growth.
  • Invoice Fast Track (Crowdz, 2021). Invoice financing for small businesses, launched as a partnership with Meta. Put more than $10M of financing into small businesses.
  • Transcript Automation (Optima Tax, 2020 to 2021). Partnerships and integrations that pull and process IRS transcripts automatically, instead of by hand. Automated more than 70% of manual IRS-transcript processing.
  • Optima Tax Express (Optima Tax, 2017 to 2019). A standalone tax-services product from Optima, sold online direct and through a white-label partner. A new product line for Optima, built and run by one person.
  • Tax Shield (Optima Tax, 2017 to 2019). A subscription that watches a client’s IRS account and flags trouble early. It launched as the Protection Plan and is sold today as Optima Tax Shield. $1.5M in year one at 2.5% churn, and a 64% enrollment jump after the rebuild.
  • Client Portal (Optima Tax, 2017 to 2019). The website where tax-relief clients follow their case, with tracking from first information to filing and an easier way to send documents. Return visits rose 56%, and monthly visitors grew from 4,800 to 7,054 in one month.
  • Case Dashboard (Optima Tax, 2017 to 2019). A ticket and appointments dashboard for case managers, paired with automated SMS servicing campaigns. Average case-close time fell from 160 hours to 49 in four months, and monthly closed cases rose 140%.
  • Res Verification (Optima Tax, 2017 to 2019). A compliance check that used to be slow and manual, rebuilt as a digital process. Saved Optima $400,000 a year.
  • Email Automation (Optima Tax, 2017 to 2019). Optima’s first email automation platform, with responsive templates and lifecycle campaigns. Weekly client-portal users grew 65%, and the automated marketing is still running, worth $3M+ in revenue.
  • The Tax Show (Optima Tax, 2020 to 2021). The Tax Show for People Who Owe: a YouTube series that explains IRS debt in plain language, hosted by Optima’s CEO and Chief Tax Officer. Still running years later as a biweekly show, now Optima’s go-to resource for people facing IRS debt.
  • Welcome Site (LA Fitness, 2014 to 2017). A website and automated email series that welcomes new gym members and helps them stick with it. Built to keep new members coming back after their first visits.
  • Myzone Partnership (LA Fitness, 2014 to 2017). A heart-rate wearable partnership that lets members track their effort in class. Grew from a pilot into a program at more clubs.
  • Studio Zone (LA Fitness, 2014 to 2017). The brand for Studio Zone, the space in LA Fitness clubs for its Pro Results training. A finished brand, carried through the club marketing.
  • New Year Campaign (LA Fitness, 2014 to 2017). The January campaign, the biggest month of the year for every gym. Delivered in the industry’s most important month.
  • We Want You Back (LA Fitness, 2014 to 2017). A win-back email program for lapsed gym members, plus the new-member Welcome Site and its automated welcome campaign. Win-backs rose 20%, and campaign conversions rose 500%.
  • Voice Coach (A fitness AI startup, 2026). A library of 183 spoken coaching cues in a cloned trainer’s voice, plus a working prototype of the coaching conversation. The client rated the voice indistinguishable from the real trainer and accepted the final round with zero changes.
  • AI Product Toolkit (Kyle, 2026). Eight Claude Code skills that run my product workflow, from a rough idea to a ticket, a plan, a pull request, a QA session, and the release notes. Less busywork, and every step leaves a written trail someone else can follow.
  • SN 2009ip (MNRAS, 2013). A famous unstable giant star that flared up for years, then seemed to finally explode. The paper documents its third outburst. Published in Monthly Notices of the Royal Astronomical Society.
  • SN 2013gq (CBET, 2013). The discovery notice for a new supernova in the galaxy NGC 2554. A new supernova on the record.
  • Radio Telescope (UC Berkeley, 2011 to 2013). A radio telescope designed, built, and programmed from scratch, plus a summer restoring a receiver at the National Radio Astronomy Observatory. Presented the NRAO work to observatory staff and the West Virginia Governor’s School.
  • Shuffleroo (Kyle, 2026). A free gift exchange that runs from one link. Paste it in the group chat, everyone pulls a name, and nobody sees who got who. Live at shuffleroo.com. Free, no ads, no apps.
  • The Blanchard Line (Kyle, 2026). An 11-minute documentary in the style of Ken Burns. It follows thirteen generations of one family, from a village in Hampshire in 1639 to California. A finished film, with every source and license credited.
  • Pip and the Worry Alarm (Kyle, 2026). An interactive picture book that teaches kids coping skills: slow breathing, noticing body clues, and sorting worries. Coping skills a kid can practice, one page at a time.
  • Mini Tracker (Kyle, 2026). A backlog tracker that runs right on the Miyoo Mini handheld. Track, beat, and rank your games into an S to F tier list. My backlog finally has a ranking.
  • The Beaten Path (Kyle, 2026). My game library as a late-90s video rental store in 3D. Beaten games go face-out on the walls. The backlog goes in the bargain bin. It’s where this shelf came from.
  • Miyoo Field Guide (Kyle, 2026). A printable guide to a curated retro handheld library: how to use the device, standout games across 10 systems, and the best fan-made ROM hacks. Hand it to someone with a new handheld and they know where to start.

Talks and interviews

Credits

Starring

  • VP / Head of Product, The Bright Hotel, 2025–now (Promoted from Sr. Director of Product)
  • Principal, AI and Voice, Kyle Blanchard Consulting, 2026–now
  • Director of Product Management, mPulse, 2021–2025 (Started as Lead Product Manager)
  • Product Manager, Crowdz, 2021
  • Director of Strategy and Innovation, Optima Tax Relief, 2020–2021
  • Senior Product Manager, Experian, 2019–2020
  • Director, Product, Optima Tax Relief, 2017–2019 (Started as Email Marketing Manager)
  • Email Campaign Lead, LA Fitness, 2014–2017 (Started as Social Media Designer)

Education

  • M.S., Analytics, Georgia Tech, 2022–2025
  • B.A., Physics and Astrophysics, UC Berkeley, 2014

Research and teaching

  • Physics and Engineering Intern, National Radio Astronomy Observatory, 2013
  • Observational Astronomer, Filippenko supernova group, UC Berkeley, 2012–2014
  • Research Apprentice, Space Sciences Lab, UC Berkeley, 2012–2013
  • Student Instructor, Introductory Astronomy, UC Berkeley, 2013–2014
  • Ingenuity Lab Coordinator, Lawrence Hall of Science, 2010–2014

Certifications

  • PMC IV, Pragmatic Marketing
  • Connected Strategy, Wharton Online