Search ResiDesk, work, writing, talks, and small tools.
Founder, builder, housing work work work
Resident texts, reviews, tickets, and calls
Arjun Kannan
I build software for teams that already hear from customers every day. The hard part is turning it into work someone can use.
At ResiDesk, that means turning resident texts, reviews, calls, and support threads into an answer, the right history, and a next step the property team can trust.
Talking with customers is still business 101. The hard part is noticing when the same problem is showing up in three places.
ResiDesk takes most of my week. Before that, I worked on outcome-based lending at Climb Credit and advisor tools at BlackRock. Different rooms, similar lesson: lose the thread and the product gets harder to trust.
I care less about a polished demo than the next day. The queue is full, the edge cases are real, and the person using the product has to live with it.
ResiDesk takes up most of my week. Before that: Climb Credit and BlackRock. I like work where the customer has already pointed at the problem.
The customer usually left a clue somewhere somewhere
Most teams already hear plenty from customers. It gets stuck in inboxes, tickets, calls, and support threads before it changes rent, renewals, maintenance, staffing, or the product itself.
The next day is where the test starts
I care about what survives once the meeting ends: the queue is full, the team is moving, and someone still needs an answer they trust.
Do not make the team guess twice twice
A good tool helps the team move faster without getting careless. It shows the policy, the history, the uncertainty, and the person who owns the next step.
The test is pretty simple
Useful software helps the person doing the work see the customer, what happened, and the next decision before the queue gets away from them.
Start here
[PATHS 03]
Start with the question you came here with
You probably came here with one question. Start there. The rest can wait for now.
We help property teams see what keeps happening while there is still time to do something about it.
Residents already tell you what is broken. The job is to answer them, notice what keeps repeating, and get it to the right person without asking someone to reread the whole thread.
What I look for
Teams do not need another place to search.
The best operators care about retention, NOI, workload, maintenance, and resident trust. The hard part is seeing the pattern and knowing who should act on it.
What I keep coming back to in writing
What happens after the answer is the part that matters.
AI can write a decent answer quickly. I still care who owns the next step, what they know, and whether the resident has to tell the story again.
Where I lose patience
An answer is not the work done.
If the answer goes out and nothing changes, I do not trust the product. That is not the work done; it is just a cleaner inbox.
Before ResiDesk
[WORK 04]
The jobs that changed the way I work
WhereMy partWhat changedMore
Climb CreditI was CTO and CPO. We put student outcomes into the product, data, and underwriting.Annual loan volume grew from $1M to $300M as the product moved closer to student outcomes.TechCrunch
BlackRockI worked on product and engineering for advisor tools. Interface quality mattered because real money sat behind every decision.The advisor analytics product reached $40M ARR in its first year, with me working across product and engineering.Work history
ResiDeskI co-founded ResiDesk and spend most of my energy on data, product, and the ordinary work of running a company.Law360 covered a reported 7% lift associated with getting resident feedback into decisions sooner.Law360
How I try to work
[MODULE 04]
How I work
My default is simple: talk to the customer, make the work visible, and see what survives a busy day.
01 / Customer
Talk to the customer before asking a model to help.
If you have customers, understanding them is still business 101. In housing, the hard part is hearing enough residents without making one operator read every thread.
02 / Context
Show me the job as it really is.
Abstractions do not move a team. Show me the stakes, the weird cases, and the person who has to live with the result every day.
03 / Adoption
A demo is not adoption. Repeated use is the test.
I learned this at BlackRock: a prototype can win the room and still lose to the spreadsheet people already trust. The test is what they open the next morning.
04 / AI
Shorten the distance between a question and an answer someone can act on.
I do not need AI to handle everything. I need it to move something from stuck to almost done while a person still owns the call.
05 / Candor
Be direct without making the person feel small.
The best teams can name what broke without blaming the person who found it. Take the pain seriously, then separate what happened from what caused it.
06 / Team
Work with people who make the room clearer.
The best people I have worked with can walk into a mess, find the few facts that matter, and move. They make the room clearer by keeping the work simple.
AI in use
[MODULE 05]
What still works once people use it?
AI made it easier to produce a good answer quickly. It did not make it easy to change the work. I care about the policy, the handoff, the risk, and what happens after the answer leaves the box.
Field note
The real product starts when people use it.
I am most useful when a team has a rollout problem in real use: the launch worked, and the actual day still fights back.
01Watch the work happen
Spend time with the people doing the job. Watch what happens when the tool is slow, weird, or just one more thing to manage.
02Name the change
Check whether it saves time, lowers risk, protects retention, frees capacity, or earns trust. If nothing changes, the work is not done.
03Design around the work people actually do
Choose the model later. First decide what history it gets, what it can do, who checks it, and where the work goes next.
04Test the cases most likely to break it
Run the cases where it misses. Show the misses without drama. Then measure whether the work moved faster, got safer, or reached the right person.
05Put what you learn back into the product
Sales credibility comes from what actually happened: the objection, the before state, the number that moved, and the sentence the buyer can repeat without editing it.
$40M ARRBlackRock advisor analytics, year one
$300MClimb Credit annual loan volume
100+Startup investments and founder work
AI work in use
[MODULE 06]
The demo is only the start
Capability is only the first question. The buyer, user, reviewer, and executive sponsor usually worry about different things.
Field note
Find the work. Show where AI helps. Name the person who owns it.
A great demo can still lose to a spreadsheet. Information helps only when the person making the decision can act on it without decoding it first.
01
Start with the work itself
Find the narrow place where AI changes the day, not just the pitch. Talk to the buyer, the user, and the person who is stuck when it breaks.
02
Make the rollout something you can repeat
Know the job, staff the messy parts, test the risk, and put what you learn back into the product.
03
Tell the story of what really happened
Help the team say the true thing: what changed, why the buyer cared, what broke, and what still worked.
04
Keep the simple version true
Make it simple without making it vague. I care about the number, the risk, the owner, and the next move.
Messages from residents
[MODULE 07]
How a resident message becomes work someone can hand off
[STEP 01 / LISTEN]
Start with what the resident actually said.
A useful product starts with ordinary apartment problems: a broken washer, a pet-policy question, Wi-Fi complaints, package-room issues, and the early signs someone may not renew.
Texts every dayActual complaints
ResiDesk
[LOOP 08]
Why housing makes this concrete
Before
Resident feedback is everywhere, and it is still hard to use.
Texts, reviews, tickets, surveys, renewal notes, and maintenance complaints still live in different places. By the time an owner sees the pattern, the problem is usually old.
After
The right person sees the issue in time.
The product brings enough history together to answer, route, report, and show what the building can change while there is still time.
01Message
A resident text, review, ticket, call, survey, or renewal note comes in.
02History
Lease, policy, unit, prior messages, tone, and what already happened.
03Owner
The person or team that can change what happens next.
04Action
Answer, escalate, repair, explain, or change the rule.
05Report
What owners need to see about retention, NOI, workload, and risk while there is still time to act.
About
[MODULE 08]
I did not start in housing
I grew up around research, so the route looked academic at first. I studied applied physics at Cornell because it sounded hard, interesting, and close to experiments.
Software came in sideways. I was in an electron microscopy lab and wrote code to speed up a magnetic-noise setup. It saved hours quickly, and software stopped feeling theoretical.
The industries changed. The habit did not. At BlackRock, it meant making institutional tools usable for advisors. At Climb Credit, it meant underwriting against outcomes. At ResiDesk, it means helping housing teams hear residents clearly enough to act.
Physics
Software clicked when it gave me real time back.
I wrote a tool in an electron microscopy lab to speed up a magnetic-noise experiment. It saved enough time that software stopped feeling like coursework and started feeling useful.
BlackRock
Real stakes make the interface matter.
I did not get the job at first. Six months later I re-interviewed, moved to New York, and learned that interface quality matters when real money sits behind the decision.
Climb Credit
Outcomes changed the question.
Instead of asking who looked safest on paper, we asked how the graduate's earnings changed after the program. That pushed outcomes into underwriting, product, and data as annual loan volume moved from $1 million to $300 million.
ResiDesk
Housing can listen more carefully.
Residents tell buildings what works and what does not every day. The work is making that clear to owners, useful for operators, and less annoying for residents.
Work history
[MODULE 09]
How I got here
The settings changed, but the habit stayed similar: understand what the customer is saying inside a messy process, then build the simplest responsible way to help.
Resident feedback result
7%
That 7% is associated with getting resident feedback into decisions sooner. Law360 wrote up the outside version.
Climb Credit
$1M → $300M
Annual loan volume grew while outcomes moved into product, data, and underwriting.
Advisor tools
$40M ARR
Advisor-facing analytics product I helped take from zero to $40M ARR in its first year.
Co-founder; data, product, and the ordinary work of running the company
We help rental-property owners and operators understand what residents ask for across renewals, rent, maintenance, and staffing. The product earns its keep when the right person knows what to do before the next thread arrives.
We underwrote against a different question: not who looked safest on paper, but what happened to a graduate's earnings. That put outcomes into the product, data, and underwriting.
The job was turning institutional infrastructure into a product advisors could use in real conversations. The information was the same underneath, but the product had to help someone explain, compare, and decide.
More detail
[MODULE 10]
More detail, if you want it
What changed
I have worked on products where the number had to mean something.
BlackRock advisor analytics reached $40M ARR in the first year. At Climb, annual loan volume grew from $1M to $300M as outcomes moved into product and data.
If you want the longer version, start with the links.
TechCrunch covered Climb. Law360, HackerNoon, TechTimes, TechBullion, BuiltWorlds, and 20for20 fill in more of ResiDesk, applied AI, talks, and property-operations work.
I write when I am trying to make a thought less fuzzy. Most pieces come back to the same test: does this help someone finish the work, or did we just make the demo easier to sell?
If a tool does not help someone finish a real task sooner, with less dropped context, it is hard for me to care.
Understand the job first
If you do not know what someone is actually trying to do, you are probably just rearranging pixels.
Build around the work
The model is one part. The surrounding tools, guardrails, checks, and handoff into someone's day decide whether anything changes.
Demos leave things out
What matters is whether people still reach for it mid-work, when nobody is watching.
FAQ
[MODULE 14]
Quick answers
What kind of AI do you build?
I build AI around work people already have to do. At ResiDesk, that means helping property teams answer residents, understand what is happening in the building, and get the right issue to someone who can act on it.
What came before ResiDesk?
Before ResiDesk, I worked at Climb Credit and BlackRock. At Climb, I helped take annual loan volume from $1M to $300M. At BlackRock, I worked on an advisor analytics product that reached $40M ARR in its first year.
What do I usually come back to?
I usually come back to the same things: agents, checks, product loops, and the gap between a strong demo and something people still reach for on a busy day. Housing makes this concrete because residents are already telling you what broke.
How do I judge AI?
I care less about whether something looks impressive and more about whether it helps someone make a better call. That means getting the sequence right, testing what good looks like, and keeping a person close enough to stop the product from doing the wrong thing.
Investing
[MODULE 15]
Investing, when I can be useful
I have invested in more than 100 startups and mentored through Techstars. I tend to back founders who are close to the problem, close to the customer, and honest about what they still do not know.
Generic advice is everywhere now. The useful version is specific: the customer, the constraint, the ask, and the next decision.
You have real customer pain and need it to change the product, GTM, or operations.
You are putting AI into work where checks, handoff, and trust are not optional.
You are a housing operator trying to spot resident issues before they become churn, extra cost, or owner surprises.
Probably not the fit
You want generic AI inspiration without a real customer or real job behind it.
You need someone to bless a demo with no owner, no metric, and no next step.
You want a broad advisory call without a specific problem to make sharper.
Small tools
[MODULE 17]
Small tools if useful
01 / Page map
Local views ready
Map the work.
Pick a view. The graphic runs anywhere. If the browser has a local model, it can add a sharper read.
02 / Ask this site
Checking browser AI
Ask a real question.
The answer uses the copy, talks, writing, links, and tools already on this page. Try: "why ResiDesk?", "what works after launch?", "where should I start?"
Try asking about ResiDesk, founder advice, BlackRock, Climb, writing, or what happens after launch.
03 / Conversation map
Start with the question.
04 / Find what matters
Skip to the parts that matter.
05 / Talk lens
Pull the point from one conversation.
06 / Useful AI check
Paste an AI idea. Check the job.
07 / Tuesday test
Put a demo through a normal Tuesday.
Pick a demo promise and where it has to run. The tool shows what has to be true before it works on a normal day.
Run the test to see what the demo is likely to miss on a real day.
08 / Resident messages
Build a small owner readout.
09 / Pattern check
Show repeated words.
Highlights the words I use a lot here: customer, measurement, handoff, follow-through, trust, and demo.
10 / Page check
Check whether the page is clear.
This checks whether the page is clear, useful, and honest about the work.
11 / Next fix
Pick the next rough spot.
Pick the part that feels roughest and get one concrete next fix.
