Deepika Salwankar

One

I explain
complicated
things.

Mostly AI and enterprise software.

Mostly positioning, launches, and the words that carry them.

Mostly by talking to people. Product, sales, design, the customer, the person who built the thing.

Before this I fixed teeth. Turns out both jobs involve reducing pain.

Things I've helped people understand

Positioning, launch narratives, websites, video, thought leadership.

Botanical plate — Polypodium vulgare
Portrait of Deepika Salwankar
Woodblock print — bird perched among leaves.
Pencil drawing — white-throated kingfisher.

the buyer's morning
is the story.

Plate i

Two

How I got
here.

I became a dentist because I loved understanding systems. Eventually I realized I cared more about why people make decisions than how teeth worked. Early on I wrote for doctors and worked with KOLs. That taught me how experts read, and what they skip.

Marketing turned out to be the perfect excuse to study psychology, technology, business and storytelling all at the same time, without having to pick.

Over the last twelve years I've launched enterprise software in regulated industries, mostly AI and customer engagement platforms. I've rebuilt websites, argued over positioning and rewritten messaging until it sounded human. Most of that work came back to one question: what does a buyer need to read before they'll take the meeting?

“Most enterprise software fails to sell because nobody explains it properly.”

— from a working note, undated

That's my favourite part. The gap between what a product does and what a buyer, a seller and the team that built it each think it does is the room I work in.

A good explanation translates between two people who both think they are being clear and both, privately, are not.

Today I work where product, research and storytelling overlap: positioning, launch narratives, websites and the campaigns that carry them, for enterprise software teams selling something genuinely complicated.

Email Deepika — →

Three

Selected remarks · on the record

LinkedIn ↗

"Deepika has a gift for taking highly technical or abstract concepts and making them accessible, engaging, and relatable."

WL

Whitney Labonte

Growth Marketing Strategist & Leader · Worked with Deepika at Aktana

04 / 04Rotating exhibit

Four

Things I've
built.

Same structure every time — problem, discovery, thought, change, outcome — because that is how the work actually happens.

Category Creation. Tact.ai
Tact.ai — pharma customer engagementPlate I

I.

Category Creation. Tact.ai.

Tact.ai was built on conversational AI and agentic capability, aimed at a broad market. We pivoted and focused on pharma exclusively. The problem was that pharma customer engagement did not exist as a category. It sat inside Veeva and Salesforce, and if a buyer wanted it, they searched for Veeva.

Problem
Buyers wanted this. They did not know how to ask for it. The tools they already had were built as process systems, not for CX, and they did not give field teams the right insight at the right moment, which is where the last mile of pharma engagement actually happens. There was a want with no words attached to it and no category to put it in.
What I discovered
This was a positioning problem before it was anything else. I fixed the positioning in a deck, which was the easy part. The harder part was that nobody was talking about the problem anywhere the buyer would encounter it. The deck could not do that work. The narrative had to exist in public.
How I thought
Engineer the narrative and then put it everywhere the buyer looks. Website, blogs, event stages, all running the same argument. Not a content calendar. A single position, published repeatedly, until the words were ours and the search results agreed.
What I changed
Rebuilt the website narrative around the category. Ran the blog as the argument rather than as SEO filler, and it earned the search authority anyway. Rewrote the event narratives to carry the same position. Backed all of it with customer stories, because a category with no proof is a claim.
Business outcome
Tact.ai led search results for pharma customer engagement, and Gartner and Everest covered the category. Competitors soon copied our names and product pages, which is the clearest signal you get that you have become the blueprint.
Aktana Website
Aktana.com homepagePlate II

II.

Aktana Website.

Everyone in pharma had decided they were agentic. Every competitor site said the same thing. None of them explained how, and none of them said what it would actually do for the buyer.

Problem
The site talked about the technology in depth. Product owned the messaging, not marketing. So the site explained model architecture to commercial leads who did not care about model architecture. Features, not outcomes.
What I discovered
Nobody in the interviews could tell me what the product did. Not one person. They could describe the meeting where they wished it existed. I went through the competitor sites and found the same gap. They said agentic. They did not say how, and they did not say what changed as a result. That was the part my buyers wanted.
How I thought
Build the site around the meeting, not the machine. Every page had to answer a question a buyer had already asked out loud, in the words they used. If I could not attribute a sentence to someone, it came out.
What I changed
New site, new narrative, all copy rewritten around distinct commercial roles. Cut most of the pages. Kept one page on product architecture and gave it to product to write. The rest of the site ran on the commercial story, the same one we were already using in the ABM campaigns.
Business outcome
Qualified pipeline from the website grew sharply. Sales stopped opening calls by explaining what the product did.
Product Adoption and PLG
Product adoption — pharma field teamsPlate III

Adoption across a pharma field force.

III.

Product Adoption and PLG.

A pharma client had rolled Tact.ai out to its whole field force. Most reps weren’t using it.

The contract was signed. The product was not.

Problem
Onboarding had been built for the demo, not for a rep’s first Monday.

The nudges were generic. And nobody really owned the question of whether anyone was opening the thing, because the deal was already done.
What I discovered
I followed the journey from first login to regular use and marked where reps stalled.

The people who adopted had something in common. Someone on their own team had shown them how to use it.

Not us. Them.
How I thought
Adoption is a funnel with a product problem inside it.

The first session has to be useful before it is effortful. The nudge has to arrive when the rep needs it, not when marketing remembers. And the people who make it stick have to be theirs.
What I changed
Rebuilt onboarding around the first useful moment. Reworked the in-app nudges around real workflow triggers. Fixed the UX problems that were costing us the second session.

Then built an internal champion network and gave them something worth carrying.
Business outcome
Adoption climbed across the whole field force.

The referrals and case studies that came out of it became the expansion story inside the account. They also became proof we could take to the next pharma buyer.
IIM Bangalore Women’s Startup Programme
NSRCEL — Women Startup ProgrammePlate IV

IV.

IIM Bangalore Women’s Startup Programme.

A national startup programme that had always been run by hand. We won an education client and decided the entire competition would sit on our platform instead. Product-led growth with an event as the front door. It turned the platform into a national user base.

Problem
The programme was manual. Applications, screening, cohort management, all of it. No product existed for this because nobody had run a startup competition at that scale on a platform before. And a competition is a one-time event. If people applied and left, we had bought traffic, not growth.
What I discovered
I ran discovery the same way I would for any product. Interviewed every stakeholder who had run the programme by hand: the people screening applications, the people managing cohorts, the people chasing incomplete forms. Nobody had a clean answer for how the process worked. Everyone owned a fragment of it, and the fragments did not fit together. I mapped what they actually did against what they said they did, found the gaps, and turned that into a real process. Then prototyped it, took the prototypes back to the same people, and rebuilt where they told me I had it wrong. That process became the product, and the friction points I found in discovery became the onboarding.
How I thought
The application is the acquisition channel. The product has to do the retention. Thousands of women were going to arrive on a platform most of them had never used, and the difference between a spike and a user base was whether the product gave them a reason to come back after they submitted.
What I changed
Owned product development end to end: roadmaps, prototypes, worked with the tech team through build. Then owned adoption on the other side. Onboarding flows, in-product usage prompts, and product tips sent to every applicant as they moved through the funnel, all of it aimed at getting them into the platform rather than past it.
Business outcome
Women applied from across India, and the platform’s user base grew many times over. Women entrepreneurs used the platform at scale for the first time.

Currently

What I'm up to, right now.

Currently

Bird watching or vibe coding.

Readings · this quarter

  1. 01The Savage Mind — Claude Lévi-Strauss
  2. 02The Molecule of More — Lieberman & Long
  3. 03Isles of the Emberdark — Brandon Sanderson
  4. 04The Strength of the Few — James Islington
  5. 05The Last Contract of Isako — Fonda Lee
  6. 06Mortedant's Peril — RJ Barker

Five

Notebook.

The lead

Essay one — 4 min

Why every AI company sounds identical.

A short investigation into what happens when a category grows faster than its vocabulary. Includes a partial glossary, one apology, and a modest proposal.

Read the essay

Also in this issue

  1. 02

    How I research competitors.

    A short method. Involves a spreadsheet, a highlighter, and no dashboards.

    Research2 min
  2. 03

    Things dentists understand about product marketing.

    Reducing pain. Building trust. Explaining a procedure before you do it.

    Essay2 min
  3. 04

    What enterprise buyers actually read.

    It is not your blog. It is a specific list of eight things, in order.

    Field notes3 min
  4. 05

    Why messaging fails.

    Common reasons, none of them fixable by adding more words.

    Messaging2 min
  5. 06

    Notes from analyst briefings.

    Overheard, misheard, learned. A running set of margin notes.

    AR2 min
  6. 07

    What makes a great demo.

    The demo is the argument. If it isn't, cut the demo, keep the argument.

    Sales3 min
  7. 08

    Why adoption is a growth problem.

    Find where users stall. Fix it in the product. Then spend.

    Growth2 min

Last page

If you've made it this far, we'll probably enjoy working together.

I'm looking for senior product marketing and omnichannel engagement roles in life sciences, where curiosity matters as much as execution.

If you're building something ambitious, I'd love to hear about it.

Correspondence

Elsewhere

Colophon

Set in Cormorant Garamond and IBM Plex Mono. Printed on paper that does not exist.

Deepika S.

Deepika Salwankar · MMXXVI · fin